__group__ ticket summary difficulty component version type owner status created _changetime _description _reporter Milestone 6.0.0 1146 create-temporary-file and create-temporary-directory are subject to race conditions medium core libraries 4.10.x defect new 2014-08-05T16:46:58+02:00 2023-11-07T00:23:48+01:00 "Both use the non-atomic strategy of checking if a file system object exists then act, which may lead to race conditions. {{{create-temporary-directory}}} is specially more fragile because it uses {{{directory-exists?}}} to check if a file system object exists. So, for example, if a regular file with the same name as the attempted one exists, create-temporary-directory will fail." Mario Domenech Goulart Milestone 6.0.0 1303 File ports should be nonblocking hard unknown 4.11.0 defect new 2016-06-30T19:14:45+02:00 2023-11-06T23:43:46+01:00 "As shown by this ugly little program, reading from regular files blocks. {{{ (use srfi-18) (thread-start! (lambda () (let lp () (print ""spinning"") (thread-sleep! 0.2) (lp)))) (thread-start! (lambda () (with-input-from-file ""/dev/tty"" (lambda () (let lp () (print (cons 'read-from-tty (read))) (lp)))))) (thread-sleep! 1000) }}}" sjamaan Milestone 6.0.0 1355 define strips hygiene from defined identifiers at top-level hard expander 4.12.0 defect new 2017-03-20T15:08:46+01:00 2023-11-06T23:52:22+01:00 "The following program with tests passes in chibi and scheme48, but fails in chicken. The expander handles inserted ""tmp"" identifiers and disambiguates correctly for readers and predicates, but it appears define-record-type treats all the setters the same. {{{ (cond-expand (chibi (import (scheme base) (chibi test))) (chicken (use test))) ;; other (,open srfi-9 srfi-23 in scheme48) ;; (define (read-string k in) ;; (let lp ((i k) (res '())) ;; (if (zero? i) ;; (list->string (reverse res)) ;; (lp (- i 1) (cons (read-char in) res))))) ;; (define (test-begin) #f) (define (test-end) #f) ;; (define (test e x) ;; (if (not (equal? e x)) ;; (error ""test failed"" e '!= x))) ;; utility primitive type consisting of a predicate, reader and writer (define-syntax define-binary-type (syntax-rules () ((define-binary-type name pred reader writer) (define-syntax name (syntax-rules (pred: read: write:) ((name pred: args) (pred args)) ((name read: args) (reader args)) ((name write: args) (writer args))))))) ;; a fixed-length string (define-binary-type fixed-string (lambda (args) (let ((len (car args))) (lambda (x) (and (string? x) (= (string-length x) len))))) (lambda (args) (let ((len (car args))) (lambda (in) (read-string len in)))) (lambda (args) (lambda (str out) (write-string str out)))) ;; wrapper around define-record-type to provide type checking and ;; (de)serialization (define-syntax defrec (syntax-rules () ;; all fields processed: expand record, reader and setters ((defrec name (make . make-fields) pred reader () (field (type . args) read-field get %set set) ...) (begin (define-record-type name (make . make-fields) pred (field get %set) ...) (define set (let ((field? (type pred: 'args))) (lambda (x val) (if (not (field? val)) (error ""invalid field"" '(type . args) val)) (%set x val)))) ... (define reader (let ((read-field (type read: 'args)) ...) (lambda (in) (let ((field (read-field in)) ...) (make . make-fields))))))) ;; step: insert read-field and %set bindings ((defrec name make pred reader ((field (type . args) get set) . rest) fields ...) (defrec name make pred reader rest fields ... (field (type . args) read-field get %set set))))) (define-syntax define-binary-record-type (syntax-rules () ((define-binary-record-type name make pred reader . fields) (defrec name make pred reader fields)))) ;; example (define-binary-record-type stuff (make-stuff foo bar) stuff? read-stuff (foo (fixed-string 2) stuff-foo stuff-foo-set!) (bar (fixed-string 3) stuff-bar stuff-bar-set!)) (test-begin) (let ((x (make-stuff ""ab"" ""bar""))) (test ""ab"" (stuff-foo x)) (test ""bar"" (stuff-bar x)) (stuff-foo-set! x ""xy"") ;; <----- this checks the right field type ;; but uses the sets the wrong %set (test ""xy"" (stuff-foo x)) (test ""bar"" (stuff-bar x))) (test-end) }}} " Alex Shinn Milestone 6.0.0 1464 ir-macro-expander's compare + define-syntax + modules don't mix expander 5.0.0 defect new 2018-05-22T08:18:26+02:00 2023-11-06T23:45:16+01:00 "This snippet is supposed to print ""it is x"", but doesn't if there's a macro called `x` in the module. {{{ module m (foo) ;; (import chicken scheme) (import (chicken syntax) scheme) ;; (define-syntax x (ir-macro-transformer (lambda _ `(begin)))) ;; (define (x) (begin)) ;; This is OK (define-syntax foo (ir-macro-transformer (lambda (e inj cmp) (apply (lambda (a) (if (cmp 'x a) ""it is x"" ""fail"")) (cdr e)))))) (import m) (print (foo x)) }}}" megane Milestone 6.0.0 1486 Reading and writing to a pipe set up by process can cause scheduler assertion errors hard core libraries 4.13.0 defect new 2018-07-15T13:52:24+02:00 2023-11-06T22:10:38+01:00 "As pointed out in [[https://stackoverflow.com/questions/51344859/connecting-to-interactive-tools-using-chicken-schemes-process|this StackOverflow post]]: {{{ (use posix) (let-values (((in out pid) (process ""bc""))) (begin (display ""2 + 2\n"" out) (print (read-line in)))) }}} Causes the following crash when evaluated: {{{ thread is registered for I/O on unknown file-descriptor: 4 (expected 3) [] ...more... <-- }}} This bug is not easy to trigger. I happened to get ""lucky"" and it crashed the first time, but after that I tried again several times and it just worked." sjamaan Milestone 6.0.0 1564 srfi-18: (mutex-unlock) Internal scheduler error extensions 5.0.0 defect new 2018-11-30T10:26:07+01:00 2023-11-06T22:11:53+01:00 "{{{ (import srfi-18) (import (chicken random)) (define m (make-mutex)) (define (rt) (/ (pseudo-random-integer 100) 1000)) (define (y) (let lp4 () (unless (= 0 (pseudo-random-integer 10)) (print (current-thread) "" "" ""yield"") (thread-yield!) (lp4)))) (thread-start! (make-thread (lambda () (let lp () (print (current-thread) "" "" ""lock"") (mutex-lock! m (rt)) (print (current-thread) "" "" ""unlock"") (mutex-unlock! m) (y) (lp))))) (let lp2 () (print (current-thread) "" "" ""lock"") (mutex-lock! m (rt)) (print (current-thread) "" "" ""sleep"") (thread-sleep! (rt)) (print (current-thread) "" "" ""unlock"") (mutex-unlock! m) (lp2)) }}} {{{ Warning (#): in thread: (mutex-unlock) Internal scheduler error: unknown thread state # ready Call history: t.scm:7: chicken.random#pseudo-random-integer t.scm:7: scheme#/ t.scm:19: srfi-18#mutex-lock! t.scm:20: srfi-18#current-thread t.scm:20: chicken.base#print t.scm:21: srfi-18#mutex-unlock! <-- # }}}" megane Milestone 6.0.0 1695 Exception handler bug when compiled with -O5? hard compiler 5.2.0 defect new 2020-04-10T18:30:26+02:00 2023-11-06T22:13:03+01:00 "Reported by NieDzejkob on IRC: Compile the following program with -O4 or lower and it terminates. With -O5 it gets stuck in an endless loop in the `with-exception-handler` lambda. {{{ (import (chicken condition)) (display (handle-exceptions exn (if (eq? exn 'two) 3 (abort exn)) (let [(orig (current-exception-handler))] (with-exception-handler (lambda (exn) (if (eq? exn 'one) (abort 'two) (orig exn))) (lambda () (abort 'one)))))) }}}" sjamaan Milestone 6.0.0 1803 Avoid getrusage system calls in GC medium core libraries 5.3.0 defect new 2022-05-31T13:16:34+02:00 2023-11-06T22:13:54+01:00 To calculate GC time statistics, `C_cpu_milliseconds` is currently used, which results in an getrusage(2) system call on many platforms. This should be avoided, preferrably by using a faster alternative like clock(3). felix winkelmann Milestone 6.0.0 1812 Modules leaking define-for-syntax variables and functions core libraries 5.3.0 defect new 2022-10-13T15:47:54+02:00 2023-11-10T16:18:45+01:00 "{{{ (module foo (foo-deluxe) (import scheme (chicken base) (chicken syntax)) (define-for-syntax (bar) (print ""hi!"")) (define-syntax foo-deluxe (syntax-rules () ((foo-deluxe) (bar)))) ) (import foo) (foo-deluxe) (bar) (set! bar #f) (foo-deluxe) }}} Is this intentional?" Idiomdrottning Milestone 6.0.0 1837 create-directory doesn't raise an error when the requested directory already exists easy core libraries 5.3.0 defect Mario Domenech Goulart assigned 2024-06-24T15:57:51+02:00 2024-06-25T11:36:51+02:00 "Changing that behavior in C5 would cause too much trouble, but we can change that in C6. {{{ $ touch foo $ csi -e '(begin (import (chicken file)) (create-directory ""foo""))' $ echo $? 0 }}} If the new behavior is implemented, please update https://wiki.call-cc.org/porting-c5-to-c6" Mario Domenech Goulart Milestone 6.0.0 1466 Try to get rid of evaluated imports in generated import libraries hard compiler 5.0.0 enhancement sjamaan new 2018-05-24T10:13:03+02:00 2023-11-06T23:53:26+01:00 "As we found out in #1457 and #1172, `merge-se` is quadratic in complexity. This means that every `import` we do is potentially really slow. Now that we only emit `eval` calls when generating import libraries of modules with syntax exports, it might be worthwhile to see if we can get rid of them altogether. As [http://lists.nongnu.org/archive/html/chicken-hackers/2018-05/msg00024.html|I proposed earlier], one way to do this would be to attach the syntax environment as an alist to each macro in the import library, instead of deriving it via imports (which will actually provide a much larger environment than necessary)." sjamaan Milestone 6.0.0 1504 Eggs compiled with -O2 medium extensions 5.0.0 enhancement new 2018-08-11T11:49:56+02:00 2023-11-06T23:36:29+01:00 "Some eggs are being compiled with -O2. Maybe their runtime performance can be improved by compiling them with -O3. Here's an initial list of the ones I observed being compiled with -O2: * base64 * defstruct * matchable * memory-mapped-files * posix-groups * sendfile * simple-md5 * simple-sha1 * srfi-18 * uri-common " Mario Domenech Goulart Milestone 6.0.0 1611 Restructure checks for inlinable intrinsics medium compiler enhancement sjamaan new 2019-04-24T17:16:19+02:00 2023-11-06T22:10:10+01:00 "Currently, we have a lot of intrinsics which are inlinable. Sometimes the safe inlineable versions are missing, sometimes the unsafe ones are missing. It also means there is a *lot* of handwritten C code with duplications. There would be less duplication if the compiler knew that the safe version is actually (roughly) an argument check followed by an unsafe call. For example: - `C_a_i_log` is just `C_check_real` followed by `C_flonum(a, log(f))` - `C_a_i_flonum_log` is `C_flonum(a, C_log(C_flonum_magnitude(x)))` (it should actually have a `_u_` in its name) Same is true for all the arithmetic operators in `(chicken flonum)`. At least for flonums, I think it's quite doable to have just the unsafe version and have some sort of ""rewrite"" in the compiler which will do the check followed by a call to the unsafe one. This can be done for a lot of other intrinsics too; think all the `char=`, `char<`, `char-upcase`, etc. This would give us a lot of faster implementations of safe operations ""for free"" by rewriting them to the unsafe ones, and if I'm not mistaken, we'd end up with a lot of calls to the ""check"" procedures, which ties in nicely with lfa2, which means all but the first can be eliminated. I'm thinking something like this: {{{ (define-intrinsic (chicken.flonum#fp+ x y) (check-flonum x) (check-flonum y) (##core#inline_allocate (""C_a_i_flonum_plus"" 4) x y) ;; to be renamed to C_a_u_i_flonum_plus? ) (define-intrinsic (scheme#integer->char n) (check-fixnum n 'integer->char) (##core#inline ""C_make_character"" (##core#inline ""C_unfix"" n)) ) }}} This could even replace the definitions in `library.scm`; then when we are compiling, we could process this code to generate an intrinsics database file, which the compiler can use like types.db. The above `define-intrinsic` would collapse to a simple `define` while generating `library.so`. When the compiler is compiling with an intrinsics database, it can replace all calls with the inlinable versions. Perhaps this is simply a more controlled way of doing `-emit-inline-file`? This isn't difficult but a lot of work and some care needs to be taken with regards to backwards C API/binary compatibility." sjamaan Milestone 6.0.0 1656 Refactor TCP unit to make ipv6 work from core medium core libraries 5.1.0 enhancement new 2019-11-26T09:37:56+01:00 2023-11-06T22:15:09+01:00 With the IPv4 address space having [https://www.ripe.net/publications/news/about-ripe-ncc-and-ripe/the-ripe-ncc-has-run-out-of-ipv4-addresses officially run out], I think we should start planning to make the [https://wiki.call-cc.org/eggref/5/tcp6 tcp6] egg part of core. sjamaan Milestone 6.0.0 1821 Have big-chicken-like feature in the core core libraries 5.3.0 enhancement new 2023-06-04T00:11:46+02:00 2023-11-07T18:18:52+01:00 "This is about having the ability to conveniently import all core modules, like what the big-chicken egg provides. " Mario Domenech Goulart Milestone 6.0.0 1828 Keep R5RS syntax-rules? easy expander 5.3.0 task new 2023-11-06T23:22:26+01:00 2023-11-06T23:22:26+01:00 For C6, syntax-rules has been changed to comply with R7RS. We could keep the old semantics as a separate export, since it can parameterized via an additional argument to make-transformer (as suggested by sjamaan). Since the default environmnet is still R5RS, providing R5RS syntax-rules by default seems to me to risk confusion, so I'm inclined to default to R7RS. felix winkelmann Milestone 6.0.0 1436 Don't disable inline file generation with -O5 (disable-interrupts) compiler 5.0.0 change request new 2018-01-02T11:27:37+01:00 2023-11-06T23:30:46+01:00 "(I'm assuming here that disable-interrupts is enough to implement critical sections in concurrency related code.) The disable-interrupts declaration correctly disables the global inline file generation. My belief is that inlining is more of an optimization than disabling the interrupt checks. So removing the disable-interrupts from O5 would be one choice. But wouldn't it be OK to always disable the checks in inlined code? The functions made available for global inlining are controlled explicitly by the user by the use of the inline declaration. So if you need interrupt checks (a function doing some heavy computation perhaps) just don't make it available for global inlining. " megane Milestone 6.0.0 125 cross-chicken should prefer loading uncompiled import libraries medium core tools 4.2.x defect assigned 2009-11-21T12:38:13+01:00 2023-11-06T22:15:30+01:00 Unless in `-host` mode, `chicken(1)` should prefer loading import libraries from the current directory in source form (`.scm`), or it will possibly load import libs compiled for the target architecture. felix winkelmann Milestone 6.0.0 1348 improvement to manual pages easy documentation 4.12.0 defect new 2017-02-27T08:28:28+01:00 2023-11-06T23:37:42+01:00 It would be greatly helpful if each manual on the website had a link to the level right above it. Right now, to know which chapter you are in, you have to scroll to the bottom of each page and click the link to go to the previous section until you get to the beginning. vigilancetech Milestone 6.0.0 1370 Build-dependencies between import libraries and core units don't seem to be completely right medium build system 5.0.0 defect new 2017-05-09T16:19:05+02:00 2023-11-06T22:15:49+01:00 "During some messing around with my working copy, I deleted all generated `*.import.c` and `*.import.scm` files. Running `make` again resulted in build errors, reporting missing import libraries (more specifically: unresolved imports.) To fix this, the unit holding the module that produces the missing import libraries has to be recompiled from the `.scm` spurce. It seems that the dependencies between Scheme files and import libraries are not fully specified in `rules.make`. " felix winkelmann Milestone 6.0.0 1619 Specialization ignores fixnum mode, leading to suboptimal code compiler 5.0.0 defect new 2019-05-29T16:33:06+02:00 2023-11-06T22:16:39+01:00 "As we found with #1604, when running with `-fixnum-arithmetic`, specialization happens before optimization. Because the scrutinizer doesn't know about fixnum mode, this is still something of an issue, because we'd specialize to code that supports bignums, which is allocating. This is problematic in two ways: one is that the generated code that doesn't get replaced will not be able to deal with bignums (resulting in extremely strange behaviour on overflow or when the code results in ratnums), and it's less than optimal because fixnum ops aren't allocating, so they should be faster still. Perhaps we could stop the scrutinizer from doing replacements that deal with numerics in fixnum mode?" sjamaan Milestone 6.0.0 1823 Ship chicken.version module unknown 5.3.0 defect Mario Domenech Goulart accepted 2023-06-04T01:01:36+02:00 2025-10-03T18:26:59+02:00 "CHICKEN 4's setup-api provides {{{version>=?}}}, whose code is in CHICKEN 5 but not exposed to users. That leads to a lot of copy&paste of that procedure into the source of eggs. This code could be an egg, but then it is needed by CHICKEN anyway, and has been stable enough not to require any modification along the entire time CHICKEN 5 has been available, so unlikely to require a lot of maintenance in the future. The following eggs copy&paste the definition from CHICKEN 4: * [https://github.com/mario-goulart/test-new-egg/blob/635eb6abe29890b23c906b50c3ecd02f4146b7ec/test-new-egg.scm#L26-L44 test-new-egg] * [https://github.com/mario-goulart/henrietta-cache-git/blob/344ed10b82c03e539d45bc953cd1bac7e43c03f6/version.scm#L6-L24 henrietta-cache-git] * [http://bugs.call-cc.org/browser/project/release/5/henrietta/trunk/henrietta.scm#L66 henrietta] * [https://code.call-cc.org/cgi-bin/gitweb.cgi?p=eggs-5-latest.git;a=blob;f=git/0.2.0/custom-build.scm;h=ac99a6441d286b2f59a36b36cbd5985f86756a2f;hb=HEAD#l18 git] * [https://github.com/mario-goulart/egg-tarballs/blob/fa18fc9d0b4ceccd08c7b4a03213aa670edf9f3c/egg-tarballs-index.scm#L24-L42 egg-tarballs] * [https://git.sr.ht/~dieggsy/chicken-update/tree/f473ed256bde58b5bfd030db4ec3cc2eaf0aac52/item/chicken-update.scm#L86-103 chicken-update] Additionally, the following CHICKEN code not packaged as eggs duplicate {{{version>=?}}}: * [https://github.com/mario-goulart/egg-layer/blob/628367314e87c0fc99d269db740eba5d7c371dde/egg-layer.scm#L71-L89 egg-layer] " Mario Domenech Goulart Milestone 6.0.0 1856 srfi-69 egg hash-table-remove! operation doesn't operate over the entire table trivial extensions 5.4.0 defect new 2025-08-20T00:34:01+02:00 2025-08-20T00:34:01+02:00 hash-table-remove! when a match will remove the bucket, but ignore the rest of the chain. Kon Lovett Milestone 6.0.0 1857 srfi-69 egg hash-table-update![/default] always attempts table resize easy extensions 5.4.0 defect new 2025-08-20T02:55:05+02:00 2025-08-20T02:55:05+02:00 assumes the entry to be updated doesn't exist so attempts to grow the table by 1 Kon Lovett Milestone 6.0.0 1858 srfi-69 egg uses list for table of primes & vector better trivial extensions 5.4.0 defect new 2025-08-20T03:02:15+02:00 2025-08-20T03:02:15+02:00 vector is better for size & access, especially since during a growth spurt the table is frequently referenced. the table is never accessed except thru one function so any representation change is isolated. Kon Lovett Milestone 6.0.0 938 Feature Request: Load type files for units medium compiler 4.8.x enhancement felix winkelmann assigned 2012-10-21T14:42:52+02:00 2023-11-06T22:18:08+01:00 "Currently type files for units not installed in the repository are not loaded. I've attached a patch that does something like this, but it's just a quick hack that works for my needs. You should probably use `##sys#resolve-include-filename` to find the types file, like it's used to find the inline files. " megane Milestone 6.0.0 1286 define-constant and define-inline should shadow imported identifiers hard compiler 4.10.x enhancement new 2016-05-22T07:28:29+02:00 2023-11-06T22:18:34+01:00 "The following should print {{{yes}}} rather than the procedure {{{exact}}}: {{{ (use (only data-structures exact)) (define-constant exact 'yes) (print exact) ; => #exact x1007)> instead of 'yes }}} The same applies for {{{define-inline}}}: {{{ (use (only numbers exact)) (define-inline (exact n) 'yes) (print (exact 1.0)) ; => 1 instead of 'yes }}} This happens because variables are module-resolved (via {{{lookup}}} against the current environment) in {{{resolve-variable}}} before they're keyed into the constant and inline tables. I think the raw identifier should be used when looking up inlined and constant idenfifiers instead ({{{x0}}} rather than {{{x}}}), but haven't looked too deeply at this issue yet." evhan Milestone 6.0.0 1605 Display compiler warnings for macros with missing imports hard compiler 5.0.0 enhancement new 2019-04-10T12:07:55+02:00 2023-11-06T23:54:29+01:00 The compiler already displays warnings for procedures with missing imports which is super useful, especially when compiling code that's not wrapped in a module (wrapping it in a module will turn the warnings into errors). However there is no equivalent for macros with missing imports and neither does wrapping them into a module produce a warning or error. This leads to less than obvious errors if you later import such a macro into other code and compile it, you'll then get an error for the import you forgot to add for the macro's module. See #1603 for an example. Vasilij Schneidermann Milestone 6.0.0 1705 Module names should be unique medium compiler 5.2.0 enhancement new 2020-07-21T09:34:10+02:00 2023-11-06T23:07:44+01:00 "When you compile a program which uses a module and defines a module of the same name, the compiler should probably give a warning or error. A practical (simplified) example of this which alicemaz ran into: http://paste.call-cc.org/paste?id=59272bcd1450867b3db94a2b7a6148fbd39d4f44 I think we can't really do this in the interpreter because re-evaluating an existing module should be possible on the REPL (think incremental development from Emacs etc)." sjamaan Milestone 6.0.0 1746 Improve chicken-install cyclic dependencies error core tools 5.2.0 enhancement new 2021-04-12T23:18:40+02:00 2023-11-06T23:35:17+01:00 "Currently with cyclic dependencies in a complex egg file, {{{chicken-install}}} throws an error like e.g.: {{{ Error: cyclic dependencies: ((math.polynomial.chebyshev math.flonum.functions math.racket-shim math.base) (math.functions.stirling-error math.base math.flonum.functions math.flonum.polyfun math.polynomial.chebyshev math.racket-shim) (math.vector.base) (math.number-theory.tangent-number) (math.number-theory.quadratic-residues math.number-theory.divisibility math.number-theory.modular-arithmetic math.number-theory.base math.... }}} This is quite hard to make sense of (and not to mention incomplete)" Diego Milestone 6.0.0 1824 Change locatives representation medium core libraries 5.3.0 enhancement sjamaan assigned 2023-06-27T09:24:25+02:00 2023-06-27T09:24:25+02:00 "As pointed out [https://lists.nongnu.org/archive/html/chicken-hackers/2023-06/msg00029.html here], we can change the representation of locatives such that they take up less space, and we only have to traverse the live '''weak''' locatives for update during GC. The idea is to drop the extra ""object"" slot, and store it in the pointer slot instead. When the locative is weak, set {{{C_SPECIALBLOCK_BIT}}}. When it is strong, do not set it (so that the object is referenced normally). For weak locatives, we traverse the live ones using the ""chained"" approach and update the object. We don't need to do the pointer recalculation on every GC. Instead, we calculate the pointer on-the-fly during {{{locative_ref}}}. NOTE: Perhaps we also want to expose a version of locative-ref that doesn't error out when the object has been collected, but returns {{{bwp-object}}} instead? This means handling of such locatives is a bit simpler and doesn't require exception handling." sjamaan Milestone 6.0.0 1354 Clean up the expander a bit hard expander 5.0.0 task sjamaan new 2017-03-18T22:40:20+01:00 2023-11-06T23:08:03+01:00 "Currently the expander is full of hooks and hacks, mostly for maintaining the line number database and various syntactic environments. It would be nice if we can clean things up a bit. A few ideas: * {{{##sys#syntax-context}}} and {{{##sys#syntax-error/context}}} are weird and can maybe be cleaned up if we can make {{{##sys#check-syntax}}} work better with line number info. * {{{##sys#syntax-error-culprit}}} seems to be unnecessary, because while expanding {{{##sys#check-syntax}}} should have access to the input form already? * {{{##sys#line-number-database}}} is defined in expander, but only really used in {{{core.scm}}} (except for getting the line number) * The new {{{expansion-result-hook}}} is yet another hook, and can perhaps be combined with the previous two. * The special-cased hack in {{{##sys#canonicalize-body}}} that prevents expansion of {{{import}}} and hands back control to the compiler after expansion of {{{##core#module}}}) (in a currently pending patch) needs to die. No clue yet how to do this..." sjamaan Milestone someday 1112 Figure out a way to safely release objects hard core libraries 4.8.x defect new 2014-04-16T17:11:17+02:00 2017-08-25T14:05:30+02:00 "The {{release-object}} procedure will just call {{free()}} on the object's pointer. This means that any Scheme variable still referring to the object will now point to free memory, which may get overwritten with arbitrary values. The GC will not touch memory outside the heap, but it's possible that the freed object's memory gets ""pulled in"" when resizing the heap upon major GC, which means the GC's tests for whether it's in the heap space will succeed, causing it to treat the object as ""valid"", even if something else has temporarily written bogus stuff to that memory space in between. We need a way for the GC to kill all references to the released object, but this is currently impossible: these references are plain pointers, and there is no way to ensure a GC won't find any references to these objects. The trick used by object-become is to collect these objects on a special stack, and force a major GC, which then treats these objects specially. This is probably too much of performance impact though, because object-evict generally tends to be used for performance reasons. Perhaps some form of ""weak evicted object pointers"" can be introduced?" sjamaan Milestone someday 1293 eq?-hash tables calculcate different value for objects after they're mutated and lose locatives after GC hard core libraries 4.10.x defect new 2016-05-27T08:43:10+02:00 2017-10-28T21:49:45+02:00 "If you mutate a pair, it's still `eq?` to itself, but the `hash-by-identity` / `eq?-hash` doesn't calculate the same hash anymore, which means the hash table cannot find the key, unless the randomization factor happens to be such that it finds the same bucket, of course, but that's highly unlikely for simple cases like the one below: {{{ #;1> (use srfi-69) ; loading /home/sjamaan/chickens/4.11.0/lib/chicken/8/srfi-69.import.so ... ; loading library srfi-69 ... #;2> (define x (cons 1 2)) #;3> (hash-by-identity x) 380849329 #;4> (set-car! x 3) #;5> (hash-by-identity x) 380718257 #;6> (= #3 #5) #f #;7> }}} Originally [[http://paste.call-cc.org/paste?id=982e393227789ab1a6c471ada66e729eb28f83b8|pointed out by John Croisant]]: {{{#!scm ;; Even though the filler slots are not used, if you remove any of ;; them, the hash of the instance will change after GC, causing the ;; hash table lookup to fail. (use srfi-69 lolevel) (define table (make-hash-table test: eq? hash: eq?-hash)) (define-record-type box (make-box contents) box? ;; Remove or comment out any of the next three lines: (filler1 box-filler1) (filler2 box-filler2) (filler3 box-filler3) (contents box-contents)) (define my-box (make-box (make-locative ""foo""))) (hash-table-set! table my-box #t) (printf ""before gc, hash table contains ~S? ~S~N"" my-box (hash-table-exists? table my-box)) (gc) (printf ""after gc, hash table contains ~S? ~S~N"" my-box (hash-table-exists? table my-box)) (printf ""hash table as alist: ~S~N"" (hash-table->alist table)) }}}" sjamaan Milestone someday 488 SLIME completion does not complete up to first non-unique point in string hard extensions 4.6.x defect Nick Gasson new 2011-01-30T13:59:06+01:00 2023-11-06T21:57:47+01:00 "When I type in {{{(with-i)}}}, I'd expect SLIME to complete {{{(with-input-from-)}}} and then offer me the options with-input-from-file, with-input-from-port, with-input-from-string and with-input-from-pipe. Right now it only offers options for completion but does not fill out the parts that can already be filled out, requiring me to switch to the completion buffer and choose an option or type it out myself until there's only one option left; this it does complete. According to Christian, this may be a problem with the way the swank server replies to the completions request." sjamaan Milestone someday 731 couchdb-view-server: segfault in tests extensions 4.7.x defect Moritz Heidkamp new 2011-11-18T18:26:22+01:00 2023-11-06T21:58:32+01:00 See http://tests.call-cc.org/master/linux/x86/2011/11/18/salmonella-report/tests/couchdb-view-server.html Mario Domenech Goulart Milestone someday 779 hyde: support trailing slashes on directories trivial extensions 4.7.x defect new 2012-01-19T03:31:25+01:00 2023-11-06T21:59:31+01:00 "hyde serve (hyde 0.15) fails to serve index content when the request url has a trailing slash on the end of the directory. For example, make a hyde project with a file like the following in the src/ dir: about/index.wiki Run hyde serve and try the following urls: http://localhost:8080/about/index.html → OK http://localhost:8080/about → OK http://localhost:8080/about/ → FAIL The problem is in the procedure 'page-by-path', which is called by 'serve' to find the page object for the given uri path. In the case where the directory has a trailing slash, the path object (a list that was returned by 'uri-path') has as its last term an empty string, and page-by-path fails to deal with this. A patch showing one way to fix this is attached. " John Foerch Milestone someday 813 Extension tiger-hash fails on big-endian machines hard extensions 4.7.x defect Kon Lovett accepted 2012-04-07T19:43:00+02:00 2023-11-06T21:59:48+01:00 Doesn't pass self-test. Kon Lovett Milestone someday 945 html-parser supports only a few entities easy extensions 4.8.x defect Alex Shinn assigned 2012-10-31T00:59:05+01:00 2017-08-25T14:10:09+02:00 The html-parser egg supports only a few entities ( , < and friends) by default. If one wants to parse html that contains other entities than these, one cannot use the simple html->sxml, but must use make-html-parser just to add entities. I propose that for purposes of html->sxml, html-parser should not validate entities at all, but simply create sxml entity forms for whatever entities it encounters. John Foerch Milestone someday 960 Need API to box and unbox FILE* pointers as Scheme ports hard core libraries 4.8.x defect new 2012-12-18T12:47:48+01:00 2016-08-25T21:07:47+02:00 "The C API should be able to box a FILE* as a Scheme input port or output port. There should also be a way to extract a FILE* from a Scheme port, or NULL if the port is not file-based. Ideally this should be done by extending the list of foreign type specifiers." johnwcowan Milestone someday 980 make input-parse more neighbourly by signalling an exception in assert-curr-char instead of calling error easy unknown 4.8.x defect new 2013-02-08T12:53:38+01:00 2016-08-25T21:10:17+02:00 The attached patch signals an exception of type (input-parse parser-error exn) instead of calling error directly. Christian Kellermann Milestone someday 991 spiffy-request-vars' test fails sometimes insane unknown 4.8.x defect reopened 2013-03-01T23:11:52+01:00 2016-08-25T21:10:56+02:00 "A failure that sometimes happens on spiffy-request-vars' tests happened yesterday while I was testing a stability release tarball provided by Jim. Here are the reports: http://tests.call-cc.org/master/linux/x86/2012/06/19/salmonella-report/test/spiffy-request-vars.html http://tests.call-cc.org/master/linux/x86/2012/09/06/salmonella-report/test/spiffy-request-vars.html http://tests.call-cc.org/master/linux/x86/2012/09/10/salmonella-report/test/spiffy-request-vars.html The report for the test I was running yesterday is here: http://parenteses.org/mario/salmonella/4.8.0.3/linux/x86-64/2013/02/28/salmonella-report/test/spiffy-request-vars.html" Mario Domenech Goulart Milestone someday 1005 mmap test doesn't correctly catch error situations medium extensions defect reopened 2013-04-11T15:12:17+02:00 2016-08-25T20:17:48+02:00 "I have just came across this issue when dealing with a failed mmap FFI call. In this case mmap (The C function) will return a MAP_FAILED which is defined as (void*)-1 on linux and OpenBSD. The mmap code checks for this by doing a (eq? -1 addr2). This is fine for 32 bit systems but not for 64 bit systems. {{{ ;; on 32 bit #;2> (address->pointer -1) # ;; on 64 bit #;2> (address->pointer -1) # #;3> (pointer->address #2) 1.84467440737096e+19 }}} Note how the pointer printing code also gets it wrong. **this has been fixed in CHICKEN 5: it will now correctly print the pointer's address, interpreted as an unsigned number.** A workaround would be to explicitly check for the -1 pointer representation: {{{ (pointer=? (address->pointer -1) addr2) }}} This works but looks icky. I am not sure how to handle this right." Christian Kellermann Milestone someday 1043 byte-blob-stream-find fails to find matches at the end of a stream extensions 4.8.x defect Ivan Raikov assigned 2013-08-08T03:20:33+02:00 2014-08-05T03:24:33+02:00 "Analogous to issue #1037, byte-blob-stream-find cannot find matches at the end of a stream: (byte-blob-stream-find (list->byte-blob '(1 31)) (list->byte-blob-stream '(0 1 31 3))) " Ivan Raikov Milestone someday 1095 AD egg compilation fails due to unsupported gcc flag on OpenBSD-current trivial unknown 4.8.x defect new 2014-02-04T13:16:01+01:00 2016-08-25T21:24:10+02:00 "The AD egg fails to compile, due to a hard coded ""-Wno-unused-but-set-variable"" which is not available on OpenBSD-current's gcc (4.2.1). It builds fine without that flag." Christian Kellermann Milestone someday 1128 stty termios struct, add feature test for speed attributes hard extensions 4.9.x defect Alex Shinn new 2014-06-03T17:03:24+02:00 2016-08-25T21:34:19+02:00 "the stty egg fails to compile on android because of missing fields. in android bionic termios is defined like this: {{{ struct termio { unsigned short c_iflag; /* input mode flags */ unsigned short c_oflag; /* output mode flags */ unsigned short c_cflag; /* control mode flags */ unsigned short c_lflag; /* local mode flags */ unsigned char c_line; /* line discipline */ unsigned char c_cc[NCC]; /* control characters */ }; }}} while on linux (x86) it's: {{{ struct termios { tcflag_t c_iflag; /* input mode flags */ tcflag_t c_oflag; /* output mode flags */ tcflag_t c_cflag; /* control mode flags */ tcflag_t c_lflag; /* local mode flags */ cc_t c_line; /* line discipline */ cc_t c_cc[NCCS]; /* control characters */ speed_t c_ispeed; /* input speed */ speed_t c_ospeed; /* output speed */ #define _HAVE_STRUCT_TERMIOS_C_ISPEED 1 #define _HAVE_STRUCT_TERMIOS_C_OSPEED 1 }; }}} The attached patch adds feature tests for c_ispeed and c_ospeed. Have tested on my workstation (arch linux x86) and embedded android (arm). My first time using feature-test so if you have any feedback please let me know. If it looks okay, please merge. Regards, Peder Refsnes" hanDerPeder Milestone someday 1131 Kill ##sys#alias-global-hook with fire insane unknown 4.9.x defect new 2014-06-09T20:57:36+02:00 2023-11-06T21:53:43+01:00 The whole concept is broken. Just by studying the code it's easy to come up with bug after bug. sjamaan Milestone someday 1141 hyde: execute permissions on hyde executable medium extensions 4.9.x defect Moritz Heidkamp assigned 2014-07-15T01:37:54+02:00 2017-08-25T14:14:06+02:00 "I installed hyde by doing: $ chicken-install -s hyde Chicken-install reported that the hyde executable was installed like this: sudo cp -r 'hyde' '/usr/bin/hyde' sudo chmod a+r '/usr/bin/hyde' This file then did not have execute permission for my user account: $ ls -la /usr/bin/hyde -rwxr-xr-- 1 root root 82326 Jul 14 19:24 /usr/bin/hyde This may or may not be related to the fact that my account's umask is 027 instead of the more commonly seen 022. $ umask 0027 I am using chicken 4.9.0 from Debian jessie (current testing)." John Foerch Milestone someday 1145 Issue with matchable and the module system hard extensions 4.9.x defect sjamaan assigned 2014-08-05T16:15:36+02:00 2016-08-25T21:36:06+02:00 Here are some cases that illustrate the problem: http://paste.call-cc.org/paste?id=4d676dd966a3e3c5b5a3d216014d07a7fa746c6f Mario Domenech Goulart Milestone someday 1165 redis relies on GNU make, which is not the default make on BSDs medium extensions 4.9.x defect 0xab new 2014-10-28T14:20:22+01:00 2016-08-25T21:36:49+02:00 See http://salmonella-freebsd-x86-64.call-cc.org/master/clang/freebsd/x86-64/2014/10/28/salmonella-report/install/redis.html for an error example. Mario Domenech Goulart Milestone someday 1168 make install: umask problem with modules.db medium build system 4.9.x defect new 2014-11-09T03:40:26+01:00 2016-08-25T21:37:00+02:00 With a UMASK of 0027, 'sudo make PLATFORM=linux install' installed the file $PREFIX/lib/chicken/7/modules.db with permissions rw-r-----, instead of the expected rw-r--r--, preventing csc from working for my user account. John Foerch Milestone someday 1180 parley seems to insert the same sexpression to the reader again when getting input while sleeping extensions 4.9.x defect Christian Kellermann new 2015-02-17T12:56:53+01:00 2017-10-14T16:59:19+02:00 See http://paste.call-cc.org/paste?id=1f10cc1f4285e1b1a4de027bc05a3ae30652121b Christian Kellermann Milestone someday 1204 chicken-install permissions with executables medium unknown 4.9.x defect new 2015-07-26T18:02:02+02:00 2016-08-25T21:42:51+02:00 "I believe I am seeing another case of issue #773, incorrect permissions on chicken-install'd files when UMASK is 0027. I installed the missbehave egg with chicken-install, which includes a binary called 'behave'. I see the following in chicken-install's output: sudo cp -r 'behave' '/usr/local/bin/behave' sudo chmod a+r '/usr/local/bin/behave' This file should have a+rx since it is an executable. CHICKEN 4.9.0.1 setup-helper: 1.5.4 " John Foerch Milestone someday 1217 Unit Posix time/date behaviors differ across operating systems insane core libraries 4.9.x defect new 2015-08-28T15:17:10+02:00 2016-08-25T21:44:33+02:00 "Procedures seconds->local-time and string->time produce different results on (Arch) Linux vs OS X. Windows behavior is unknown. This likely reflects a difference in the behavior of strptime on the two systems---see: [http://stackoverflow.com/questions/31984002/why-does-strptime-behave-differently-on-osx-and-on-linux] Giving a fully specified and unambiguous time string in Chicken does not fix this problem. For example, {{{ (string->time ""2015-08-25 12:00:00 EDT"" ""%Y-%m-%d %H:%M:%S %Z"") }}} returns different vectors on OS X vs Linux." nemo Milestone someday 1235 Issue with parameters medium core libraries 4.10.x defect new 2015-12-15T11:36:17+01:00 2018-04-27T20:59:51+02:00 "Reported by Joo ChurlSoo via e-mail: {{{ CHICKEN (c) 2008-2015, The CHICKEN Team (c) 2000-2007, Felix L. Winkelmann Version 4.10.0 (rev b259631) windows-mingw32-x86 [ manyargs dload ptables ] compiled 2015-08-04 on yves.more-magic.net (Linux) #;1> (define a (make-parameter 1)) #;2> (a 10 23) 10 #;3> (a) 10 #;4> (define b (make-parameter 2 number->string)) #;5> (b 20 23) 20 #;6> (b) 20 #;7> (b 20) ""20"" #;8> (b) ""20"" }}}" Mario Domenech Goulart Milestone someday 1238 mw (tests): openssl egg is needed for HTTPS trivial extensions 4.10.x defect Tony Sidaway new 2015-12-31T12:21:32+01:00 2016-08-25T20:23:57+02:00 See http://salmonella-linux-x86-64.call-cc.org/master-debugbuild/gcc/linux/x86-64/2015/12/30/salmonella-report/test/mw.html Mario Domenech Goulart Milestone someday 1241 pty: test failure extensions 4.10.x defect Alex Shinn new 2015-12-31T12:26:51+01:00 2015-12-31T12:26:51+01:00 See http://salmonella-linux-x86-64.call-cc.org/master-debugbuild/gcc/linux/x86-64/2015/12/30/salmonella-report/test/pty.html Mario Domenech Goulart Milestone someday 1247 twilio's license is BSD, but it depends on a GPL-3 egg (s) extensions 4.10.x defect Peter Danenberg new 2015-12-31T12:53:30+01:00 2015-12-31T12:53:30+01:00 "See http://salmonella-linux-x86-64.call-cc.org/master-debugbuild/gcc/linux/x86-64/2015/12/30/salmonella-report/dep-graphs/twilio.html [https://wiki.call-cc.org/egg/s s]'s license is GPL-3." Mario Domenech Goulart Milestone someday 1265 AD: installation failure easy extensions 4.10.x defect 0xab new 2016-02-25T22:25:24+01:00 2016-08-25T20:30:35+02:00 "See http://salmonella-linux-x86-64.call-cc.org/master-debugbuild/gcc/linux/x86-64/2016/02/25/salmonella-report/install/AD.html Note that the installation failure for AD breaks four eggs: http://salmonella-linux-x86-64.call-cc.org/master-debugbuild/gcc/linux/x86-64/2016/02/25/salmonella-report/rev-dep-graphs/AD.html" Mario Domenech Goulart Milestone someday 1291 lowdown: markdown->sxml produces unnormalised SXML medium extensions 4.10.x defect Moritz Heidkamp assigned 2016-05-26T18:51:39+02:00 2016-08-25T20:32:44+02:00 "{{{ % csi CHICKEN (c) 2008-2015, The CHICKEN Team (c) 2000-2007, Felix L. Winkelmann Version 4.10.0 (rev b259631) macosx-unix-clang-x86-64 [ 64bit manyargs dload ptables ] compiled 2015-08-04 on yves.more-magic.net (Linux) #;1> (use lowdown sxpath) ; loading /Data/tools/chicken-4.10.0/lib/chicken/7/lowdown.import.so ... [...] #;2> (define testdoc ""Title\n=====\n\n * [link](uri)\n"") #;3> (define sx-test (with-input-from-string testdoc (lambda () (markdown->sxml (current-input-port))))) #;4> sx-test ((h1 (""Title"")) (ul (li ((a (@ (href ""uri"")) ""link""))))) #;5> ((sxpath '(// a)) sx-test) () #;6> (define (tidy-sxml sx) (cond ((not (list? sx)) sx) ((and (= (length sx) 1) (list? (car sx))) ;; turn ((node ...)) into (node ...) (tidy-sxml (car sx))) (else (map tidy-sxml sx)))) #;7> ((sxpath '(// a)) (tidy-sxml sx-test)) ((a (@ (href ""uri"")) ""link"")) #;8> }}} I think the parsed `sx-test` is not invalid, but it does appear to be sufficiently unnormalised that it confuses `sxpath`. See http://okmij.org/ftp/Scheme/SXML.html#Normalized%20SXML The function `tidy-sxml` works in this case, but is probably too simple-minded to work in general." Norman Gray Milestone someday 1305 chmod of modules.db on new installation easy build system 4.11.0 defect new 2016-07-17T00:44:49+02:00 2016-08-25T20:45:34+02:00 "I use a umask of 0027 on my debian system, so files created with my user account have no 'other' permissions by default. I built and installed CHICKEN 4.11.0 as follows: $ make PLATFORM=linux PREFIX=/usr/local $ sudo make PLATFORM=linux PREFIX=/usr/local install The file /usr/local/lib/chicken/8/modules.db was installed with the following permissions: -rw-r----- 1 root staff 50046 Jul 16 16:18 /usr/local/lib/chicken/8/modules.db This prevents chicken-install from working properly." John Foerch Milestone someday 1321 async-io tests hang unknown 4.11.0 defect new 2016-08-27T00:08:12+02:00 2017-11-25T19:11:24+01:00 "See http://lists.nongnu.org/archive/html/chicken-users/2016-07/msg00040.html for more context. Basically, async-io tests hang sometimes. To reproduce the issue, just run the tests repeatedly until they hang. It was making salmonella hang on the linux machines. I installed a hack to kill the hanging test process after some time. Although the hack is in place and working, salmonella runs take longer when the test process hangs, as the hack polls to detect hanging processes." Mario Domenech Goulart Milestone someday 1323 Teach compiler that fold and map with clean procedure is itself clean hard compiler 4.11.0 defect new 2016-08-29T12:30:29+02:00 2016-08-29T12:30:29+02:00 "It looks like CHICKEN currently doesn't understand that `fold` and `map` are `#:clean` if the argument procedure is `#:clean`. This prevents it from recognising that something like `(fold + 0 (iota 10000))` is foldable. We could hardcode these procedures, but I'd prefer not to, so this might require some kind of new annotation in `types.db`. This idea was inspired by http://pramode.in/2016/08/25/rust-add-billion-numbers/ - if Rust can do it, so can we!" sjamaan Milestone someday 1325 qt-light egg: timers may put chicken into an inconsistent state extensions 4.11.0 defect new 2016-09-02T17:36:22+02:00 2017-08-25T14:22:59+02:00 "I have found that a Qt::Timer may be used to make callbacks into chicken faster than the callbacks themselves can finish. This puts chicken into an inconsistent state and causes a crash. Here is a test case - callback rate and count may need to be adjusted for faster machines. {{{ (import chicken scheme) (use qt-light) (qt:init) (let* ((thread-rate 0.025) (timer (qt:timer thread-rate))) (define (timer-callback) (let loop ((i 0)) (if (< i 1e6) (loop (+ i 1)) #t))) (qt:connect timer ""timeout()"" timer-callback) (qt:start timer) (qt:run)) }}} {{{ $ ./minimal.scm Warning: excluded identifier doesn't exist in module foreign: foreign-declare Warning: excluded identifier doesn't exist in module foreign: foreign-declare csi: runtime.c:2797: C_save_and_reclaim: Assertion `av > C_temporary_stack_bottom || av < C_temporary_stack_limit' failed. Aborted }}} Since Qt must be used for the main event loop in a Qt program, making timers safe to use would be fairly important for the qt-light egg. Without knowing much about the internals, I think that maybe a semaphore could be used to block a callback if a previous callback has not finished. When this occurs, it would be nice if the api also exposed a way for a chicken program to know that a callback was dropped, so that it could take an appropriate action." John Foerch Milestone someday 1335 Building on OS X: egg uses 1 deprecated function and 1 missing function easy unknown 4.11.0 defect Thomas Chust assigned 2016-10-31T11:49:09+01:00 2021-08-13T14:33:24+02:00 "Building the sqlite3 egg on OS X (10.11.6) produces a warning and an error -- see below for full trace. Specifically: {{{ 'sqlite3_enable_shared_cache' has been explicitly marked deprecated here }}} The SQLite docs note https://www.sqlite.org/c3ref/enable_shared_cache.html that ‘This method is disabled on MacOS X 10.7 and iOS version 5.0 and will always return SQLITE_MISUSE’ Also, the function sqlite3_enable_load_extension appears not to be available on OS X. The SQLite docs https://www.sqlite.org/c3ref/enable_load_extension.html seem to recommend a different route to disabling this, on security grounds, but I confess I don't understand the implications of that. This function is still mentioned in the system sqlite3.h header, but it appears to be absent from (been deliberately removed from?) the system sqlite3 library. There appear to be alternative documented routes to both of the intended functionalities, so I've tentatively marked this as 'easy' difficulty. Building with: {{{ chicken-install -n -D disable-shared-cache,disable-load-extension }}} appears to work OK, though I haven't tested fully. Original build output... {{{ % chicken-install -version 4.11.0 % chicken-install -n retrieving ... checking platform for `sqlite3' ... checking dependencies for `sqlite3' ... install order: (""sqlite3"") installing sqlite3: ... changing current directory to . '/Data/tools/chicken-4.11.0/bin/csi' -bnq -setup-mode -e ""(require-library setup-api)"" -e ""(import setup-api)"" -e ""(setup-error-handling)"" -e ""(extension-name-and-version '(\""sqlite3\"" \""\""))"" -e ""(keep-intermediates #t)"" -e ""(setup-install-mode #f)"" 'sqlite3.setup' '/Data/tools/chicken-4.11.0/bin/csc' -feature compiling-extension -setup-mode -k -O2 -d1 -s sqlite3.scm -lsqlite3 -j sqlite3 Warning: imported identifier doesn't exist in module chicken: dynamic-wind sqlite3.c:310:23: warning: 'sqlite3_enable_shared_cache' is deprecated: first deprecated in macOS 10.7 [-Wdeprecated-declarations] C_r=C_int_to_num(&C_a,sqlite3_enable_shared_cache(t0)); ^ /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.12.sdk/usr/include/sqlite3.h:5424:31: note: 'sqlite3_enable_shared_cache' has been explicitly marked deprecated here SQLITE_API int SQLITE_STDCALL sqlite3_enable_shared_cache(int) __OSX_AVAILABLE_BUT... ^ 1 warning generated. Undefined symbols for architecture x86_64: ""_sqlite3_enable_load_extension"", referenced from: _f_8972 in sqlite3.o ld: symbol(s) not found for architecture x86_64 clang: error: linker command failed with exit code 1 (use -v to see invocation) Error: shell command terminated with non-zero exit status 256: '/Applications/Xcode.app/Contents/Developer/usr/bin/gcc' 'sqlite3.o' -o 'sqlite3.so' -m64 -bundle -headerpad_max_install_names -lsqlite3 -L/Data/tools/chicken-4.11.0/lib -lchicken -lm Error: shell command failed with nonzero exit status 256: '/Data/tools/chicken-4.11.0/bin/csc' -feature compiling-extension -setup-mode -k -O2 -d1 -s sqlite3.scm -lsqlite3 -j sqlite3 Error: shell command terminated with nonzero exit code 17920 ""'/Data/tools/chicken-4.11.0/bin/csi' -bnq -setup-mode -e \""(require-library set... }}}" Norman Gray Milestone someday 1350 Static linking is broken on Solaris medium unknown 4.12.0 defect new 2017-03-08T05:06:23+01:00 2017-03-12T22:14:04+01:00 "While testing Peter's fix for #1347, I found that static linking is totally broken on Solaris. In particular, our linking tests fail on both Solaris 10 and 11 (gcc, i386 -- I haven't tried with any other compiler or architecture), albeit with slightly different error messages. The underlying issue might be the same, however, since they both fail in the same situations. Note that this problem isn't caused by the fix for #1347; both c4c60ce5 and 4.12.0 fail in the same way, so this was a preexisting bug. Try running the following command to see the problem: {{{ $ echo '(print 1)' | csc -static -o /tmp/a.out - }}} Test output on Solaris 10: {{{ ==> solaris-solaris-10-i386: ======================================== linking tests ... ==> solaris-solaris-10-i386: [... snip ...] ==> solaris-solaris-10-i386: /export/home/vagrant/chicken-core/tests/../chicken 'linking-tests.scm' -output-file 'linking-tests.c' -feature chicken-compile-static -verbose -include-path /export/home/vagrant/chicken-core/tests/.. -uses reverser ==> solaris-solaris-10-i386: 'gcc' 'linking-tests.c' -o 'linking-tests.o' -c -fno-strict-aliasing -fwrapv -DHAVE_CHICKEN_CONFIG_H -DC_ENABLE_PTABLES -Os -fomit-frame-pointer -I/export/home/vagrant/chicken-core/tests/.. -I/export/home/vagrant/chicken-core/tests/../include/chicken/ ==> solaris-solaris-10-i386: rm linking-tests.c ==> solaris-solaris-10-i386: 'gcc' 'linking-tests.o' 'reverser.o' -o 'linking-tests' -L/export/home/vagrant/chicken-core/tests/.. -L/export/home/vagrant/chicken-core/tests/../lib/ -Wl,-R'/export/home/vagrant/chicken-core/lib/' -static -lchicken -lsocket -lnsl -lm -ldl -lrt ==> solaris-solaris-10-i386: ld: fatal: library -lsocket: not found ==> solaris-solaris-10-i386: ld: fatal: library -lnsl: not found ==> solaris-solaris-10-i386: ld: fatal: library -lm: not found ==> solaris-solaris-10-i386: ld: fatal: library -ldl: not found ==> solaris-solaris-10-i386: ld: fatal: library -lrt: not found ==> solaris-solaris-10-i386: ld: fatal: library -lc: not found ==> solaris-solaris-10-i386: ld: fatal: file processing errors. No output written to linking-tests ==> solaris-solaris-10-i386: Error: shell command terminated with non-zero exit status 256: 'gcc' 'linking-tests.o' 'reverser.o' -o 'linking-tests' -L/export/home/vagrant/chicken-core/tests/.. -L/export/home/vagrant/chicken-core/tests/../lib/ -Wl,-R'/export/home/vagrant/chicken-core/lib/' -static -lchicken -lsocket -lnsl -lm -ldl -lrt }}} Test output on Solaris 11: {{{ ==> solaris-solaris-11-i386: ======================================== linking tests ... ==> solaris-solaris-11-i386: [... snip ...] ==> solaris-solaris-11-i386: /export/home/vagrant/chicken-core/tests/../chicken 'linking-tests.scm' -output-file 'linking-tests.c' -feature chicken-compile-static -verbose -include-path /export/home/vagrant/chicken-core/tests/.. -uses reverser ==> solaris-solaris-11-i386: 'gcc' 'linking-tests.c' -o 'linking-tests.o' -c -fno-strict-aliasing -fwrapv -DHAVE_CHICKEN_CONFIG_H -DC_ENABLE_PTABLES -Os -fomit-frame-pointer -I/export/home/vagrant/chicken-core/tests/.. -I/export/home/vagrant/chicken-core/tests/../include/chicken/ ==> solaris-solaris-11-i386: rm linking-tests.c ==> solaris-solaris-11-i386: 'gcc' 'linking-tests.o' 'reverser.o' -o 'linking-tests' -L/export/home/vagrant/chicken-core/tests/.. -L/export/home/vagrant/chicken-core/tests/../lib/ -Wl,-R'/export/home/vagrant/chicken-core/lib/' -static -lchicken -lsocket -lnsl -lm -ldl -lrt ==> solaris-solaris-11-i386: ld: fatal: option '-R/export/home/vagrant/chicken-core/lib/' is incompatible with building a static executable ==> solaris-solaris-11-i386: ==> solaris-solaris-11-i386: Error: shell command terminated with non-zero exit status 256: 'gcc' 'linking-tests.o' 'reverser.o' -o 'linking-tests' -L/export/home/vagrant/chicken-core/tests/.. -L/export/home/vagrant/chicken-core/tests/../lib/ -Wl,-R'/export/home/vagrant/chicken-core/lib/' -static -lchicken -lsocket -lnsl -lm -ldl -lrt }}}" evhan Milestone someday 1353 posix-semaphore egg won't build on OS X: uses deprecated API easy extensions 4.12.0 defect dleslie assigned 2017-03-14T00:45:58+01:00 2017-08-25T14:59:37+02:00 "Building the posix-semaphore egg fails on OS X, as it uses a deprecated API; see the log below. This is on OS X 10.11.6, with its associated version of clang (Apple LLVM version 8.0.0 (clang-800.0.42.1)). It appears to be only the support for unnamed semaphores (which are apparently optional in POSIX) that's deprecated; the support for named ones seems to be present. There's a StackOverflow discussion about the deprecation at (1) below, which points to some fairly authoritative discussion, including why the deprecation exists, at (2). (1) http://stackoverflow.com/questions/27736618/why-are-sem-init-sem-getvalue-sem-destroy-deprecated-on-mac-os-x-and-w (2) https://lists.apple.com/archives/darwin-kernel/2009/Apr/msg00010.html {{{ % chicken-install posix-semaphore retrieving ... connecting to host ""chicken.kitten-technologies.co.uk"", port 80 ... requesting ""/henrietta.cgi?name=posix-semaphore&mode=default"" ... reading response ... HTTP/1.1 200 OK Date: Mon, 13 Mar 2017 23:30:47 GMT Server: Apache/2.2.31 (Unix) DAV/2 PHP/5.5.36 mod_fastcgi/2.4.6 Connection: close Transfer-Encoding: chunked Content-Type: text/plain reading chunks . reading files ... ./readme.md ./LICENSE ./posix-semaphore.release-info ./posix-semaphore.meta ./posix-semaphore.setup ./posix-semaphore.scm ./test.scm posix-semaphore located at /var/folders/7b/1fldmb290mzbgxppkpwwffx40000gn/T/temp3c46.40550/posix-semaphore checking platform for `posix-semaphore' ... checking dependencies for `posix-semaphore' ... install order: (""posix-semaphore"") installing posix-semaphore:0.5.1 ... changing current directory to /var/folders/7b/1fldmb290mzbgxppkpwwffx40000gn/T/temp3c46.40550/posix-semaphore '/Data/tools/chicken-4.12/bin/csi' -bnq -setup-mode -e ""(require-library setup-api)"" -e ""(import setup-api)"" -e ""(setup-error-handling)"" -e ""(extension-name-and-version '(\""posix-semaphore\"" \""0.5.1\""))"" 'posix-semaphore.setup' '/Data/tools/chicken-4.12/bin/csc' -feature compiling-extension -setup-mode -O3 -C -O3 -lpthread -s -d1 -j posix-semaphore posix-semaphore.scm posix-semaphore.c:63:11: warning: 'sem_getvalue' is deprecated [-Wdeprecated-declarations] int err = sem_getvalue(sem, &val); ^ /usr/include/sys/semaphore.h:54:5: note: 'sem_getvalue' has been explicitly marked deprecated here int sem_getvalue(sem_t * __restrict, int * __restrict) __deprecated; ^ posix-semaphore.c:92:1: warning: implicit declaration of function 'clock_gettime' is invalid in C99 [-Wimplicit-function-declaration] clock_gettime(CLOCK_REALTIME, &tm); ^ posix-semaphore.c:92:15: error: use of undeclared identifier 'CLOCK_REALTIME' clock_gettime(CLOCK_REALTIME, &tm); ^ posix-semaphore.c:95:7: warning: implicit declaration of function 'sem_timedwait' is invalid in C99 [-Wimplicit-function-declaration] CHECK(sem_timedwait(sem, &tm)) ^ posix-semaphore.c:180:7: warning: 'sem_destroy' is deprecated [-Wdeprecated-declarations] CHECK(sem_destroy(sem)) ^ /usr/include/sys/semaphore.h:53:5: note: 'sem_destroy' has been explicitly marked deprecated here int sem_destroy(sem_t *) __deprecated; ^ posix-semaphore.c:192:7: warning: 'sem_init' is deprecated [-Wdeprecated-declarations] CHECK(sem_init(sem, shared, value)) ^ /usr/include/sys/semaphore.h:55:5: note: 'sem_init' has been explicitly marked deprecated here int sem_init(sem_t *, int, unsigned int) __deprecated; ^ 5 warnings and 1 error generated. Error: shell command terminated with non-zero exit status 256: '/Applications/Xcode.app/Contents/Developer/usr/bin/gcc' 'posix-semaphore.c' -o 'posix-semaphore.o' -c -fno-strict-aliasing -fwrapv -fno-common -DHAVE_CHICKEN_CONFIG_H -m64 -DC_ENABLE_PTABLES -Os -fomit-frame-pointer -fPIC -DPIC -DC_SHARED -O3 -I/Data/tools/chicken-4.12/include/chicken Error: shell command failed with nonzero exit status 256: '/Data/tools/chicken-4.12/bin/csc' -feature compiling-extension -setup-mode -O3 -C -O3 -lpthread -s -d1 -j posix-semaphore posix-semaphore.scm Error: shell command terminated with nonzero exit code 17920 ""'/Data/tools/chicken-4.12/bin/csi' -bnq -setup-mode -e \""(require-library setup... }}}" Norman Gray Milestone someday 1372 mailbox timeouts and thread signaling extensions 4.12.0 defect Kon Lovett assigned 2017-05-23T11:31:33+02:00 2017-08-25T16:56:34+02:00 "There appears to be a scheduler problem in the mailbox egg. If I signal a thread waiting on mailbox-receive! with a timeout, the program hangs and csi eats all my cpu. {{{#!scheme (use srfi-18 mailbox) (define mbox (make-mailbox)) (define primordial (current-thread)) (define t (thread-start! (lambda () (thread-sleep! 1) (thread-signal! primordial 'example)))) ;; this hangs forever and eats all my cycles (with timeout) (print (mailbox-receive! mbox 4)) ;; this exits as expected with the 'example exception (no timeout) (print (mailbox-receive! mbox)) }}}" Caolan McMahon Milestone someday 1399 Scrutinizer produces incorrect procedure types after merge medium scrutinizer 4.12.0 defect evhan reopened 2017-09-09T10:00:20+02:00 2023-11-06T22:05:01+01:00 "Suppose we get this expression after scrutinizing some code: {{{ (or (procedure (|#!rest|) noreturn) (procedure (* |#!rest|) . *)) }}} At the moment the result of the union is `(procedure (#!rest))` which is clearly wrong because it has arity N >= 0 while the second procedure in the `or` expression has arity N >= 1. Moreover the return type is lost, I'd expect the `noreturn` to be dropped and `*` to be used instead. The result of the miscalculation of the resulting type may or may not be dangerous." LemonBoy Milestone someday 1404 aima: Error: (match) during expansion of (make-debug-environment ...) - no matching pattern extensions 4.12.0 defect Peter Danenberg new 2017-09-12T18:52:08+02:00 2017-09-12T18:52:08+02:00 See https://salmonella-linux-x86-64.call-cc.org/master-debugbuild/gcc/linux/x86-64/2017/09/12/salmonella-report/install/aima.html Mario Domenech Goulart Milestone someday 1411 ssql-record (tests): Error: unbound variable: test-table-rec-update (and other failures/errors) extensions 4.12.0 defect arthurmaciel new 2017-09-12T19:39:20+02:00 2017-09-12T19:39:20+02:00 See https://salmonella-linux-x86-64.call-cc.org/master-debugbuild/gcc/linux/x86-64/2017/09/12/salmonella-report/test/ssql-record.html Mario Domenech Goulart Milestone someday 1412 msgpack (tests): Error: segmentation violation extensions 4.12.0 defect hugoarregui new 2017-09-12T19:42:07+02:00 2017-09-12T19:42:07+02:00 See https://salmonella-linux-x86-64.call-cc.org/master-debugbuild/gcc/linux/x86-64/2017/09/12/salmonella-report/test/msgpack.html Mario Domenech Goulart Milestone someday 1416 chunk-vector: many unbound identifiers extensions 4.12.0 defect pluizer new 2017-09-16T08:40:00+02:00 2017-09-16T08:40:00+02:00 See https://salmonella-linux-x86-64.call-cc.org/master-debugbuild/gcc/linux/x86-64/2017/09/15/salmonella-report/install/chunk-vector.html Mario Domenech Goulart Milestone someday 1450 TCP connections can cause process to hang hard core libraries 4.13.0 defect new 2018-03-19T08:42:09+01:00 2023-11-06T22:11:22+01:00 "As reported by Jim Ursetto, his Spiffy server (which is running on localhost behind a NGINX proxy) will stop responding at some point. It is currently not 100% clear when exactly this happens. As Jim says: > I believe I’ve tracked down the problem, although not the solution yet. It seems the file descriptor table is filled up with half open sockets. lsof shows: > > … > chickadee 13361 jim 1019u sock 0,7 0t0 2365037 can't identify protocol > chickadee 13361 jim 1020u sock 0,7 0t0 2366414 can't identify protocol > chickadee 13361 jim 1021u sock 0,7 0t0 2368047 can't identify protocol > chickadee 13361 jim 1022u sock 0,7 0t0 2368343 can't identify protocol > > And this message on Linux seems to occur when sockets are half-open (or half-closed, if you are a pessimist). And later: > my hunch is that this happens when the connecting side hangs up while we are still sending. Which is probably obvious from my previous description of the problem. I was going to start attacking it by inserting strategic sleeps at various locations to try and manually trigger it, although you probably have a better way. Obviously you do appear to catch all errors in spiffy, so I’m not sure if one of those error handlers is neglecting to close a socket, if the socket close fails and isn’t reported, or if this is a deeper bug inside the tcp unit. I hear faint echoes of #340..." sjamaan Milestone someday 1453 Running MinGW-produced Chicken from a MSVC-produced executable causes a stack overflow during initialization unknown 4.13.0 defect new 2018-04-18T17:43:14+02:00 2018-04-23T15:32:31+02:00 "I am on Windows. I have a program, compiled via MSVC, that has a dependency on libchicken.dll, compiled via MinGW, both 64 bit. It successfully calls CHICKEN_run with the default toplevel, but during the loading of the modules, the program segfaults with a stack overflow error in various places, usually around the time it hits loading modules.scm. It seems that Chicken underestimates the size of the stack when ran in this manner, causing it to run out of stack space much sooner than it anticipates, causing the segfault when it fails to resize the stack. Attached is a set of files that can replicate the issue. Running make should produce call_from_cl.exe, which will segfault when run. The libchicken.dll it needs should be produced via the mingw-msys platform. You'll need MSVC's cl.exe and lib.exe installed to compile." jrobbins Milestone someday 1454 process-wait does not function corrently with nohang on Windows trivial core libraries 4.13.0 defect new 2018-04-23T15:23:30+02:00 2018-04-23T15:23:30+02:00 On Windows, if you specify #t to nohang when calling process-wait, and the process is still running, you will get an error with a bogus error message. This is because process-wait uses WaitForSingleObject on Windows, and the result code WAIT_TIMEOUT is not handled, being treated as an error even though there was never an error in the first place. Attached is a patch to fix this bug. jrobbins Milestone someday 1465 get-keyword fails to operate correctly if keywords in first and second arguments come from different environments trivial core libraries 4.13.0 defect new 2018-05-23T20:08:24+02:00 2018-05-23T20:08:24+02:00 "Run the following code in csi, and you'll get a failed assertion: {{{ (define-syntax keyword-from-syntax (syntax-rules () ((_) #:k))) (assert (get-keyword (keyword-from-syntax) '(h i j #:k l))) }}} This should return {{{l}}}, but instead it returns {{{#f}}}. The bug occurs when a keywords from different environments are compared; the {{{#:k}}} from the syntax-rules is not being treated as equal to the {{{#:k}}} from the normal code for some reason. A working replacement (and the replacement I use in my code) could be as follows: {{{ (define (get-keyword kw args . default) (let ( (tail (memq kw args)) ) (if (and tail (not (null? (cdr tail)))) (cadr tail) (if (null? default) #f ((car default)) ) ) ) ) }}}" jrobbins Milestone someday 1475 (scrutinizer) Types for global variables are not refined by predicates easy scrutinizer 5.0.0 defect new 2018-06-15T09:08:36+02:00 2023-11-06T22:12:45+01:00 "{{{ (: g1 (list-of fixnum)) (define g1 '()) ;; Aliased global gets refined (let ((a g1)) (compiler-typecase a ((list-of fixnum) 1)) (if (null? a) (compiler-typecase a (null 1)) (compiler-typecase a ((not null) 1)))) (: g2 (list-of fixnum)) (define g2 '()) (if (null? g2) (compiler-typecase g2 (null 1)) ;; <- error (compiler-typecase g2 ((pair fixnum (list-of fixnum)) 1))) ;; $ csc -O3 globals.scm ;; ;; Error: at toplevel: ;; (globals.scm:13) no clause applies in `compiler-typecase' for expression of type `(list-of fixnum)': ;; null }}} I don't think this is an issue of non-local mutation (e.g. from other threads). You should use mutexes or some other measures to prevent mutation while, you're working on a global." megane Milestone someday 1496 zlib: checksum mismatch in tests starting on 2018-08-05 extensions 4.13.0 defect joseph.gay new 2018-08-06T08:14:48+02:00 2021-04-01T19:44:17+02:00 "See http://salmonella-linux-x86.call-cc.org/chicken-4-debugbuild/gcc/linux/x86/2018/08/05/salmonella-report/test/zlib.html That happened with CHICKEN at 5f2c3ae6. The day before this error didn't occur and CHICKEN was at d32b83d5." Mario Domenech Goulart Milestone someday 1503 Fix dev-snapshot script infrastructure 5.0.0 defect Mario Domenech Goulart new 2018-08-11T11:19:00+02:00 2018-08-11T11:19:00+02:00 Peter reported it is broken. Mario Domenech Goulart Milestone someday 1513 Type declarations in modules do not namespace structs easy core libraries 5.0.0 defect assigned 2018-08-20T21:49:00+02:00 2023-11-06T22:05:49+01:00 "As reported by Jörg Wittenberger, the following program gives a compilation warning in CHICKEN 5 but not in CHICKEN 4: {{{ (module foo * (import scheme) (cond-expand (chicken-5 (import (chicken base) (chicken type) )) (else (import chicken))) (: make-foo (string --> (struct foo))) (define-record foo bar) ) }}} Presumably this is because the low-level name of the struct is `foo#foo` here (because we're inside a module). I think the correct fix for this would be to look at struct identifiers, and when unqualified, just prefix it." sjamaan Milestone someday 1519 salt: test failure: can't read MLB path map file extensions 5.0.0 defect Ivan Raikov new 2018-08-21T21:31:55+02:00 2018-08-22T19:16:39+02:00 See http://salmonella-linux-arm64.call-cc.org/master-debugbuild/gcc/linux/arm64/2018/08/21/salmonella-report/test/salt.html Mario Domenech Goulart Milestone someday 1524 environments 1.6.2 breaks the installation of eggs which depend on it extensions 4.13.0 defect new 2018-08-25T10:09:34+02:00 2018-08-25T10:16:47+02:00 "Attempting to install environments with recent CHICKENs break with {{{ Error: Sorry, but CHICKEN 4.7.4 and later is not supported anymore }}} This [https://salmonella-linux-x86-64.call-cc.org/chicken-4-debugbuild/gcc/linux/x86-64/2018/08/24/salmonella-report/rev-dep-graphs/environments.html breaks the installation of eggs which depend on environments]: * http://salmonella-linux-x86.call-cc.org/chicken-4-debugbuild/gcc/linux/x86/2018/08/24/salmonella-report/install/uri-dispatch.html * http://salmonella-linux-x86.call-cc.org/chicken-4-debugbuild/gcc/linux/x86/2018/08/24/salmonella-report/install/couchdb-view-server.html I don't know if these eggs actually hard-depend on environments to work. In case they do, I think they should be moved to the ""obsolete"" category as well." Mario Domenech Goulart Milestone someday 1527 Procedures grouping in the wiki is confusing easy wiki defect new 2018-08-31T23:31:45+02:00 2018-09-01T20:59:22+02:00 "Compared to the source file of the documentation, or to chickadee, the wiki is very confusing when showing grouping of multiple procedures. It groups procedures even when they are separated by a new line in the source, which looks like these procedures are linked together even when they are not. As an example, look at the ncurses egg documentation, more precisely at the halfdelay procedure. https://wiki.call-cc.org/eggref/4/ncurses https://api.call-cc.org/doc/ncurses In the wiki, it’s grouped with has_colors, has_ic and has_il, and everything is documented to return a boolean value. In chickadee, halfdelay is by itself, and only has_colors, has_ic and has_il are grouped together with their documentation entry. " Kooda Milestone someday 1561 Rewrite .egg file before installation medium extensions 5.0.0 defect new 2018-11-20T11:20:10+01:00 2018-11-20T11:20:10+01:00 "This was suggested by Kooda: install the expanded egg file in the extension repository, after all `cond-expand` forms have been processed. This reflects the configuration that was used during installation (respecting `-f` options to `chicken-install`) and does not require re-processing conditionals. " felix winkelmann Milestone someday 1566 ports-tests are not testing file-lock procedures trivial core libraries 5.0.0 defect Christian Kellermann new 2018-11-30T22:46:38+01:00 2023-11-08T21:17:24+01:00 Our ports-tests script does call file-lock and friends but only to test whether they raise an exception on already closed ports. There is no test whether these procedures actually do anything. That's why #1565 has gone unnoticed. Christian Kellermann Milestone someday 1571 Sources for blowfish egg no longer available in bitbucket extensions 4.13.0 defect rivo new 2018-12-22T18:09:36+01:00 2018-12-22T18:09:36+01:00 "We've been seeing the following error from henrietta-cache: {{{ Could not fetch release-info file for egg blowfish (CHICKEN release 4) from https://bitbucket.org/rivo/blowfish-egg/raw/tip/blowfish.release-info -- list-terminator mismatch }}} The error happens because https://bitbucket.org/rivo/blowfish-egg/raw/tip/blowfish.release-info redirects to a login page. Apparently, blowfish is not available in bitbucket anymore. Where can we find it?" Mario Domenech Goulart Milestone someday 1574 Redefining reverse in csi makes lambda behave differently, and even segfault hard compiler 5.0.0 defect new 2018-12-30T15:25:58+01:00 2018-12-30T15:25:58+01:00 "Running this in csi triggers a segfault in ##sys#decompose-lambda-list (eval.scm), even though reverse is bound in a let. {{{ (define reverse 1) (lambda (foo . bar) bar) }}}" Kooda Milestone someday 1596 "envsubst: test failure: Error: (include) cannot open file: ""tests/call-with-environment-variables.scm""" trivial extensions 5.0.0 defect Robert C Jensen assigned 2019-03-22T18:37:31+01:00 2021-08-29T21:08:04+02:00 See http://salmonella-linux-x86.call-cc.org/master-debugbuild/gcc/linux/x86/2019/03/22/salmonella-report/test/envsubst.html Mario Domenech Goulart Milestone someday 1597 dbus 0.96: installation failure trivial extensions 4.13.0 defect Shawn Rutledge new 2019-03-22T20:17:54+01:00 2019-03-22T20:17:54+01:00 "See https://salmonella-linux-x86-64.call-cc.org/chicken-4-debugbuild/gcc/linux/x86-64/2019/02/23/salmonella-report/install/dbus.html Note that this is for CHICKEN 4." Mario Domenech Goulart Milestone someday 1599 "spock: failure in tests: Error: library not found: ""syntax.scm""" extensions 5.0.0 defect Mario Domenech Goulart assigned 2019-03-28T22:34:04+01:00 2019-03-29T19:07:28+01:00 See http://salmonella-linux-x86.call-cc.org/master-debugbuild/gcc/linux/x86/2019/03/28/salmonella-report/test/spock.html Mario Domenech Goulart Milestone someday 1607 zlib: failure in tests hard extensions 5.0.0 defect Robert C Jensen new 2019-04-13T16:18:10+02:00 2021-07-01T02:37:39+02:00 See http://salmonella-linux-x86.call-cc.org/master-debugbuild/gcc/linux/x86/2019/04/13/salmonella-report/test/zlib.html Mario Domenech Goulart Milestone someday 1615 hidden/gensymed variables are registered for export easy expander 5.0.0 defect sjamaan new 2019-05-07T20:54:17+02:00 2023-11-14T18:51:09+01:00 "This might cause problems if an interned symbol with the same name is defined. Find out if this really is a problem. Also: alias global hook strikes again ;) Example: {{{ (module foo (x123) (import scheme (chicken syntax)) (define-syntax define-x123 (ir-macro-transformer (lambda (e i c) (let ((name (gensym 'x))) `(define ,name 'blah)))) (define x123 5) (define-x123) ;; Should not overwrite the interned version of x123, which happens if both collide on foo#x123! ) (import foo) (print x123) ;; Should print 5, not blah }}} Of course this requires some fiddling to get the numbers right, which would differ per CHICKEN version (a proper test would require reading the gensym counter to force the collision, or something like that)." sjamaan Milestone someday 1632 atom: failure in tests extensions 5.1.0 defect Jim Ursetto new 2019-07-13T09:15:14+02:00 2019-07-13T09:15:14+02:00 See https://salmonella-linux-x86-64.call-cc.org/master/gcc/linux/x86-64/2019/07/13/salmonella-report/test/atom.html Mario Domenech Goulart Milestone someday 1685 shell-variable in egg-compile should quote environment variables core libraries 5.2.0 defect evhan assigned 2020-03-09T20:43:47+01:00 2023-11-06T22:08:50+01:00 "Without quoting shell variables, install scripts end up with commands like {{{ mkdir -p ${DESTDIR}'/home/mario/local/chicken-head/lib/chicken/11' }}} If {{{DESTDIR}}}} happens to contain a space, the command above will not do what it is expected to do. Double-quoting shell variables on Unix systems is not harmful, as far as I can tell. I have no clue about Windows. At the moment, {{{shell-variable}}} seems to be only applied to {{{DESTDIR}}}. The following patch would improve the situation on Unix: {{{ diff --git a/egg-compile.scm b/egg-compile.scm index 4a72d5d0..e6be5d1d 100644 --- a/egg-compile.scm +++ b/egg-compile.scm @@ -1227,7 +1227,7 @@ EOF (define (shell-variable var platform) (case platform - ((unix) (string-append ""${"" var ""}"")) + ((unix) (string-append ""\""${"" var ""}\"""")) ((windows) (string-append ""%"" var ""%"")))) ;; NOTE `cmd' must already be quoted for shell }}}" Mario Domenech Goulart Milestone someday 1696 mutex-lock! doesn't always lock mutex insane unknown 5.2.0 defect new 2020-04-18T14:24:09+02:00 2020-04-18T14:30:06+02:00 "{{{ ;; From C4 tests/mutex-test.scm (import scheme) (cond-expand (chicken-5 (import (chicken base) (chicken type) (chicken format) srfi-18)) (else (import chicken) (use srfi-18))) (define-record-printer (mutex x out) (format out """" (mutex-name x) (if (##sys#slot x 5) ""LOCKED"" ""FREE"") (if (##sys#slot x 4) ""/ABANDONED"" """") (mutex-state x) (if (##sys#slot x 2) (##sys#slot x 2) ""none"") (##sys#slot x 3) )) (let ((m1 (make-mutex))) ;; (print m1) (mutex-lock! m1) (print m1) (print* ""l"") ;; (thread-sleep! 0.001) (let ((t1 (thread-start! (lambda () (print* ""t"") (mutex-lock! m1 0.001) (print ""\noh no, we got the mutex"") (print m1) (exit 1))))) (print* ""y"") (thread-yield!) (print* ""Y"") (when (eq? 'ready (##sys#slot t1 3)) ;; <- #3 thread state (print* ""X"") (thread-sleep! 1) (print* ""Z"")))) (print* ""."") ;; OK run: ;; lytY. (owner #) waiting ()> ;; Failing run: ;; lytYX ;; oh no, we got the mutex ;; (owner #) waiting ()> }}}" megane Milestone someday 1714 types.db not found on mingw32 unknown 5.2.0 defect new 2020-08-08T15:35:38+02:00 2020-12-13T15:46:43+01:00 "Hi, On my build (Windows, with an old mingw32) `types.db` is not found anymore. {{{ D:\tmp\chicken\'types.db' not found>type str.scm (display ""Sunday"") (newline) D:\tmp\chicken\'types.db' not found>chicken str.scm -output-file ""str.c"" Error: default type-database `types.db' not found }}} Chicken is installed in ""C:\Program Files (x86)\Chicken Scheme"", and types.db is in ""c:\Program Files (x86)\Chicken Scheme\lib\chicken\11\"" . Where does csc look for `types.db`?" Răzvan Rotaru Milestone someday 1716 coops + s11n malfunction extensions 5.1.0 defect new 2020-08-19T21:38:08+02:00 2020-08-24T16:20:38+02:00 "{{{ #!scheme (define-class frotz () ((name accessor: name))) (define xyzzy (make frotz 'name ""blue jeans"")) (name xyzzy) (with-output-to-file ""/tmp/frotz"" (lambda () (serialize xyzzy))) (define nitfol (with-input-from-file ""/tmp/frotz"" deserialize)) nitfol ; <= evals to # (name nitfol); <= fails with Error: (name) no method defined for given argument classes: (#'>) }}} This used to work and print ""blue jeans"" but I'm not sure when it stopped working. I especially hope I can rescue all my other saved objects because I have a lot of precious in there♥" Idiomdrottning Milestone someday 1728 chicken-install postgresql fails on Windows unknown 5.2.0 defect new 2021-01-29T05:07:50+01:00 2021-05-18T17:24:21+02:00 "{{{ c:\>chicken-install postgresql building postgresql C:\Users\j\AppData\Local\chicken-install\postgresql\build-postgresql.bat -host -D compiling-extension -J -s -regenerate-import-libraries -setup-mode -I C:\Users\j\AppData\Local\chicken-install\postgresql -C -IC:\Users\j\AppData\Local\chicken-install\postgresql -O2 -d1 postgresql.scm -o C:\Users\j\AppData\Local\chicken-install\postgresql\postgresql.so creating subprocess failed Error: shell command terminated with nonzero exit code 1 ""C:\\Users\\j\\AppData\\Local\\chicken-install\\postgresql\\postgresql.build.bat"" c:\> }}} " Josh Helzer Milestone someday 1729 [fmt egg]: Num fails with numbers near 0 extensions 5.2.0 defect foof assigned 2021-03-01T15:59:51+01:00 2021-03-01T15:59:51+01:00 "{{{ $ csi5 -P '(import fmt) (fmt #f (num (/ 1.0 1e300) 10 2))' ;; Error: (quotient) bad argument type - not an integer: +inf.0 }}}" megane Milestone someday 1745 apropos doesn't report macros medium extensions 5.2.0 defect Kon Lovett assigned 2021-04-12T05:25:27+02:00 2021-04-21T00:06:26+02:00 Kon Lovett Milestone someday 1759 Module system weirdness unknown 5.2.0 defect new 2021-06-01T18:18:40+02:00 2021-06-01T18:18:40+02:00 "So the trippy thing is that if I start a fresh REPL and (import brev-separate) without importing matchable nor import-for-syntax anything, {{define-ir-syntax*}} works. When I try to do that inside a module, it doesn't work. {{define-ir-syntax*}} is a macro that makes macros, using matchable to do so. OK, so let me try something else: When I make a file not in a module but just something.scm and in that file only put: (import brev-separate) (define-ir-syntax* (jolly) '(print ""This is jolly good"")) (jolly) and then csc it and run it, it works. But inside of a module it does not work. I need {{(import-for-syntax matchable)}}. This behavior, I don't know if it's a bug or not but it was unintuitive enough to send me on an 8 hour rabbit hole trek. Which is fine. I was embarrassed in the IRC channel but I'm glad everyone was cool. I wish I understood the module system better. I feel like I'm just klutzing around, testing out what works and doesn't work. " Idiomdrottning Milestone someday 1784 Modify chickadee to say what module an identifier is in easy documentation 5.2.0 defect new 2021-09-10T20:59:48+02:00 2021-09-10T23:00:05+02:00 Right now, if you want to call `features`, you have to know or guess which module to load. Changing chickadee to provide that information will help a great deal. (Looking through all the individual module pages works, but ugh.) johnwcowan Milestone someday 1792 Some eggs install into system /share and /bin (even when CHICKEN_INSTALL_REPOSITORY is set); this fails when the directory is RO unknown 5.2.0 defect new 2021-10-14T17:58:15+02:00 2021-10-14T23:20:15+02:00 "I'm using Nixpkgs (hence the paths below), but the same would be true in another case when the system /share is for whatever reason not expected to be writable by chicken-install. Since .../share is expected to be read-only in this context, I've defined a separate repository for eggs to be installed in: CHICKEN_REPOSITORY_PATH=$HOME/.chicken-install:/nix/store/1sy3rv81crwlviknmqnfs1cwz3bbqr1g-chicken-5.2.0/lib/chicken/11 CHICKEN_INSTALL_REPOSITORY=$HOME/.chicken-install CHICKEN_INCLUDE_PATH=/nix/store/na72pj8gy1gyqd0llz6myvk8dzd4q5i0-compiler-rt-libc-7.1.0/share:/nix/store/1sy3rv81crwlviknmqnfs1cwz3bbqr1g-chicken-5.2.0/share That works OK for several eggs, but srfi-29 fails: {{{ installing srfi-29 mkdir: cannot create directory ‘/nix/store/1sy3rv81crwlviknmqnfs1cwz3bbqr1g-chicken-5.2.0/share/chicken/srfi-29-bundles’: Permission denied }}} and sure enough, $HOME/.cache/chicken-install/srfi-29/srfi-29.install.sh includes {{{ mkdir -p ${DESTDIR}'/nix/store/1sy3rv81crwlviknmqnfs1cwz3bbqr1g-chicken-5.2.0/share/chicken' mkdir -p ${DESTDIR}'/nix/store/1sy3rv81crwlviknmqnfs1cwz3bbqr1g-chicken-5.2.0/share/chicken/srfi-29-bundles' install -m 644 '/Users/norman/.cache/chicken-install/srfi-29/srfi-29-bundles/srfi-29' ${DESTDIR}'/nix/store/1sy3rv81crwlviknmqnfs1cwz3bbqr1g-chicken-5.2.0/share/chicken/srfi-29-bundles' }}} That is, it seems to unconditionally want to install in the system share directory. The same is true of some other eggs (I'm not sure this is an exhaustive list: Kon Lovett on-list suggested that these, like srfi-29, specify a 'data files’ component to the extension, which chicken-install places in '(chicken-home)’): sqlite3: tried to install stuff into system /share spock and manual-labor: tried to install stuff into system /bin and /share In case anyone isn't aware NixOS, and Nixpkgs as a package system for other OSs, is predicated on the idea that the system directories are owned entirely by the Nix system, so that .../share/FOO is out of bounds. I'm using Nixpkgs 2.3.16 on macOS 11.6, though I don't think either of those versions is relevant to the issue). " Norman Gray Milestone someday 1808 Srfi-42 index keyword hygiene extensions 5.2.0 defect new 2022-07-29T21:04:10+02:00 2022-07-30T00:48:29+02:00 "Found by Chris (so ignore this duplicate if he already got to posting it): {{{ (import sequences srfi-42) (do-ec (:vector elt (index i) '#(a b c)) (print i elt)) }}} (Because of the collision with `index`.)" Idiomdrottning Milestone someday 1810 ezd man page not installed unknown 5.3.0 defect new 2022-08-23T13:57:26+02:00 2022-08-23T13:57:26+02:00 The ezd.1 man page is in the doc directory of the ezd egg, but it is discarded along with the rest of the cached files when the build is complete. It should be copied into /usr/local/share/man/man1 (or the local equivalent), as it is the only good documentation of the whole ezd command set. johnwcowan Milestone someday 1815 (chicken irregex): Use condition kinds core libraries 5.3.0 defect new 2023-02-10T19:59:18+01:00 2023-02-10T19:59:18+01:00 The (chicken irregex) module should raise conditions with kinds. It's hard to handle exceptions automatically without them. For example, 'irregex' raises an exception when called with a syntactically-invalid regular expression or with an object of the wrong type (say, a vector). The only way to tell which of these problems occurred is through a nasty kluge: examining the condition's message. It should instead raise a condition with kind '(exn type)' for the type error (like the rest of CHICKEN's core forms) and with kind '(exn irregex)' (e.g.) for the regular expression syntax error. Zipheir Milestone someday 1818 Exporting a macro inhibits compiled program from locating libraries compiler 5.3.0 defect new 2023-03-30T16:22:01+02:00 2023-11-08T23:04:37+01:00 "The bug I found makes the compiled program unable to locate libraries based on the environment variables `CHICKEN_INCLUDE_PATH` and `CHICKEN_REPOSITORY_PATH`. The bug is quite fiddly to reproduce so I've added a 7z archive that contains the neccessary file structure and sets the environment variables appropriately. To compile and run it execute the make.sh script. The program runs as expected as long as the 5. line (exporting the macro) in the test.scm file is commented out, but fails to load the testlib library if it's included. What makes the bug so fiddly and likely the reason why it hasn't been noticed so far is that everything works as expected as long as the include *.import.scm files and the library *.so files are in the same directory, as pointed by CHICKEN_REPOSITORY_PATH. This can be observed by commenting out the 8. line in the make.sh script. Therefore I hope it's a simple error in using a wrong variable to get the search directory for the *.import.scm files. Lastly, I've run it on Arch Linux updated as of 2023.03.30: {{{#!sh $ chicken-csc -version CHICKEN (c) 2008-2021, The CHICKEN Team (c) 2000-2007, Felix L. Winkelmann Version 5.3.0 (rev e31bbee5) linux-unix-gnu-x86-64 [ 64bit dload ptables ] }}} " Filip Jagiełłowicz Milestone someday 1829 CHICKEN nixos package is incorrectly configured easy build system 5.3.0 defect new 2023-12-21T22:36:57+01:00 2023-12-21T22:36:57+01:00 "See https://github.com/NixOS/nixpkgs/issues/275901 " felix winkelmann Milestone someday 1832 Exported syntaxes are not fully hygienic hard compiler 5.3.0 defect new 2024-02-22T21:44:24+01:00 2025-11-03T06:45:56+01:00 "Modules exporting syntaxes generated by syntaxes are not fully hygienic as the serialization into module-name.import.scm uses strip-syntax on module's sexports. == Minimal (Working) Example Attached as follows: mwe-ok.scm - single file where the macros show expected behavior mwe-ko-common.scm - module with the macro generating macros mwe-ko-one.scm - first macro generated by the macro from common module mwe-ko-two.scm - second macro generated by the macro from module ""one"" mw-ko-run.scm - should behave like mwe-ok.scm, shows binding collisions == How to run Reference (working) module: {{{#!sh csi -q -b mwe.ok.scm }}} Multiple modules with problems: {{{#!sh csc -P -J mwe-ko-common.scm csc -P -J mwe-ko-one.scm csc -P -J mwe-ko-two.scm csi -q -b mwe-ko-run.scm }}} " Dominik Joe Pantůček Milestone someday 1835 Error in the utf8 egg's substring=? extensions 5.4.0 defect siiky reopened 2024-04-12T00:10:31+02:00 2024-04-23T18:23:06+02:00 "First found it using tab-completion in {{{csi}}}: {{{ Error: (substring=?) out of range 3 0 3 Call history: utf8-srfi-13.scm:374: proc utf8-srfi-13.scm:374: proc utf8-srfi-13.scm:374: proc utf8-srfi-13.scm:374: proc utf8-srfi-13.scm:374: proc utf8-srfi-13.scm:374: proc utf8-srfi-13.scm:374: proc utf8-srfi-13.scm:374: proc utf8-srfi-13.scm:374: proc utf8-srfi-13.scm:374: proc utf8-srfi-13.scm:374: proc utf8-srfi-13.scm:374: proc utf8-srfi-13.scm:374: proc utf8-srfi-13.scm:374: proc utf8-srfi-13.scm:374: proc utf8-srfi-13.scm:374: proc <-- }}} This is seen in the egg's tests as well: http://salmonella-linux-x86.call-cc.org/master/clang/linux/x86/2024/04/11/salmonella-report/test/utf8.html ({{{chicken-install}}}'s exit code is 0, I'm not seeing why -- there's a {{{test-begin}}} at the top and a {{{test-end}}} at the bottom -- though this is another problem): {{{ (string-prefix? ""元麻布"" ""麻布十番"" 1) ........................ [ERROR] Error: (substring=?) out of range 6 0 6 (string-prefix? ""元麻布"" ""麻"" 1 2) ............................... [ERROR] Error: (substring=?) out of range 3 0 3 }}} It's almost certain these errors are related to the recent changes made to {{{substring=?}}} and friends: https://code.call-cc.org/cgi-bin/gitweb.cgi?p=chicken-core.git;a=commit;h=0e97b648407d750f1e476e3c193e136126558346" siiky Milestone someday 1836 Chicken doesn't build out of the box on recent macosx trivial build system 5.3.0 defect new 2024-06-07T12:29:40+02:00 2024-06-07T12:29:40+02:00 "A friend wanted to build Chicken 5.3.0 on his ""Mac OS 14"" mac, and complained that it looked in the wrong place for tools. He had to override stuff to make it build: {{{ make PLATFORM=macosx PREFIX=...redacted... C_COMPILER=/usr/bin/gcc \ LIBRARIAN=/usr/bin/ar POSTINSTALL_PROGRAM=/usr/bin/install_name_tool \ install }}} He has the XCode cli tools installed but not the UI. The makefile was trying to find stuff under /Developer/... but it ain't there no more." Alaric Snell-Pym Milestone someday 1840 Chicken Scheme fails to build for xtensa unknown 5.3.0 defect new 2024-07-22T15:03:16+02:00 2024-07-22T15:03:16+02:00 "I'm trying to fix a problem detected by Buildroot's autobuilder, in which Chicken does not build for the xtensa architecture. Xtensa is not a supported architecture, but the patch to build there is very small and non-invasive. Please consider merging. This bug was originally discovered under version 5.3.0, but I have tested this patch against 5.4.0 and it works. a reference to the buildroot autobuild failure is here, for reference: http://autobuild.buildroot.net/results/0428595153756ab37a7b8eb421b0d58c3b357b9b/ " Woodrow E Douglass Milestone someday 1842 Possibly inconsistent install/lookup behavior in the presence of CHICKEN_INSTALL_PREFIX core libraries 5.4.0 defect new 2024-08-26T18:30:43+02:00 2024-08-26T18:33:22+02:00 "The title is very poor and the description is going to be a paste of the discussion in #chicken. Basically, the issue is highlighted by one of the tests of srfi-29 {{{ -- testing Installed Test Bundle ------------------------------------- (localized-template/default 'srfi-29 'srfi-29) ............... [ FAIL] expected ""SRFI 29"" but got srfi-29 }}} Discussion in #chicken: {{{ klovett: apparently srfi-29 is installing the srfi-29 bundle via the `data' section in the .egg file. That makes the bundle be installed under PREFIX/share/chicken when CHICKEN_INSTALL_PREFIX is *unset*. When CHICKEN_INSTALL_PREFIX *is set* (that's what salmonella does), the srfi-29 bundle is installed under PREFIX/share (no chicken directory). Then srfi-29 looks for bundles under include-path, which resolves to C_INSTALL_SHARE_HOME when CHICKEN_INCLUDE_PATH is *unset*. In that case, PREFIX/share/$(PROGRAM_PREFIX)chicken$(PROGRAM_SUFFIX). However when CHICKEN_INCLUDE_PATH *is set*, chicken-install installs the bundle under PREFIX/share, but in runtime it looks for bundles under include-path, which is the value of CHICKEN_INCLUDE_PATH, which is set by salmonella as PREFIX/share/chicken. hm... Now I don't know if salmonella is setting CHICKEN_INCLUDE_PATH wrongly of if the default values given to override-prefix in egg-compile.scm is wrong. possible The default values given to override-prefix definitely don't match the values of C_INSTALL_INCLUDE_HOME and C_INSTALL_SHARE_HOME. originally it only looked in (chicken-home) I think that was the same as what now (include-path) is, if I'm not mistaken. well, one is a set I _think_ (don't quote me on that) the root cause of the issue is the inconsistency between the values of C_INSTALL_INCLUDE_HOME/C_INSTALL_SHARE_HOME and the way that override-prefix assembles paths when when CHICKEN_INSTALL_PREFIX is set. is something is chicken.platform returning PREFIX/share? i can code the envvar test & subsequent path assumption but rather not. That's part of the problem I'm implying. You can get PREFIX/share *if* CHICKEN_INCLUDE_PATH is set to PREFIX/share. If CHICKEN_INCLUDE_PATH is *not set* you'll get PREFIX/share/chicken (or, more accurately, PREFIX/share/$(PROGRAM_PREFIX)chicken$(PROGRAM_SUFFIX)). This whole thing is very confusing. there is not a '(data-path)' so i will assume if CHICKEN_INSTALL_PREFIX is set then push ""...PREFIX/share"" onto the search list I think it would be better to wait a bit. I _think_ srfi-29 is probably doing the right thing. The issue _might_ be in salmonella or in CHICKEN. ok looks like a job for chicken.platform#data-repository or similar }}} " Mario Domenech Goulart Milestone someday 1849 http-client: with-input-from-request fails when uri-or-request is an intarweb request easy extensions 5.3.0 defect new 2025-01-20T10:56:22+01:00 2025-01-20T13:32:37+01:00 "{{{ (import (chicken io) (intarweb) (http-client)) (let ((req (make-request uri: ""https://google.com""))) (with-input-from-request req #f read-string)) }}} {{{ Error: (call-with-input-request) The first argument must be either an uri-common object, an intarweb request object, or an URI string # #f # Call history: (##core#begin (with-input-from-request req #f read-string)) simple-req.scm:3 (with-input-from-request req #f read-string) simple-req.scm:2 (make-request #:uri ""https://google.com"") simple-req.scm:2 (make-request #:uri ""https://google.com"") simple-req.scm:3 (with-input-from-request req #f read-string) http-client.scm:903: call-with-input-request* http-client.scm:803: uri-common#uri-reference? http-client.scm:805: intarweb#request? http-client.scm:805: intarweb#request-uri http-client.scm:807: uri-common#uri? http-client.scm:810: uri-common#uri-reference? http-client.scm:808: http-client-error http-client.scm:344: chicken.condition#make-property-condition http-client.scm:345: chicken.condition#make-property-condition http-client.scm:343: chicken.condition#make-composite-condition http-client.scm:343: srfi-18#raise <-- }}} Here, we take the request-uri out of the intarweb request. That results in a string. Two lines below, we expect it to be an uri and we fail. https://code.call-cc.org/cgi-bin/gitweb.cgi?p=eggs-5-latest.git;a=blob;f=http-client/1.2.2/http-client.scm;h=58d045d661cdae7924afe4983f24dfb822d49dc3;hb=HEAD#l805 " Pietro Cerutti Milestone someday 1855 Egg spec's version format is stricter than release-info's unknown 5.4.0 defect new 2025-08-14T14:52:45+02:00 2025-08-14T14:52:45+02:00 "= Problem = `chicken-install` strictly validates the `version` property of egg specs (i.e. what goes into `.egg` files) [https://wiki.call-cc.org/man/5/Egg%20specification%20format#version according to the documented format]: Specifies version string for this egg. `STRING` should have the format `..`, where only the `` part is mandatory. However, the same validation does not apply to eggs released through the central egg repository. Here, the version specified in the `.release-info` file is taken verbatim. This means it's not possible to install an egg from a local source directory when its version doesn't match the stricter egg spec format. = Proposed solution = Since there are already a lot of egg releases in existence which don't match the egg spec (e.g. `fmt` version `0.8.11.3`), enforcing the same strict format for `.release-info` appears too disruptive a change. Instead, we should make the egg spec's format less strict, effectively allowing any string as version. = CHICKEN 6 = For C6, we could still consider enforcing the strict format everywhere. = Reproduction = {{{ svn checkout https://anonymous@code.call-cc.org/svn/chicken-eggs/release/5/fmt/tags/0.8.11.3 cd 0.8.11.3 csi -e ""(write (cons '(version ""0.8.11.3"") (read)))"" < fmt.egg > fmt.egg.new mv fmt.egg.new fmt.egg chicken-install }}} Output: {{{ Error: egg information item has invalid structure: (version 0.8.11.3) }}}" Moritz Heidkamp Milestone someday 1859 Type information doesn't survive calls to define-inline procedures medium scrutinizer 6.0.0 defect new 2025-08-20T08:48:57+02:00 2025-08-20T09:03:21+02:00 "Compile the following to C: {{{#!scheme (module foo () (import scheme (chicken base) (chicken plist)) (let ((x (gensym))) (put! 'foo x 1)) (define-inline (put-one! sym prop) (put! sym prop 1)) (let ((y (gensym))) (put-one! 'foo y))) }}} Note that the first call to put! gets rewritten to {{{C_a_i_putprop(&a,3,lf[1],t1,C_fix(1))}}}, but the second call remains a CPS call to {{{chicken.plist#put!}}} via procedure lookup. This is unexpected, as inlining by hand should give the same results as calling a {{{define-inline}}} procedure. If you restore the typing info by hand, it also works as expected: {{{#!scheme (define-inline (put-one! sym prop) (put! (the symbol sym) (the symbol prop) 1)) }}}" sjamaan Milestone someday 1861 [R7RS] Initial environment of a CHICKEN program should be empty easy unknown 5.4.0 defect new 2025-11-29T00:14:41+01:00 2025-11-29T00:14:41+01:00 "R7RS (§5.1) requires that no identifiers be bound in the initial environment of a program. Yet there seems to be no way to prevent CHICKEN from importing all of (chicken base) into a CHICKEN program. According to the Report, `print` should be unbound in the following program: {{{ (import (scheme base)) (print ""hello, world"") }}} csc 5.4.0 compiles this (and the resulting binary runs) without an error, since `print` is in `(chicken base)`, which is automatically imported. Another result of this asymmetry is that the import qualifiers `only`, `prefix`, `rename`, etc. are effectively useless for `(chicken base)` and `(scheme)`. R7RS is right, at least for CHICKEN in r7rs mode: any identifiers used but not defined by a program should be explicitly imported. Addendum: The wiki.call-cc.org page for `(chicken base)` suggests that the `-explicit-use` compiler option can be used to avoid ""using"" the module, but Bunny351 says this option is unrelated to this issue." zaifir Milestone someday 1862 Chicken 5.4.0 doesn't build on Windows unknown 5.4.0 defect new 2025-12-22T20:29:12+01:00 2025-12-23T03:53:26+01:00 "I've been following this guide to get up and running with Chicken on Windows: https://wiki.call-cc.org/msys2#installing-chicken-50-on-msys2 When trying to compile Chicken I get compiler errors: {{{ gcc -fno-strict-aliasing -fwrapv -DHAVE_CHICKEN_CONFIG_H -DC_ENABLE_PTABLES -c -Os -DC_BUILDING_LIBCHICKEN posixwin.c -o posixwin-static.o -I. -I./ In file included from posixwin.c:9: posixwin.c: In function 'f_4799': posixwin.c:185:57: error: passing argument 2 of 'execve' from incompatible pointer type [-Wincompatible-pointer-types] 185 | #define C_u_i_execve(f,a,e) C_fix(execve(C_c_string(f), (const char *const *)C_c_pointer_ve, (const char *const *)C_c_pointer_vector_or_null(e))) | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | | | const char * const* chicken.h:1046:56: note: in definition of macro 'C_fix' 1046 | #define C_fix(n) ((C_word)((C_uword)(n) << C_FIXNUM_SHIFT) | C_FIXNUM_BIT) | ^ posixwin.c:9646:17: note: in expansion of macro 'C_u_i_execve' 9646 | t6=(C_truep(t4)?C_u_i_execve(t2,t3,t4):C_u_i_execvp(t2,t3)); | ^~~~~~~~~~~~ In file included from C:/msys64/ucrt64/include/unistd.h:11, from chicken.h:134: C:/msys64/ucrt64/include/process.h:187:64: note: expected 'char * const*' but argument is of type 'const char * const*' 187 | _CRTIMP int __cdecl execve(const char *_Filename,char *const _ArgList[],char *const _Env[]) __MINGW_ATTRIB_DEPRECATED_MSVC2005; | ~~~~~~~~~~~~^~~~~~~~~~ posixwin.c:185:109: error: passing argument 3 of 'execve' from incompatible pointer type [-Wincompatible-pointer-types] 185 | #define C_u_i_execve(f,a,e) C_fix(execve(C_c_string(f), (const char *const *)C_c_pointer_vector_or_null(a), (const char *const *)C_c_pointer_vector_or_null(e))) | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | | | const char * const* chicken.h:1046:56: note: in definition of macro 'C_fix' 1046 | #define C_fix(n) ((C_word)((C_uword)(n) << C_FIXNUM_SHIFT) | C_FIXNUM_BIT) | ^ posixwin.c:9646:17: note: in expansion of macro 'C_u_i_execve' 9646 | t6=(C_truep(t4)?C_u_i_execve(t2,t3,t4):C_u_i_execvp(t2,t3)); | ^~~~~~~~~~~~ C:/msys64/ucrt64/include/process.h:187:87: note: expected 'char * const*' but argument is of type 'const char * const*' 187 | _CRTIMP int __cdecl execve(const char *_Filename,char *const _ArgList[],char *const _Env[) __MINGW_ATTRIB_DEPRECATED_MSVC2005; | ~~~~~~~~~~~~^~~~~ posixwin.c:184:57: error: passing argument 2 of 'execvp' from incompatible pointer type [-Wincompatible-pointer-types] 184 | #define C_u_i_execvp(f,a) C_fix(execvp(C_c_string(f), (const char *const *)C_c_pointer_ve)) | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | | | const char * const* chicken.h:1046:56: note: in definition of macro 'C_fix' 1046 | #define C_fix(n) ((C_word)((C_uword)(n) << C_FIXNUM_SHIFT) | C_FIXNUM_BIT) | ^ posixwin.c:9646:40: note: in expansion of macro 'C_u_i_execvp' 9646 | t6=(C_truep(t4)?C_u_i_execve(t2,t3,t4):C_u_i_execvp(t2,t3)); | ^~~~~~~~~~~~ C:/msys64/ucrt64/include/process.h:188:64: note: expected 'char * const*' but argument is of type 'const char * const*' 188 | _CRTIMP int __cdecl execvp(const char *_Filename,char *const _ArgList[]) __MINGW_ATTRIB_DEPRECATED_MSVC2005; | ~~~~~~~~~~~~^~~~~~~~~~ posixwin.c: In function 'f_4882': posixwin.c:189:75: error: passing argument 3 of 'spawnvpe' from incompatible pointer type [-Wincompatible-pointer-types] 189 | #define C_u_i_spawnvpe(m,f,a,e) C_fix(spawnvpe(C_unfix(m), C_c_string(f), (const char *cons, (const char *const *)C_c_pointer_vector_or_null(e))) | ^~~~~~~~~~~~~~~~~ | | | const char * const* chicken.h:1046:56: note: in definition of macro 'C_fix' 1046 | #define C_fix(n) ((C_word)((C_uword)(n) << C_FIXNUM_SHIFT) | C_FIXNUM_BIT) | ^ posixwin.c:9768:17: note: in expansion of macro 'C_u_i_spawnvpe' 9768 | t6=(C_truep(t4)?C_u_i_spawnvpe(((C_word*)t0)[2],t2,t3,t4):C_u_i_spawnvp(((C_word*)t0)[2],t2,t3)); | ^~~~~~~~~~~~~~ C:/msys64/ucrt64/include/process.h:199:75: note: expected 'char * const*' but argument is of type 'const char * const*' 199 | _CRTIMP intptr_t __cdecl spawnvpe(int,const char *_Filename,char *const _ArgList[],char *const _Env[]) __MINGW_ATTRIB_DEPRECATED_MSVC2005; | ~~~~~~~~~~~~^~~~~~~~~~ posixwin.c:189:127: error: passing argument 4 of 'spawnvpe' from incompatible pointer type [-Wincompatible-pointer-types] 189 | #define C_u_i_spawnvpe(m,f,a,e) C_fix(spawnvpe(C_unfix(m), C_c_string(f), (const char *const *)C_c_pointer_vector_or_null(a), (const char *const *)C_c_pointer_vector_or_null(e))) | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | | | const char * const* chicken.h:1046:56: note: in definition of macro 'C_fix' 1046 | #define C_fix(n) ((C_word)((C_uword)(n) << C_FIXNUM_SHIFT) | C_FIXNUM_BIT) | ^ posixwin.c:9768:17: note: in expansion of macro 'C_u_i_spawnvpe' 9768 | t6=(C_truep(t4)?C_u_i_spawnvpe(((C_word*)t0)[2],t2,t3,t4):C_u_i_spawnvp(((C_word*)t0)[2],t2,t3)); | ^~~~~~~~~~~~~~ C:/msys64/ucrt64/include/process.h:199:98: note: expected 'char * const*' but argument is of type 'const char * const*' 199 | _CRTIMP intptr_t __cdecl spawnvpe(int,const char *_Filename,char *const _ArgList[],char *) __MINGW_ATTRIB_DEPRECATED_MSVC2005; | ~~~~~~ posixwin.c:188:74: error: passing argument 3 of 'spawnvp' from incompatible pointer type [-Wincompatible-pointer-types] 188 | #define C_u_i_spawnvp(m,f,a) C_fix(spawnvp(C_unfix(m), C_c_string(f), (const char *const)) | ^~~~~~~~~~~~~~~~~~ | | | const char * const chicken.h:1046:56: note: in definition of macro 'C_fix' 1046 | #define C_fix(n) ((C_word)((C_uword)(n) << C_FIXNUM_SHIFT) | C_FIXNUM_BIT) | ^ posixwin.c:9768:59: note: in expansion of macro 'C_u_i_spawnvp' 9768 | t6=(C_truep(t4)?C_u_i_spawnvpe(((C_word*)t0)[2],t2,t3,t4):C_u_i_spawnvp(((C_word*)t0)[2],t2,t3)); | ^~~~~~~~~~~~~ C:/msys64/ucrt64/include/process.h:198:74: note: expected 'char * const*' but argument is of type 'const char * const*' 198 | _CRTIMP intptr_t __cdecl spawnvp(int,const char *_Filename,char *const _ArgList[]) __MINGW_ATTRIB_DEPRECATED_MSVC2005; | ~~~~~~~~~~~~^~~~~~~~~~ make: *** [rules.make:121: posixwin-static.o] Error 1 }}}" MickeyKnox Milestone someday 1867 Cannot display literal values over 2^30 but within fixnum 64-bit range when declaring fixnum-arithmetic unknown 5.4.0 defect new 2026-03-15T17:57:25+01:00 2026-03-19T17:49:11+01:00 "Hi! I encountered a problem using the Chicken Scheme Compiler that I have trouble understanding. When I use `(declare (fixnum-arithmetic))`, literal values that are `(expt 2 30)` or over but within the fixnum 64-bit range cause the following error: {{{ Error: cannot coerce inexact literal `2166136261' to fixnum }}} What is stranger is that printing the `most-positive-fixnum` value from `(import (chicken fixnum))` works fine and is a value way larger than 2^30. I included the various tests I performed. Uncomment the relevant parts for one particular test, the results I got are included after each test: {{{ ;File: test-chicken.scm (declare (fixnum-arithmetic)) ;; always there for all tests ;(import (chicken fixnum)) ;(display most-positive-fixnum) ;;; works fine, prints 4611686018427387903 ;(display (exact? 1073741824)) ;;; works fine, prints #t ;(display 1073741824) ;;; craches with Error: cannot coerce inexact literal `1073741824' to fixnum ;(display 1073741824) ;;; works fine, prints 1073741824 }}} I compiled and ran the tests using the following command: `csc test-chicken.scm -o test-chick && ./test-chick` For context, I found this bug working on Ribbit, another scheme implementation because of the inclusion of the literal for a hash function: https://github.com/udem-dlteam/ribbit/blob/7666b6c017a73a0a82e8df7f9de119d2408da404/src/rsc.scm#L1035 Here are infos about my `csc` version: {{{ > csc -version CHICKEN (c) 2008-2022, The CHICKEN Team (c) 2000-2007, Felix L. Winkelmann Version 5.4.0 (rev 1a1d1495) macosx-unix-clang-arm64 [ 64bit dload ptables ] }}}" leo-ard Milestone someday 1868 vector->string adds \x00 to the output easy core libraries 6.0.0 defect new 2026-03-21T02:01:04+01:00 2026-03-21T02:01:04+01:00 "vector->string adds NULL bytes to the output in increasing frequency: {{{#!scheme (vector->string '#(#\a #\b #\c #\d #\e)) ;;=> ""ab\x00;c\x00;\x00;\x00;d\x00;\x00;\x00;\x00;\x00;\x00;\x00;e\x00;\x00;\x00;\x00;\x00;\x00;\x13;\x00;"" }}} Even worse, this: {{{#!scheme (vector->string (string->vector (vector->string '#(#\a #\b #\c #\d #\e)))) }}} results in a segmentation violation error." zilti Milestone someday 1869 vector->string adds \x00 to the output easy core libraries 6.0.0 defect new 2026-03-21T02:01:53+01:00 2026-03-21T02:01:53+01:00 "vector->string adds NULL bytes to the output in increasing frequency: {{{#!scheme (vector->string '#(#\a #\b #\c #\d #\e)) ;;=> ""ab\x00;c\x00;\x00;\x00;d\x00;\x00;\x00;\x00;\x00;\x00;\x00;e\x00;\x00;\x00;\x00;\x00;\x00;\x13;\x00;"" }}} Even worse, this: {{{#!scheme (vector->string (string->vector (vector->string '#(#\a #\b #\c #\d #\e)))) }}} results in a segmentation violation error." zilti Milestone someday 1870 vector->string adds \x00 to the output easy core libraries 6.0.0 defect new 2026-03-21T02:05:53+01:00 2026-03-21T02:05:53+01:00 "vector->string adds NULL bytes to the output in increasing frequency: {{{#!scheme (vector->string '#(#\a #\b #\c #\d #\e)) ;;=> ""ab\x00;c\x00;\x00;\x00;d\x00;\x00;\x00;\x00;\x00;\x00;\x00;e\x00;\x00;\x00;\x00;\x00;\x00;\x13;\x00;"" }}} Even worse, this: {{{#!scheme (vector->string (string->vector (vector->string '#(#\a #\b #\c #\d #\e)))) }}} results in a segmentation violation error." zilti Milestone someday 1222 Rework error handling to make it work better with non-POSIX functions medium core libraries enhancement new 2015-10-31T15:50:36+01:00 2023-11-08T21:21:28+01:00 "The core system is extremely ANSI/POSIX-centered. This means for example, that {{{##sys#update-errno}}} and strerror are used all over the place. Unfortunately, this makes handling errors unnecessarily hard when dealing with native Windows functions (and potentially similar on other platforms). Michele La Monaca [gave an example of how to deal with this http://lists.gnu.org/archive/html/chicken-hackers/2014-10/msg00072.html] in a thread where he pointed out that {{{rename-file}}} is inconsistent across platforms when the target file already exists." sjamaan Milestone someday 1259 slow polling insane core libraries 4.10.x enhancement new 2016-02-19T18:09:42+01:00 2016-08-25T20:32:57+02:00 Try integrating the modified scheduler from http://ball.askemos.org joergw Milestone someday 1553 sqlite3: The PREPARE procedure does not make it clear when it has and has not parsed a statement from a string. easy extensions 4.13.0 enhancement new 2018-10-21T13:10:01+02:00 2018-10-21T13:16:01+02:00 "The PREPARE procedure does not make it clear when it has and has not parsed a statement from a string. The calls `(prepare db """")` and `(prepare db ""select 1;""))` (note no trailing whitespace) both produce a `sqlite3:statement` and `""""`, and there is no way of distinguishing between these. In the former case, the statement displays as `#`, but there is no way of detecting this in code. The documentation for PREPARE seems to suggest that a way of parsing a sequence of SQL statements from a string is to call PREPARE recursively on PREPARE's `rest` return, until `rest` returns `""""`. Thus: {{{ (define (execute-statements* sql-string) (let-values (((stmt rest) (prepare db sql-string))) (if (string-null? rest) (finalize! stmt) (begin (dynamic-wind values (lambda () (update stmt)) (lambda () (finalize! stmt))) (execute-statements* rest))))) }}} This will work //if and only if// the SQL string ends in whitespace, but will not work if the input string has no trailing whitespace or if it ends in a comment. Spotting when `sql-string` is all whitespace, and special-casing this, helps in many cases, but doesn't help in the case where `sql-string` contains, or has through this process ended up containing, only a comment. There //might// be a way of wrapping this up which solves the problem, and there might be a way of deducing that a statement is a 'zombie' by (mis)using other procedures, but more straightforward would be one of the following: * Create a STATEMENT-ZOMBIE? predicate for the case where (in the implementation) `statement-ptr` is `#f`, and a statement would display as `zombie`. Advantages: straightforward to implement; quite specific. Disadvantages: introduces an extra concept which isn't really present in the underlying SQLite library, and which appears to be meaningful only for this case. * Have STATEMENT? return `#f` for 'zombie' statements. Advantages: matches the real semantics of the situation; so is probably the least surprising thing. Disadvantages: might in principle confuse some future type reasoning, which might assume that if something is a statement, then it can be passed to UPDATE without error. * Have PREPARE return `#f` for the statement when no statement is parsed. Advantages: also unsurprising, and easy to check for without changing the definition of STATEMENT?. Disadvantates: some complication to the type signature of PREPARE. Of the three, I think the first is unattractive, and the second slightly more attractive than the third. I've marked this as an enhancement request, though I think it's on the border of being a bug. I don't know what the criteria are for major vs minor priority. " Norman Gray Milestone someday 1667 Overhaul c-backend.scm hard compiler 5.1.0 enhancement felix winkelmann assigned 2020-01-24T21:42:43+01:00 2023-11-06T22:09:50+01:00 "The C code generation stuff is a terrible, rotten mess. It's incomprehensible and making changes in the code has become rather difficult. I have a ""backend"" branch that uses an intermediate representation (S-expressions, roughly C-like), originally intended for an alternative backend (""qbe"") that seems to work, at least for the core system + tests. I can try to bring it up to date and present it once it is ready. " felix winkelmann Milestone someday 1796 Allow type definitions to be exported. core libraries 5.3.0 enhancement new 2022-01-04T20:07:21+01:00 2022-01-04T20:07:21+01:00 "Currently, there is no way for a module to export type definitions for public use. (i.e. if a module M defines a type T, there is no way for programs that import M to make use of T; they must use * instead). This severely limits the usefulness of the (chicken type) system, since type checking cannot be used on types outside of the CHICKEN core. I'd propose an `export`-like form, e.g. `(export-type TYPE ...)` which makes type definitions created with `define-type` available beyond the scope of the containing module." zaifir Milestone someday 1807 with-input-from-string scsh-process extensions 5.3.0 enhancement new 2022-07-11T17:09:39+02:00 2022-07-11T19:48:56+02:00 "With-input-from-string doesn't work with stdports: Error: (port->fileno) port has no attached file - Success: # With-input-from-string works fine with other stuff like read-lines and stdports work fine with with-input-from-file." Idiomdrottning Milestone someday 1813 Test w64devkit easy build system 5.3.0 task new 2022-10-31T12:27:53+01:00 2022-10-31T12:27:53+01:00 "This may be the canonical windows toolchain that we are looking for: https://github.com/skeeto/w64devkit " felix winkelmann Milestone someday 1763 chicken-install should pass features defined on the command line on to the build script easy core tools 5.2.0 change request new 2021-06-08T11:42:58+02:00 2021-06-08T11:42:58+02:00 "Currently, features defined using the `-feature` option of `chicken-install` are only visible in `cond-expand` forms in the egg, not in code compiled via `csc` in the subsequent build steps. It would probably be more intuitive and convenient to add `-feature FEATURE` options to the build scripts. " felix winkelmann Milestone someday 749 honu: handle octal format escapes in string literals extensions 4.7.x defect new 2011-12-13T13:35:40+01:00 2023-11-06T22:00:46+01:00 Currently not done but would be needed to handle all kinds of C strings. felix winkelmann Milestone someday 784 ssax's `ssax:warn' ignores its first argument (port) extensions 4.7.x defect new 2012-02-09T12:11:30+01:00 2023-11-06T22:01:02+01:00 "From ssax-chicken.scm: {{{ (define (ssax:warn port msg . other-msg) (apply cerr (cons* nl ""Warning: "" msg other-msg))) }}} {{{port}}} doesn't seem to be used anywhere." Mario Domenech Goulart Milestone someday 934 improve documentation for functors and named interfaces easy documentation 4.8.x defect new 2012-10-18T16:03:26+02:00 2016-08-25T21:21:22+02:00 "The documentation for functors and interfaces (as defined by `define-interface`) needs to be expanded and improved, and examples added. (Suggested by Matt Gushee) " felix winkelmann Milestone someday 985 read-line blocks intermittantly on Atom based netbook hard unknown 4.8.x defect Matt Welland assigned 2013-02-18T01:03:10+01:00 2016-08-25T21:25:43+02:00 "This problem first shows up in: 9eef92115fba6034a98cc21ef740fd82ea52387b is the first bad commit commit 9eef92115fba6034a98cc21ef740fd82ea52387b Author: Peter Bex Date: Sun Nov 18 21:03:51 2012 +0100 Fix select() buffer overrun vulnerability by using POSIX poll() on systems that support it, leaving only those few systems vulnerable that don't (ie, only Windows). I used the following code snippet to test. It took as much as 10 runs to trigger the bug sometimes: =============CODE============= (use posix) (define (conservative-read port) (let loop ((res '())) (if (not (eof-object? (peek-char port))) (loop (cons (read-char port) res)) (apply conc (reverse res))))) (define (cmd-run-with-stderr->list cmd . params) (let-values (((fh fho pid fhe) (if (null? params) (process* cmd) (process* cmd params)))) (let loop ((curr (read-line fh)) (result '())) (print ""GOT HERE 1, curr="" curr) (let ((errstr (conservative-read fhe))) (print ""GOT HERE 3, errstr="" errstr) (if (not (string=? errstr """")) (set! result (cons errstr result)))) (print ""GOT HERE 2, result="" result) (if (not (eof-object? curr)) (begin (print ""GOT HERE 4"") (loop (read-line fh) (cons curr result))) (begin (print ""GOT HERE 5"") (close-input-port fh) (close-input-port fhe) (close-output-port fho) (reverse result)))))) (print ""Got: "" (cmd-run-with-stderr->list ""ls"")) " Matt Welland Milestone someday 1007 socket egg: socket-close should track closed status unknown defect Jim Ursetto new 2013-04-12T08:59:39+02:00 2013-04-12T08:59:39+02:00 "We should probably track socket closed status, possibly by setting the fileno to #f when socket-close is called. This should prevent accidental reuse of the socket FD when it is reopened by someone else. It's not clear whether socket-close should stop throwing an error on double-close, or if this should be done instead in socket-close*. This would be appropriate behavior to use in a finalizer. But socket-close* is targeted for (immediate) use after network error. Also note there are a bunch of direct calls to _close_socket which might have to be corrected to track the closed state." Jim Ursetto Milestone someday 1148 Fix setup-helper usage in eggs trivial extensions 4.9.x defect new 2014-08-14T17:25:35+02:00 2016-08-25T21:45:26+02:00 "Since version 1.5.3, the setup-helper egg uses the module system as the deployment method, before that it was to be (include)d by the users into their .setup files. A lot of eggs still use the include statement to use setup-helper and that breaks it when a custom CHICKEN_REPOSITORY is set. To fix this, all eggs should use the new import method. The fix is very simple, just replace `(include ""setup-helper"")` by (use setup-helper-mod) wherever necessary. Here is a (automaticaly generated) list of possibly affected eggs: 9ML-toolkit AD F-operator R aima alist-lib amb animation apropos blob-utils bloom-filter box call-with-environment-variables call-with-query call-with-sqlite3-connection check-errors cock cock-utils coerce combinatorics combinators condition-utils coops-utils crc csp csv-xml debug define-record-and-printer define-structure directory-utils dollar dsssl-utils error-utils expand-full ffmpeg-video freetds graphviz hashes heap image-processing kvlists levenshtein linear-algebra list-utils locale lookup-table macosx mailbox mathh message-digest moremacros mw mw-core neo4j nondeterminism number-limits posix-utils procedure-decoration qobischeme-ui redis remote-mailbox ripemd scheme2c-compatibility setup-helper sha1 sha2 sicp simple-units srfi-102 srfi-19 srfi-27 srfi-29 srfi-40 srfi-41 srfi-45 srfi-60 srfi-9-ext srfi-95 stack string-utils sxml-templates symbol-utils synch thread-utils tiger-hash timed-resource token-substitution traversal twilio uuid-lib uuid-ossp variable-item yasos " Kooda Milestone someday 1233 hexgrid: test failure easy extensions 4.9.x defect estevo new 2015-12-13T17:43:01+01:00 2016-08-25T21:46:38+02:00 See http://salmonella-linux-x86-64.call-cc.org/master-debugbuild/gcc/linux/x86-64/2015/12/13/salmonella-report/test/hexgrid.html Mario Domenech Goulart Milestone someday 1236 equal? can break at random moments hard core libraries 4.10.x defect new 2015-12-22T15:37:34+01:00 2021-04-09T11:56:03+02:00 "This is really bad. C_equal is defined to run inline, so it can't clear up stack space, but it consumes stack space! It will just bail out with an exception when it runs out of stack. This is really really bad, because it basically means that you're playing russian roulette every time you call {{{equal?}}}; it can break or not, which depends on the state of the stack, over which you have very little control. It is also very hard to predict when exactly it will fail. I found this out [[http://salmonella-linux-x86-64.call-cc.org/master-debugbuild/gcc/linux/x86-64/2015/12/22/salmonella-report/test/numbers.html|the hard way]]. Here's a standalone version that boils down to the code in the numbers tests: {{{ #!scm ;; Creates lists like (((1))) and ((((1)))) for 3 and 4, respectively. (define-syntax recompose (er-macro-transformer (lambda (e r c) (let ((n (cadr e)) (op (caddr e)) (arg (cadddr e))) (define (recompose-1 n) (if (= n 1) `(,op ,arg) `(,op ,(recompose-1 (- n 1))))) (recompose-1 n))))) (let lp ((i 0)) (print i "" "" (equal? (recompose 1000 list '(1)) (recompose 1000 list (list 1)))) (lp (add1 i))) }}} You'll see this blows up after only 16 or so iterations. I had to up the list nesting to 1000 because 100 only seems to crash once every blue moon. A fix could involve dynamically allocating (with {{{malloc}}} or so) a ""working list"" stack and use that to convert the recursion to a loop. Alternatively, we could ""de-inline"" it so it can reclaim memory. On the plus side, this allows us to convert it to Scheme. On the downside, we'd be taking a CPS hit. Originally I thought it might be possible to use the scratch space in CHICKEN 5, but I'm not sure that's suitable in this case because there's nothing in the heap or nursery that points directly to it. We could create one and make an actual Scheme list of ""todo"" objects, but it would be such a long chain that it would become pretty slow I think." sjamaan Milestone someday 1276 alist-lib alist-update!/default not useful for adding keys extensions 4.10.x defect Peter Danenberg assigned 2016-04-08T00:07:39+02:00 2017-08-25T15:20:03+02:00 I noticed that alist-update! and alist-update!/default in alist-lib have call signatures suggestive of being able to add keys to an alist (not just updating existing keys) because they accept a default OLD value if the given key is not found. However, both procedures have an unspecified return value, so it's not clear what the intent was for the case of adding new entries to the alist. If they don't return an updated reference to the head of the list, there is no way to add new entries, and it doesn't make sense to have a DEFAULT in the call signature. John Foerch Milestone someday 1339 char-ready? returns #f even when input is available insane core libraries 4.11.0 defect new 2017-01-11T09:41:53+01:00 2019-08-25T17:47:08+02:00 "The current implementation of char-ready? on stdio files returns #f if you have buffering enabled, didn’t consume the whole buffer and no more input is available from the system’s point of view, as it only asks the system via select or poll (see C_char_ready_p and C_check_fd_ready). I’m afraid there’s no way to do better than this with stdio files, as there is no way of checking the state of the buffer. The workaround I found was to simply disable buffering altogether, like so: {{{ (set-buffering-mode! (current-input-port) #:none) }}} This bug probably requires a pretty big rework of ports. Here is a little program that '''might''' show the erroneous behaviour: {{{ (use posix) (call-with-input-pipe ""echo coucou"" (lambda (p) (let loop ((c (read-char p))) (unless (eof-object? c) (print c) (when (char-ready? p) (loop (read-char p))))))) }}}" Kooda Milestone someday 1340 Unable to compile a mathh gama instruction at or above optimize-level 3 medium unknown 4.11.0 defect new 2017-01-16T02:02:34+01:00 2021-04-12T13:40:17+02:00 "Compiling the following code at or above optimize-level 3 results in an error. {{{(use mathh) (define (factorial x) (gamma (+ 1 x)) ) (factorial 9)}}} csc -optimize-level 3 test.scm test.c:213:4: warning: implicit declaration of function 'stub177' is invalid in C99 [-Wimplicit-function-declaration] t6=stub177(t4,t5); ^ 1 warning generated." rca Milestone someday 1371 locative table is only grown, never shrunk medium core libraries 4.12.0 defect new 2017-05-12T09:39:49+02:00 2023-11-06T22:16:11+01:00 "From a quick look at the code, it looks like we only realloc the `locative_table` when making a locative and the table is full. In extreme cases, this means the locative table will grow and never shrink back to a manageable size. After some consideration, I think this is not a huge issue, as on every GC, collected locatives are thrown out of the table, so the maximum size of the locative table is bounded by the heap size. Still, this is not desirable. I think the best place to do this is at the end of `update_locative_table` since we're already keeping track of the table size. We may also change this function to compact the table while walking it, so we can get by with a simple `realloc()`." sjamaan Milestone someday 1382 sql-de-lite ignores new versions of sqlite3 on Cygwin trivial extensions 4.12.0 defect Jim Ursetto new 2017-06-27T17:15:31+02:00 2017-06-28T13:00:37+02:00 "The `version-check` program does not run under Cygwin, so `sql-de-lite` installation conservatively assumes it should use the bundled 3.0.17 version of sqlite3 instead of Cygwin's current 3.0.18. The root cause of this is that Cygwin ''really isn't Windows'' and shouldn't set the `windows:` feature (see #1381). As a workaround, change line 49 of the setup file with `s/windows/(and windows (not cygwin))/`. A more portable resolution is to get rid of `version-check` altogether and run `sqlite3 --version` instead: the first few characters of the output (up to a space) are the formatted sqlite3 version number." johnwcowan Milestone someday 1427 gochan's test failure extensions 4.12.0 defect kristianlm new 2017-11-27T08:30:29+01:00 2017-11-27T08:30:29+01:00 "I noticed a failure in gochan's tests today: http://salmonella-linux-x86.call-cc.org/master-debugbuild/gcc/linux/x86/2017/11/26/salmonella-report/test/gochan.html {{{ -- testing timers: each gochan-tick gets consumed by only one recv ----------- 10ms messages for 105ms means 10 messages ............................ [ FAIL] expected 10 but got 12 (length results) hopefully different senders: (1 1 1 1 1 2 3 3 4 4 4 4) 1 test completed in 0.124 seconds. 1 failure (100%). 0 out of 1 (0%) tests passed. -- done testing timers: each gochan-tick gets consumed by only one recv ------ }}} I've never seen this one before. I'm reporting in case it can help to spot a difficult to reproduce issue." Mario Domenech Goulart Milestone someday 1452 [test egg]: test-group eats the call-chain on error extensions defect new 2018-03-23T19:28:26+01:00 2018-03-23T19:28:26+01:00 "Below is code that calls an ill behaving function `foo` three times. Only the third version prints the call-chain. It should be printed in the two calls inside `test-group`s as well. {{{ (use test) (define (f) (+ 1 'a)) (test-group ""foo"" (f)) (test-group ""bar"" (test 2 (f))) (test-begin ""baz"") (f) (test-end ""baz"") ;; $ csi -qbn test-group-call-chains.scm ;; ;; Warning: error in group outside of tests ;; ;; Error: (+) bad argument type: a ;; 1 test completed in 0.0 seconds. ;; 1 error (100%). ;; 0 out of 1 (0%) tests passed. ;; -- done testing foo ---------------------------------------------------------- ;; ;; ;; -- testing bar --------------------------------------------------------------- ;; (f) .................................................................. [ERROR] ;; ;; Error: (+) bad argument type: a ;; 1 test completed in 0.0 seconds. ;; 1 error (100%). ;; 0 out of 1 (0%) tests passed. ;; -- done testing bar ---------------------------------------------------------- ;; ;; ;; Error: (+) bad argument type: a ;; ;; Call history: ;; ;; (with-exception-handler89 (##core#lambda (exvar79) (k86 (##core#lambda () (##core#let ((kvar80 (and8... ;; (##sys#call-with-values (##core#lambda () (begin67 (test 2 (f)))) (##core#lambda args87 (k86 (##core... ;; (test-run102 (lambda103 () 2) (lambda103 () (f)) (cons104 (cons104 (quote105 name106) #f) (quote105 ... ;; (cons104 (cons104 (quote105 name106) #f) (quote105 ((source107 f) (source101 f)))) ;; (cons104 (quote105 name106) #f) ;; (f) ;; [f] (+ 1 (quote a)) ;; (k86 (##core#lambda () (##sys#apply ##sys#values args87))) ;; (##sys#apply ##sys#values args87) ;; (test-end76 name56) ;; (current-test-group58 old-group57) ;; (test-begin ""baz"") ;; (test-begin ""baz"") ;; (f) ;; (f) ;; [f] (+ 1 (quote a)) <-- }}}" megane Milestone someday 1491 ##sys#expand-multiple-values-assignment works for some reason medium expander 5.0.0 defect new 2018-07-23T09:39:26+02:00 2018-07-23T14:06:15+02:00 "Replace the original with this. I have only added the `print` and the `import`. {{{ (import chicken.type) (define (##sys#expand-multiple-values-assignment formals expr) (##sys#decompose-lambda-list formals (lambda (vars argc rest) (let ((aliases (if (symbol? formals) '() (map gensym formals))) (rest-alias (if (not rest) '() (gensym rest)))) (print ""formals: "" formals "" list?: ""(list? (the * formals)) "" aliases: "" aliases "" rest "" rest-alias) `(##sys#call-with-values (##core#lambda () ,expr) (##core#lambda ,(append aliases rest-alias) ,@(map (lambda (v a) `(##core#set! ,v ,a)) vars aliases) ,@(cond ((null? formals) '((##core#undefined))) ((null? rest-alias) '()) (else `((##core#set! ,rest ,rest-alias)))))))))) }}} Now check the output of this. Pay attention to the `(d . e)` case. It's not a list, but somehow the `(map gensym formals)` expression runs without errors. {{{ (let ([e '(0 1 2 3 (4) (5 6))] [r (letrec-values ((() (values)) ((a) (values 0)) ((b c) (values 1 2)) ((d . e) (values 3 4)) (f (values 5 6))) (list a b c d e f))]) (print e) (print r) (assert (equal? e r))) --> $ csc ../letrecvalues.scm && ../letrecvalues formals: () list?: #t aliases: () rest () formals: (a) list?: #t aliases: (a11) rest () formals: (b c) list?: #t aliases: (b12 c13) rest () formals: (d . e) list?: #f aliases: (d14) rest e15 formals: f list?: #f aliases: () rest f16 (0 1 2 3 (4) (5 6)) (0 1 2 3 (4) (5 6)) }}} I think the correct implementation is something more like this: {{{ (define (##sys#expand-multiple-values-assignment formals expr) (##sys#decompose-lambda-list formals (lambda (vars argc rest) (let* ((aliases (if (symbol? formals) '() (map gensym vars))) (rest-alias (and rest (gensym rest)))) `(##sys#call-with-values (##core#lambda () ,expr) (##core#lambda ,(cond [(symbol? formals) rest-alias] [(and rest (null? (cdr vars))) rest-alias] [(null? aliases) aliases] [rest (let lp ([vs aliases]) (if (null? (cdr vs)) (car vs) (cons (car vs) (lp (cdr vs)))))] [else aliases]) ,@(map (lambda (v a) `(##core#set! ,v ,a)) vars aliases) ,@(cond [(symbol? formals) `((##core#set! ,rest ,rest-alias))] [else '((begin))]))))))) }}}" megane Milestone someday 1526 gochan: test failure: 10ms messages for 105ms means 10 messages: expected 10 but got 11 extensions defect kristianlm new 2018-08-30T07:42:06+02:00 2018-11-09T15:19:50+01:00 "See http://salmonella-linux-x86.call-cc.org/master-debugbuild/gcc/linux/x86/2018/08/30/salmonella-report/test/gochan.html Doesn't seem to happen in all runs. Tests in the previous days passed." Mario Domenech Goulart Milestone someday 1541 Use reentrant library functions, if available medium core libraries 5.0.0 defect new 2018-10-01T10:54:12+02:00 2018-10-01T10:54:12+02:00 Use reentrant C library functions, if possible (like `strerror_r`), where available. Figuring out what is available where is the hard part. felix winkelmann Milestone someday 1582 List of available eggs in RSS format contain incomplete URLs easy wiki 5.0.0 defect new 2019-02-02T15:34:33+01:00 2019-02-02T15:34:33+01:00 The link elements in http://eggs.call-cc.org/rss-4.xml and http://eggs.call-cc.org/rss-5.xml are broken. A https: prefix is missing. sven Milestone someday 1621 apropos support disjoint keyword easy extensions 5.0.0 defect Kon Lovett new 2019-06-09T20:10:43+02:00 2021-09-09T01:46:12+02:00 need to filter keyword table for apropos Kon Lovett Milestone someday 1628 hopefully: deadlock in tests extensions 5.1.0 defect joergw new 2019-07-02T07:35:56+02:00 2019-07-22T17:34:18+02:00 "See http://salmonella-linux-arm64.call-cc.org/master/clang/linux/arm64/2019/07/01/salmonella-report/test/hopefully.html Note that tests usually work. That's the first time I see this failure." Mario Domenech Goulart Milestone someday 1634 copy-file and move-file do not preserve permissions easy core libraries 5.1.0 defect new 2019-07-15T08:01:56+02:00 2019-07-28T23:17:49+02:00 both procedures create a new output file using open, so permissions are always (assuming normal umask) 644. this patch sets the permissions to those of the input file after writing all bytes and before closing the ports. tested on PLATFORM=linux. I do not know the implications of this on windows; the fact that C_chmod and C_fchmod have definitions in posixwin.scm makes me assume this is safe, but I cannot be sure without a machine to test on alicemaz Milestone someday 1635 Include multiple times in a row 'interrupts' definitions hard expander 5.1.0 defect new 2019-07-24T06:28:34+02:00 2019-07-24T13:20:40+02:00 "Given file `include.scm`: {{{#!scheme (define (foo) (print that)) }}} and a file ""include2.scm"" with arbitrary contents (including an empty file), the following code works (interpreted and compiled): {{{#!scheme #!/usr/bin/csi -s (define (main #!optional args) (include ""include.scm"") (define that ""hey"") (foo)) (main) }}} whereas this does not (interpreted or compiled): {{{#!scheme #!/usr/bin/csi -s (define (main #!optional args) (include ""include.scm"") (include ""include2.scm"") (define that ""hey"") (foo)) (main) }}} The latter throws an error about `that` being unbound." Diego Milestone someday 1669 9p: wrong test directory trivial extensions 5.4.0 defect Shawn Rutledge assigned 2020-02-02T20:35:58+01:00 2024-12-23T14:37:17+01:00 9p uses a {{{test}}} directory for tests, which is not used by chicken-install (thus not used by salmonella either). The correct directory should be {{{tests}}}. Mario Domenech Goulart Milestone someday 1670 address-info: wrong test directory trivial extensions 4.13.0 defect Thomas Chust assigned 2020-02-02T20:36:58+01:00 2020-02-02T20:58:08+01:00 address-info uses a {{{test}}} directory for tests, which is not used by chicken-install (thus not used by salmonella either). The correct directory should be {{{tests}}}. Mario Domenech Goulart Milestone someday 1671 aes: wrong test directory trivial extensions 4.13.0 defect Alaric Snell-Pym assigned 2020-02-02T20:37:50+01:00 2020-02-07T20:13:25+01:00 aes uses a {{{test}}} directory for tests, which is not used by chicken-install (thus not used by salmonella either). The correct directory should be {{{tests}}}. Mario Domenech Goulart Milestone someday 1672 crypto-tools: wrong test directory trivial extensions 4.13.0 defect Alaric Snell-Pym assigned 2020-02-02T20:38:48+01:00 2020-02-02T20:58:38+01:00 crypto-tools uses a {{{test}}} directory for tests, which is not used by chicken-install (thus not used by salmonella either). The correct directory should be {{{tests}}}. Mario Domenech Goulart Milestone someday 1673 srfi-99: wrong test directory trivial extensions 4.13.0 defect Thomas Chust assigned 2020-02-02T20:39:46+01:00 2020-02-02T20:58:51+01:00 srfi-99 uses a {{{test}}} directory for tests, which is not used by chicken-install (thus not used by salmonella either). The correct directory should be {{{tests}}}. Mario Domenech Goulart Milestone someday 1674 tweetnacl: wrong test directory trivial extensions 4.13.0 defect Thomas Chust assigned 2020-02-02T20:40:57+01:00 2020-02-02T20:59:07+01:00 tweetnacl uses a {{{test}}} directory for tests, which is not used by chicken-install (thus not used by salmonella either). The correct directory should be {{{tests}}}. Mario Domenech Goulart Milestone someday 1701 Egg SRFI-37 - wrong result extensions 5.2.0 defect new 2020-07-06T23:14:26+02:00 2020-07-06T23:14:26+02:00 "Egg srfi-37 seems to have a problem: {{{ Type ,? for help. #;1> (import (srfi 37)) ; loading /home/jeronimo/pkg/scheme/chicken/ROOT-git/lib/chicken-git/11/srfi-37.import.so ... ; loading /home/jeronimo/pkg/scheme/chicken/ROOT-git/lib/chicken-git/11/srfi-37.so ... #;2> (args-fold '(""-I"" ""the-argument"") (list (option '(#\I) #f #t (lambda (opt name arg result) (and (eqv? name #\I) arg)))) (lambda args (error ""unrecognized"")) (lambda args #f) #f) #f #;4> }}} But it should have returned the argument value (""the-argument""). If, instead of '(""-I"" ""the-argument""), you pass it '(""-Ithe-argument""), then it will correctly recognize -I's argument, ""the-argument"". Several implementations seem to do the same, so perhaps this is a bug in the reference implementation? " Jeronimo Pellegrini Milestone someday 1719 unbounded growth from parameterize core libraries 5.2.0 defect new 2020-08-28T19:38:59+02:00 2023-11-08T21:18:08+01:00 "I know of no practical use case for an unbounded number of created and forgotten parameter objects, so assign the appropriate level of priority to this -- I just thought it was an interesting case. I recently realized this was possible on my own toy scheme and fixed it by switching to ephemerons in my continuation-marks cell representation. chibi, mit-scheme, chez, guile, gauche -- all have unbounded growth on this test as well. #;1> (let loop () (parameterize (((make-parameter #f) (make-vector 10000))) (loop))) [panic] out of memory - heap has reached its maximum size - execution terminated " Jim Rees Milestone someday 1726 chicken.base#implicit-exit-handler reporting segmentation violation with non-default heap size core libraries 5.1.0 defect new 2021-01-15T04:19:31+01:00 2023-11-14T18:48:38+01:00 "I was experimenting with the ""Man-or-boy"" test (http://rosettacode.org/wiki/Man_or_boy_test) using Chicken Scheme; my variant looks like this: {{{ vjo@glaucus/pts/11:chicken/man_or_boy % cat test.scm (import (chicken process-context) (chicken fixnum)) (define (A k x1 x2 x3 x4 x5) (let ([+ fx+] [- fx-] [<= fx<=]) (define (B) (set! k (- k 1)) (A k B x1 x2 x3 x4)) (if (<= k 0) (+ (x4) (x5)) (B)))) (define k (let ([args (command-line-arguments)]) (if (> (length args) 0) (string->number (car args)) 10))) (print (A k (lambda () 1) (lambda () -1) (lambda () -1) (lambda () 1) (lambda () 0))) }}} This compiles fine, and runs great with the default heap size up to a value of k=27: {{{ vjo@glaucus/pts/11:chicken/man_or_boy % csc -O4 test.scm vjo@glaucus/pts/11:chicken/man_or_boy % time ./test 26 -21051458 ./test 26 25.20s user 1.31s system 99% cpu 26.577 total vjo@glaucus/pts/11:chicken/man_or_boy % time ./test 27 -46750171 ./test 27 101.91s user 4.42s system 99% cpu 1:46.66 total }}} At k=28, however, we run out of heap, with the default settings: {{{ vjo@glaucus/pts/11:chicken/man_or_boy % time ./test 28 [panic] out of memory - heap has reached its maximum size - execution terminated ...more... test.scm:14: B test.scm:11: A test.scm:14: B test.scm:11: A test.scm:13: x4 test.scm:11: A test.scm:14: B test.scm:11: A test.scm:14: B test.scm:11: A test.scm:13: x4 test.scm:11: A test.scm:14: B test.scm:11: A test.scm:14: B test.scm:11: A <-- ./test 28 22.39s user 3.79s system 99% cpu 26.282 total }}} No problem, say I; let me re-compile with an increased heap. That works well - and the code gives back the correct answer - but, I now encounter the bug I wish to report: {{{ vjo@glaucus/pts/11:chicken/man_or_boy % csc -O4 -heap-size 4096M test.scm vjo@glaucus/pts/11:chicken/man_or_boy % time ./test 28 -103821058 Error: segmentation violation Call history: test.scm:13: x5 test.scm:11: A test.scm:13: x4 test.scm:13: x5 test.scm:13: x5 test.scm:11: A test.scm:13: x4 test.scm:11: A test.scm:13: x4 test.scm:13: x5 test.scm:13: x5 test.scm:11: A test.scm:13: x4 test.scm:13: x5 test.scm:20: chicken.base#print chicken.base#implicit-exit-handler <-- ./test 28 90.91s user 1.71s system 99% cpu 1:32.65 total }}} It *appears* that the implicit-exit-handler triggers a segmentation violation for any binary compiled with a non-default heap size specified; as an illustrative example, here is the same code compiled with two different heap sizes and executed using k=26. The first shows what happens with an insufficient heap size, the second shows what happens when the heap size is *just barely* sufficient, but non-default value: {{{ vjo@glaucus/pts/11:chicken/man_or_boy % csc -O4 -heap-size 928M test.scm vjo@glaucus/pts/11:chicken/man_or_boy % time ./test 26 [panic] out of memory - heap full - execution terminated ...more... test.scm:14: B test.scm:11: A test.scm:14: B test.scm:11: A test.scm:13: x4 test.scm:11: A test.scm:14: B test.scm:11: A test.scm:14: B test.scm:11: A test.scm:13: x4 test.scm:11: A test.scm:14: B test.scm:11: A test.scm:14: B test.scm:11: A <-- ./test 26 5.01s user 0.44s system 99% cpu 5.471 total vjo@glaucus/pts/11:chicken/man_or_boy % csc -O4 -heap-size 929M test.scm vjo@glaucus/pts/11:chicken/man_or_boy % time ./test 26 -21051458 Error: segmentation violation Call history: test.scm:13: x5 test.scm:13: x5 test.scm:11: A test.scm:13: x4 test.scm:11: A test.scm:13: x4 test.scm:11: A test.scm:13: x4 test.scm:13: x5 test.scm:13: x5 test.scm:13: x5 test.scm:11: A test.scm:13: x4 test.scm:13: x5 test.scm:20: chicken.base#print chicken.base#implicit-exit-handler <-- ./test 26 25.87s user 0.73s system 99% cpu 26.620 total }}} The versions of all software involved have been determined below: {{{ vjo@glaucus/pts/11:chicken/man_or_boy % csc -O4 -kv test.scm '/usr/bin/chicken' 'test.scm' -output-file 'test.c' -optimize-level 4 -verbose 'x86_64-linux-gnu-gcc' 'test.c' -o 'test.o' -c -fno-strict-aliasing -fwrapv -DHAVE_CHICKEN_CONFIG_H -DC_ENABLE_PTABLES -g -O2 -fdebug-prefix-map=/build/chicken-8b5V5S/chicken-5.1.0=. -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -I/usr/include/chicken 'x86_64-linux-gnu-gcc' 'test.o' -o 'test' -L/usr/lib -Wl,-R/usr/lib -lchicken -lm -ldl vjo@glaucus/pts/11:chicken/man_or_boy % x86_64-linux-gnu-gcc --version x86_64-linux-gnu-gcc (Ubuntu 9.2.1-9ubuntu2) 9.2.1 20191008 Copyright (C) 2019 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. vjo@glaucus/pts/11:chicken/man_or_boy % csc -version CHICKEN (c) 2008-2019, The CHICKEN Team (c) 2000-2007, Felix L. Winkelmann Version 5.1.0 (rev 8e62f718) linux-unix-gnu-x86-64 [ 64bit dload ptables ] }}} " vjorlikowski Milestone someday 1739 math: failure in tests extensions 5.2.0 defect Diego assigned 2021-04-01T19:06:33+02:00 2021-04-11T16:09:44+02:00 See http://salmonella-linux-x86.call-cc.org/master-debugbuild/gcc/linux/x86/2021/03/31/salmonella-report/test/math.html Mario Domenech Goulart Milestone someday 1740 ephem: failures in tests (on 32bit systems) extensions 5.2.0 defect rca assigned 2021-04-01T19:10:12+02:00 2021-04-01T19:10:12+02:00 "Some ephem tests fail on 32bit systems. See http://salmonella-linux-x86.call-cc.org/master-debugbuild/gcc/linux/x86/2021/03/31/salmonella-report/test/ephem.html Tests pass on 64bit systems (e.g., https://salmonella-linux-x86-64.call-cc.org/master/tcc/linux/x86-64/2021/04/01/salmonella-report/test/ephem.html )" Mario Domenech Goulart Milestone someday 1769 Svn-client egg: changes needed to build under musl extensions 5.2.0 defect Peter Bex assigned 2021-07-11T15:42:44+02:00 2021-07-11T15:42:44+02:00 Christopher Brannon Milestone someday 1775 fuse tests hang extensions defect evhan assigned 2021-08-07T22:48:21+02:00 2021-10-27T23:18:11+02:00 "The test suite of fuse hangs sometimes. I don't know a way to reproduce that, but they hang from time to time. Hangs have been observed with: * CHICKEN 4 jobs on the salmonella-linux-x86-64 machine (Debian 9) * CHICKEN 5 jobs on the salmonella-linux-x86 machine (Debian 10) Below is some info: {{{ $ cat /etc/debian_version 9.13 $ clang --version clang version 3.8.1-24 (tags/RELEASE_381/final) Target: x86_64-pc-linux-gnu Thread model: posix InstalledDir: /usr/bin $ strace -p 23651 strace: Process 23651 attached restart_syscall(<... resuming interrupted poll ...> }}} " Mario Domenech Goulart Milestone someday 1778 Typo or incorrect export in fmt egg extensions defect Alex Shinn assigned 2021-08-25T16:54:30+02:00 2021-08-26T10:56:57+02:00 The documentation for the fmt egg describes a fmt-color procedure. The signature matches the fmt-color**ed** procedure, which is currently internal. I believe fmt-color should be internal or at least fmt-colored should be additionally exported and the documentation updated. Diego Milestone someday 1795 Types for srfi-69 egg extensions 5.3.0 defect new 2022-01-04T19:23:33+01:00 2025-07-25T18:30:34+02:00 srfi-69 is a widely-used CHICKEN egg, but it does not provide types--in particular, programs using srfi-69 cannot use hash-table as a type in their own code, but must use the non-specific * type instead. This makes it harder to catch mistakes when using this egg. Please add type information to srfi-69. zaifir Milestone someday 1801 parley constructs an internally inconsistent input port trivial extensions 5.3.0 defect new 2022-04-26T11:50:50+02:00 2022-04-27T06:58:43+02:00 "The parley egg still hasn't been ported successfully to C5. Most of the steps are done already (egg file, imports, tests), but interactive usage fails: {{{ #;2> (define p (let ((old (current-input-port))) (make-parley-port old))) #;3> p # #;4> (input-port? (current-input-port)) #t #;5> (input-port? p) #t #;6> (current-input-port p) Error: (current-input-port) bad argument type - not a port of the correct type: # Call history: (current-input-port p) (current-input-port p) <-- }}} Looking at the code performing checks, it seems that the I/O direction slot is checked whether it's an input one. Both `input-port?` and `current-input-port` seem to look at the same slot, but arrive to different conclusions. Perhaps C5 changed some port internals? One way of side-stepping this would be to avoid low-level port construction, but this might break C4 compatibility completely." Vasilij Schneidermann Milestone someday 1805 `html->sxml` with escaped quotes breaks text into multiple nodes extensions 5.3.0 defect Alex Shinn assigned 2022-06-10T20:53:06+02:00 2022-06-10T21:07:06+02:00 "There's some weirdness with escaping quotes in text when using `html->sxml`. Perhaps a short example would be sufficient to explain the problem I'm encountering: {{{ (html->sxml ""

foo'bar"baz

"") ;=> (*TOP* (p ""foo"" ""'"" ""bar"" ""\"""" ""baz"")) }}} As a counter-example, I'll use the [https://wiki.call-cc.org/eggref/5/ssax: ssax egg]: {{{ (call-with-input-string ""

foo'bar"baz

"") ;=> (*TOP* (p ""foo'bar\""baz"")) }}} I guess fundamentally it's a question of whether there should be one text node or not. I would argue that in this particular case, it should be a single node. I have been using html-parser to try and scrape some web pages, and this is extremely unexpected! Especially so if one uses `txpath` / `sxpath` on the final result, as `//p/text()` queries will not necessarily behave as expected. You would have to `(apply string-append ((txpath ""//p/text()"") sxml))` to the result to get the full contents of the text. Is there a rationale for this, or is that some kind of limitation of the parser? I know that tags may also contain sub-tags in HTML, but I'm not sure a new node should be made if a tag's contents are not HTML tags themselves." Jeremy Steward Milestone someday 1814 OS X / PowerPC failing test: remainder return -0.0 unknown 5.3.0 defect new 2023-02-06T08:57:28+01:00 2023-02-09T00:31:05+01:00 "I recently added CHICKEN 5.3.0 to my package manager for OS X Tiger/PowerPC and Leopard/PowerPC: https://github.com/cellularmitosis/leopard.sh/commit/192014b920692434337783199742ceb217694f27 When I ran the tests, there was only one failure: {{{ (FAIL) remainder: flo/flo: expected 0.0 but got -0.0 }}} I also verified it by hand: {{{ $ csi CHICKEN (c) 2008-2021, The CHICKEN Team (c) 2000-2007, Felix L. Winkelmann Version 5.3.0 (rev e31bbee5) macosx-unix-gnu-ppc [ 32bit dload ] Type ,? for help. #;1> (remainder 1.0 1.0) -0.0 #;2> }}} Thanks for all of your work on CHICKEN! " cellularmitosis Milestone someday 1833 inaccuracy in exponential functions unknown 5.3.0 defect new 2024-02-29T14:58:37+01:00 2024-02-29T14:58:37+01:00 "I wish to bring to your kind notice the following: (sin -1e-6i), whose magnitude should be sinh, gives inaccurate answer, in comparison SCM scheme for the same expression, gives the correct C function answer. The discrepancy is related to the C functions expm1 and log1p which are needed for small x. Another thing I noticed is the expression (sin 1e-6i) gives the error 1e-6 unbound variable, whereas in the expression (sin +1e-6i) it is correctly recognized as a number. Thanks." sjain59 Milestone someday 1845 srfi-69 .types errors extensions 5.4.0 defect new 2024-12-14T01:47:23+01:00 2024-12-14T01:47:23+01:00 " ===== hashes return fixes & are not pure ((srfi-69#number-hash (#(procedure clean: enforce:) srfi-69#number-hash (new: number old: fixnum #!optional fixnum fixnum) fixnum))) ((srfi-69#hash (new: #(procedure clean: enforce:) old: #(procedure pure: enforce:) srfi-69#hash (* #!optional fixnum fixnum) fixnum))) ((srfi-69#string-hash (#(procedure clean: enforce:) srfi-69#string-hash (string #!optional fixnum fixnum fixnum fixnum) new: fixnum old: number))) ((srfi-69#string-ci-hash (#(procedure clean: enforce:) srfi-69#string-ci-hash (string #!optional fixnum fixnum fixnum fixnum) new: fixnum old: number))) ((srfi-69#hash-by-identity (new: #(procedure clean: enforce:) old: #(procedure pure: enforce:) srfi-69#hash-by-identity (* #!optional fixnum fixnum) fixnum))) ((srfi-69#string-hash-ci (#(procedure clean: enforce:) srfi-69#string-hash-ci (string #!optional fixnum fixnum fixnum fixnum) new: fixnum old: number))) ===== min/max load is float & equiv-func is a slot ((srfi-69#hash-table-max-load (#(procedure clean: enforce:) srfi-69#hash-table-max-load ((struct hash-table)) new: float old: fixnum) (((struct hash-table)) (##sys#slot #(1) (quote 6))))) ((srfi-69#hash-table-min-load (#(procedure clean: enforce:) srfi-69#hash-table-min-load ((struct hash-table)) new: float old: fixnum) (((struct hash-table)) (##sys#slot #(1) (quote 5))))) ((srfi-69#hash-table-equivalence-function (#(procedure clean: enforce:) srfi-69#hash-table-equivalence-function ((struct hash-table)) (procedure (* *) *)) new: (((struct hash-table)) (##sys#slot #(1) (quote 3))))) " Kon Lovett Milestone someday 1851 utf8 egg: Missing char sets and outdated tables unknown 5.4.0 defect new 2025-04-08T08:31:26+02:00 2025-07-04T17:53:52+02:00 "The unicode-char-sets module of the utf8 egg is missing several character sets. In particular, there is no set for characters with the Numeric property (making it impossible to implement a Unicode-aware 'char-numeric?' in CHICKEN) or for any of the punctuation properties. The utf8-srfi-14 module includes char-set:digit and char-set:punctuation, but these are throwaway ASCII-only implementations (in a file that begins with ""Unicode capable char-sets"", no less!). These sets should be added. Furthermore, the sets that unicode-char-sets does provide seem to be built on data that is extremely out-of-date. The header comment in unicode-char-sets.scm claims the tables were generated in 2007." Zipheir Milestone someday 1853 utf8 egg: Missing character type predicates easy extensions 5.4.0 defect new 2025-07-07T16:58:15+02:00 2025-07-07T16:58:15+02:00 "The utf8 egg doesn't provide Unicode-aware versions of the R5RS character type predicates (`char-whitespace?`, etc.). The attached patch provides them; it is based on my updated version of the unicode-char-sets extension (see [http://bugs.call-cc.org/ticket/1851]). It can easily be modified to work with other versions of the egg by simply replacing the multiple char-set imports with `(import (unicode-char-sets))`. " zaifir Milestone someday 1863 delete-file and delete-file* inconsistency trivial unknown 5.3.0 defect new 2026-01-01T09:53:46+01:00 2026-01-01T09:53:46+01:00 "While trying to delete an invalid symbolic link file (for which the linked file is deleted), delete-file function succeeds, while delete-file* returns false. I have conducted a research which showed me that file-exists? on such invalid link also returns false, as it calls stat, and, therefore, follows symbolic links. I have prepared a patch which replaces stat with lstat. If that approach is not acceptable, I can try to fix it in a more preferred way. How to reproduce: $ touch some-file $ ln -s some-file some-link $ rm some-file $ csi -e '(import (chicken file)) (print (file-exists? ""some-link""))' " fa11-1eaf Milestone someday 1866 pretty-print and pp does not print UTF-8 characters unknown 6.0.0 defect new 2026-01-08T00:52:01+01:00 2026-01-08T00:52:01+01:00 "csi 6 (installed from https://code.call-cc.org/dev-snapshots/2025/10/25/chicken-6.0.0pre2.tar.gz): Version 6.0.0pre2 (rev e4e9ea4d) linux-unix-gnu-x86-64 [ 64bit dload ptables ] (import (chicken pretty-print)) (pp ""été"") => ""\xe9;t\xe9;"" (pretty-print ""été"") has the same problem in Chicken 6, Unicode code point for é is U+00E9 (display ""\xe9;"") => é (display ""é"" => é both work. print and write also work fine. Could it be that interpretations of backslashes is disabled ? csi 5: Version 5.2.0 (rev 317468e4) (import (chicken pretty-print)) (pp ""été"") => ""été"" " Kirill Lisovskiy Milestone someday 258 z3 egg improvements medium extensions enhancement Jim Ursetto new 2010-05-29T07:48:59+02:00 2023-11-06T22:01:15+01:00 "z3 implements RFC1951 (deflate) and RFC1952 (gzip) but not RFC1950 (zlib), which is basically deflate + a header and adler32 checksum. zlib format is used by some applications in lieu of a raw deflate stream -- for example, git objects are stored in this format. The default python and ruby zlib compression functions read and write zlib format as well, relegating raw deflate to second class status. That's probably because this is the standard behavior of libz itself, but support seems warranted. Also as in closed ticket #82, the interface should be improved to be more agnostic to input or output source, allowing any format to be read or written to a port or string. Backward compatibility should be retained where possible. " Jim Ursetto Milestone someday 1162 gethostbyname() is deprecated in favour of getaddrinfo() easy core libraries 4.9.x enhancement new 2014-10-13T22:34:36+02:00 2016-08-25T21:47:22+02:00 "The [http://lists.gnu.org/archive/html/chicken-users/2014-10/msg00007.html message by Moritz Wilhelmy] is copied below: {{{ While building the OpenSuSE package for Chicken, I discovered the following: [ 473s] RPMLINT report: [ 473s] =============== [ 482s] chicken-runtime.i586: I: binary-or-shlib-calls-gethostbyname /usr/lib/libchicken.so.7 [ 482s] The binary calls gethostbyname(). Please port the code to use getaddrinfo(). [ 482s] [ 482s] 5 packages and 0 specfiles checked; 0 errors, 0 warnings. [ 482s] (Full build log here: http://build.opensuse.org/package/live_build_log/home:mwilhelmy:chicken/chicken/openSUSE_12.3/i586 ) Do you think it would be a good idea to port chicken to use getaddrinfo() instead, not that it's top priority or anything? getaddrinfo is part of POSIX.1-2001 so it should be present on most systems, while POSIX.1-2008 removes gethostbyname() from the specification altogether. }}}" Mario Domenech Goulart Milestone someday 1176 Support capturing multiple value type in `forall`, etc. medium scrutinizer 4.9.x enhancement new 2015-01-12T06:08:45+01:00 2016-08-25T21:47:34+02:00 "As requested by alaricsp in [https://lists.nongnu.org/archive/html/chicken-users/2014-11/msg00016.html this thread], it would be nice if one could capture and refer to multiple-valued return types with `forall`, e.g.: {{{ (: call-with-foo (forall (a) (foo (-> . a) -> . a))) (define (call-with-foo foo thunk) ; do some stuff and return whatever thunk does, ; even if it's multi-valued: (thunk)) }}} (The above is just a sketch, not necessarily the right way to express this..)" evhan Milestone someday 1215 Warn about too specific type declarations hard scrutinizer 4.10.x enhancement new 2015-08-25T21:57:46+02:00 2018-04-27T16:45:24+02:00 "The following program produces code that may segfault when compiled with specialization: {{{ (: foo (-> string)) (define (foo) (read)) (print (string-ref (foo) 0)) }}} {{{ $ csc -specialize foo.scm echo '""foo""' | ./foo f $ echo 123 | ./foo Error: segmentation violation Call history: baz.scm:5: foo baz.scm:3: read <-- }}} I think the root cause is that the declared return type of {{{foo}}} is more specific than that of {{{read}}}. Perhaps this could be detected by the scrutinizer and warned about accordingly." Moritz Heidkamp Milestone someday 1310 feathers: show number of debug-events for each line easy core tools 4.11.0 enhancement felix winkelmann new 2016-08-10T14:07:54+02:00 2016-08-25T20:51:57+02:00 In the source-view, prefix each line with a count of debug-events that take place in that line. felix winkelmann Milestone someday 1311 "feathers: Implement ""Next""" medium core tools 4.11.0 enhancement new 2016-08-11T14:46:03+02:00 2016-08-25T20:52:11+02:00 ...which should execute in single-stepping mode until an event in the same source file occurs. felix winkelmann Milestone someday 1557 openssl load system certs easy extensions 4.13.0 enhancement Jim Ursetto assigned 2018-11-11T01:55:30+01:00 2021-04-12T20:17:53+02:00 "Patch attached to read the system default certs (configurable via parameter). It also removes (nullifies the action of) the default CA directory parameter because the default of /etc/ssl/certs only works on Debian based systems, and it doesn’t let you select a default certs.pem file anyway. This is tested on OS X Sierra and should work fine on Debian and RedHat, please test. This lets henrietta-cache automatically work on OS X with Homebrew openssl, which it could not before, as the cert path was invalid." Jim Ursetto Milestone someday 1606 Add code to lfa2 for extinguishing SRFI-4 predicates medium scrutinizer 5.0.0 enhancement felix winkelmann assigned 2019-04-11T13:57:05+02:00 2023-11-06T23:06:07+01:00 "Now that we have inlineable procedures for SRFI-4 vector access, since #757, we should also add code to lfa2 to extinguish predicates for these vectors completely. If possible, we should add unboxing for f32vectors and f64vectors, too." sjamaan Milestone someday 1702 Switch to srfi-115 medium core libraries 5.2.0 enhancement new 2020-07-10T09:46:12+02:00 2020-07-10T09:46:12+02:00 "We're currently using the irregex library, but srfi-115 is probably better-maintained as it's officially part of Chibi. Also it's standardized, which is always nice. As I understand it, SRFI-115 has complete feature parity with Irregex, including DFA compilation of submatches and full SRE support. However, IIRC there are tiny variations in the SRE dialect used by both libraries. We should do some performance comparisons and check what the differences are. Then, if both are equally fast, provide a shim library so we can still support the irregex API as well as the srfi-115 API." sjamaan Milestone someday 1732 [zmq egg] barebones curve encryption suport easy extensions 5.2.0 enhancement Moritz Heidkamp assigned 2021-03-11T22:30:16+01:00 2021-03-12T08:14:57+01:00 "I've been fiddling around with zmq for a project of mine and needed encryption support, so I tried my hand at implementing the necessary socket options for curve encryption Again, it's barebones, but authentication between two sockets seems to be working! I'm attaching my modifications and some testing code Cheers!" Ariela Wenner Milestone someday 1735 Strip away prefix when importing medium core libraries 5.2.0 enhancement new 2021-03-16T16:41:56+01:00 2021-04-22T09:22:47+02:00 "I know we can easily add prefix when importing but some eggs already come with too much prefix on all of their procedures. Wouldn't it be pretty sweet if we could import sans them? (import (sans filepath ""filepath:"") (sans irregex ""irregex-"")) I tried my jolly best to implement it in https://idiomdrottning.org/chicken-core To support both sans and prefix at the same time (when you want to remove their prefix and add your own much cooler prefix) is left as an exercise for the reader. This just removes it for now." Idiomdrottning Milestone someday 1776 More scrutable bootstrapping process insane unknown 5.2.0 enhancement new 2021-08-12T20:29:25+02:00 2021-08-12T20:29:25+02:00 "We currently have two ways to build CHICKEN: * from release tarballs. In that case we have the C code generated by CHICKEN. In this case, the only requirement to build CHICKEN is a supported C toolchain and GNU Make. The C code in release tarballs is not really ""source"" code, as it can be considered inscrutable by most humans. * from git. In this case we require a functional CHICKEN installation (i.e., binary format, compiled out of a release tarball -- see previous item) to compile Scheme code to C. Currently the most scrutable way to build CHICKEN is using the C code generated by CHICKEN to build CHICKEN. But then, as previously mentioned, we assume the C code we are compiling is thrustworthy, which can hardly be assessed as the C code is the output of a compiler, so very hard to follow. This ticket is about improving this situation, so that we can build CHICKEN from git in a way that the whole process is more scrutable. References: * http://issues.guix.gnu.org/22366 points out the issue reported in this ticket. * https://bootstrappable.org/ has more information on the benefits of transparent bootstrapping processes. " Mario Domenech Goulart Milestone someday 1798 No promise type. core libraries 5.3.0 enhancement new 2022-02-11T23:16:24+01:00 2023-11-06T23:39:46+01:00 Although promises are a Scheme type, `(chicken type)` does not provide a `promise` type. What ''does'' work is `(struct promise)`, which seems like an abstraction leak. Please expose a type for promises; even better, provide a parametric promise type, so that `(promise t)` is a type of delayed `t` values. Zipheir Milestone someday 1831 Use a parameter for the directory to be used by create-temporary-{file,directory} easy core libraries 6.0.0 enhancement new 2024-01-02T21:01:54+01:00 2024-01-02T21:01:54+01:00 "Currently, {{{create-temporary-file}}} and {{{create-temporary-directory}}} use the value of the {{{TMPDIR}}}, {{{TMP}}} and {{{TEMP}}} environment variables to determine the directory where to create files/directories, falling back to {{{/tmp}}} if none of them is set. Relying on environment variables to configure {{{create-temporary-file}}} and {{{create-temporary-directory}}} is not ideal, as changing those variables affect child processes as well. A better configuration option would be a parameter." Mario Domenech Goulart Milestone someday 747 coops: integrate primitive type classes extensions 4.7.x task felix winkelmann new 2011-12-13T13:30:40+01:00 2023-11-06T22:01:33+01:00 "Currently this is a separate file since it was considered to be an optional feature and would save some space. I'm not sure if this is worthwhile and writing methods on primitive classes is something that can be considered fundamental. IMHO `coops-primitive-objects.scm` should simply be integrated into `coops.scm` (or better: `coops-module.scm`). " felix winkelmann Milestone someday 752 test and document qt-light enhancements by satosi extensions 4.7.x task new 2011-12-13T13:42:29+01:00 2023-11-06T22:01:48+01:00 "See branch `satosi-enhancements`. " felix winkelmann Milestone someday 1476 (scrutinizer) Should global types be smashed in local contexts? scrutinizer 5.0.0 task new 2018-06-15T10:31:08+02:00 2023-11-06T23:08:34+01:00 "Currently types for globals are not smashed. Should they be smashed in, say functions? Similar to #1475. {{{ (define-type type (vector (or fixnum symbol))) (: g type) (define g (vector (the * 1))) ;; define pure alias, so printing doesn't smash (: p (* --> undefined)) (define p print) (define (foo) (compiler-typecase g (type (p ""global ok""))) (vector-set! g 0 'foo) (compiler-typecase g (type (p ""global not smashed""))) (let ([l g]) (compiler-typecase l (type (p ""alias ok""))) (compiler-typecase l (type (p ""alias ok after printing""))) (vector-set! g 0 'foo) (compiler-typecase g (type (p ""global not smashed""))) (compiler-typecase l (type 'oop) ((vector *) (p ""alias smashed"")))) (compiler-typecase g (type (p ""global not smashed"")) ((vector *) (p ""global smashed"")))) (foo) ;; $ csc -O3 global-smashing.scm && ./global-smashing ;; global ok ;; global not smashed ;; alias ok ;; alias ok after printing ;; global not smashed ;; alias smashed ;; global not smashed }}}" megane Milestone someday 1820 Clean up C preprocessor macros easy core libraries 5.3.0 task new 2023-06-03T11:13:11+02:00 2023-11-06T21:39:41+01:00 "We can replace `C_externimport` and `C_externexport` with `C_extern`. Also, `C_fcall` can be removed. Possibly also `C_regparm` and `C_ccall` as these refer to obsolete features. `C_TLS` was never used and should go. Note: `C_externimport` and `C_externexport` are produced bz the compiler, so need to be retained for at least one release." felix winkelmann Milestone someday 1155 values and the scrutinizer medium scrutinizer 4.9.x defect new 2014-09-19T00:37:12+02:00 2016-08-25T21:48:08+02:00 "The scrutinizer complains a little when faced with values under certain circumstances. Here are some examples: {{{ $ echo '(module val () (import scheme) (values))' > val.scm $ csc val.scm Warning: at toplevel: expected in `let' binding of `t10' a single result, but were given zero results }}} {{{ $ echo '(module val () (import chicken scheme) (and-let* ((#t)) (values)))' > al.scm $ csc al.scm Warning: at toplevel: branches in conditional expression differ in the number of results: (if #t (values) #f) }}} Of course, those warnings make sense, but I wonder if there is anything we can do to improve that situation." Mario Domenech Goulart Milestone someday 1351 syntax-rules has difficulty handling ellipsis in some cases? easy expander 5.0.0 defect sjamaan new 2017-03-10T10:59:09+01:00 2023-11-14T18:50:10+01:00 "As [http://lists.nongnu.org/archive/html/chicken-users/2017-03/msg00003.html reported by Sascha Ziemann], this macro application fails while it ought to succeed: {{{ (define-syntax define-facts (syntax-rules () ((_ (name a0 a1 ...) ((v00 v01 ...) (v10 v11 ...) ...)) '(define (name a0 a1 ...) (conde ((== a0 v00) (== a1 v01) ...) ((== a0 v10) (== a1 v11) ...) ...))))) }}} The failing application: {{{ (define-facts (fathero f c) ((""Abraham"" ""Ismael"") (""Abraham"" ""Isaac"") (""Isaac"" ""Jacob"") (""Jacob"" ""Benjamin""))) }}} A few other Schemes have difficulty with it, notably Chibi, Gauche, Scheme48 and Gambit (the former two returning the wrong result and the latter two giving an error). Racket, MIT, Ikarus and Guile work fine." sjamaan Milestone someday 1485 "edit of wiki entries - specific ""Description of your changes"" leads to error" wiki defect new 2018-07-12T11:29:39+02:00 2018-07-12T11:29:39+02:00 "Steps to reproduce (anonymous user - I created my user page for testing purposes): 1. https://wiki.call-cc.org/users/martin-schneeweis?action=edit 2. Any change in ""Article contents"" 3. ""Description of your changes"" - copy-paste the following 2 lines {{{ use cells -> use simple-cells examples: added missing object-type-declarations and comments }}} 4. Spam-Control-Math 5. Save ==> {{{ Error! Something went wrong while storing your changes. Please try again. If this error keeps up, please notify a system administrator about it. }}} " snoewchen Milestone someday 1731 Lowdown some minor issues extensions 5.1.0 defect Moritz Heidkamp assigned 2021-03-11T15:43:30+01:00 2021-04-14T21:16:03+02:00 "These are already fixed and on their way to upstream, but just in case someone else has the same problems. I changed one cons to be one append (in tight lists only); the bug was that (pp (markdown->sxml ""* like this"")) produced (ul (li (""like"" "" "" ""this""))) instead of, what it does now, (ul (li ""like"" "" "" ""this"")) This also fixes nested lists too, when you have a new bullet list or ordered list inside the item. The other problem was that it believed to be an in line element so hi did not work while in an or something worked. Just adding html among the block elements worked as a workaround but then
hi worked (it already did so that wasn't a regression. Just not right). So now it's a root element, it's own thing basically. My repo with these fixes are at https://idiomdrottning.org/lowdown" Idiomdrottning Milestone someday 776 shared lib link arguments in makefile are not transmitted to eggs medium build system 4.7.x enhancement new 2012-01-13T09:56:02+01:00 2023-11-06T21:59:16+01:00 "Linker options for shared libraries are all hardcoded in csc (or passed in via CSC_OPTIONS) -- any options set in the Makefile, e.g. via LINKER_LINK_SHARED_DLOADABLE_OPTIONS, are not honored. Specifically, on OS X, the {{{-headerpad 128}}} option is used to reserve extra space for {{{install_name_tool}}} to do its work. Without it, egg installation may fail once the Chicken PREFIX gets too long. The current workaround is to set {{{CSC_OPTIONS='-Wl,-headerpad -Wl,128'}}} prior to calling {{{chicken-install}}}. It might be nice to add variables into the makefile which are accessible by csc and used to provide default link options to eggs. This should also allow some of the hardcoded options in csc to be moved out into the Makefiles." Jim Ursetto Milestone someday 777 unix-sockets: implement FD sending and receiving easy extensions 4.7.x enhancement felix winkelmann assigned 2012-01-13T22:37:20+01:00 2023-11-06T22:02:03+01:00 Unix domain sockets allow sending and receiving FDs through them. It would be nice to add this functionality to the unix-sockets egg. Ruby has a pretty convenient implementation of this, maybe it could be used as a reference. See [http://apidock.com/ruby/UNIXSocket/send_io] and [http://apidock.com/ruby/UNIXSocket/recv_iop]. Moritz Heidkamp Milestone someday 809 allow to specify argument-type enforcement in type signatures medium compiler 4.7.x enhancement felix winkelmann assigned 2012-04-07T12:48:53+02:00 2023-11-06T22:02:21+01:00 "For example: `T1 ... ->! T2 ...` could specify a procedure type that is ""enforcing""." felix winkelmann Milestone someday 1020 matchable: Add conversion pattern(s) extensions 4.8.x enhancement Alex Shinn accepted 2013-07-12T10:48:47+02:00 2013-07-13T11:49:34+02:00 "The attached patch adds two new pattern types to matchable: {{{(-> convert pat-1 … pat-n)}}} {{{(-?> convert pat-1 … pat-n)}}} These patterns work like the {{{(? predicate pat-1 … pat-n)}}} pattern in that {{{convert}}} must be an expression evaluating to a single argument function which is applied to the corresponding value. However, in contrast to {{{?}}}, the following patterns {{{pat-1 … pat-n}}} are applied to the value _returned_ by {{{convert}}}. In addition, {{{-?>}}} only matches when {{{convert}}} does not return {{{#f}}}, so it can be expressed as combination of {{{->}}} and {{{?}}} like this: {{{(-> convert (? (lambda (x) x) pat-1 … pat-n))}}} The rationale is that it allows to match against a converted value and at the same time being able to bind that value to a variable instead of having to apply the conversion again in the body. One might argue that {{{-?>}}} is not really necessary so I'd be happy with only adding {{{->}}}, too. It should be noted that this patch is not backwards compatible. Some examples: {{{ (match '(""foo"" ""10"") ((""foo"" (-> string->number x)) (+ x x))) ;; => 20 (match '(""foo"" ""bar"") ((""foo"" (-> string->number x)) (+ x x))) ;; Error: (+) bad argument type: #f (match '(""foo"" ""10"") ((""foo"" (-?> string->number x)) (+ x x)) (else 'nothing)) ;; => 20 (match '(""foo"" ""bar"") ((""foo"" (-?> string->number x)) (+ x x)) (else 'nothing)) ;; => nothing }}}" Moritz Heidkamp Milestone someday 1061 syntax-rules should be able to automatically determine implicit exports hard expander 4.8.x enhancement new 2013-11-08T11:42:23+01:00 2016-08-25T21:31:33+02:00 "Given a module which exports some syntax which expands to an identifier which isn't explicitly exported and isn't itself a syntax identifier, implciit export of that identifier doesn't work: {{{ ;;; foo.scm (module foo (some-syntax) (import scheme) (define (some-implicitly-used-procedure x) (list x)) (define-syntax some-syntax (syntax-rules () ((_ x) (some-implicitly-used-procedure 'x)))) ) }}} To trigger the error: {{{ $ csc -s -J foo.scm $ csi -R foo -p '(some-syntax bar)' Error: unbound variable: some-implicitly-used-procedure }}} As can be witnessed in #1060, it works fine for implicitly exported syntax." Moritz Heidkamp Milestone someday 1134 with-input-from-pipe should be rewritten so it does not block hard core libraries 4.9.x enhancement new 2014-06-25T12:05:56+02:00 2016-08-25T21:48:19+02:00 "The current implementation of with-input-from-pipe is straightforward and uses the underlying POSIX functions popen and pclose. However the usage of pclose will wait for the child process of the pipe to exit and thus will hang the whole chicken process. This gets in the way if threading is used. Therefore it should be implemented in a way that we explicitly wait for the SIGCHLD to arrive in the scheduler. Maybe a thread-wait-for-signal mechanism needs to be implemented. This will complicate a lot of things so I will put this on the back burner for now." Christian Kellermann Milestone someday 1384 Simplify resolution of include/extend/types/inline/extension files compiler 5.0.0 enhancement new 2017-07-05T11:47:45+02:00 2017-07-05T11:47:45+02:00 "There are currently a bunch of different implementations of ""file resolution"", many of which amount to the same thing and some of which seem to be (mis|ab)used in certain places. It would be good to audit these and simplify things where possible. At a glance, {{{load-identifier-database}}} and {{{load-types-file}}} both have their own logic for locating files; {{{##sys#find-extension}}} is similar to these in that it searches the working directory and the repository, but also considers ""setup-mode"" (speaking of which, is this still necessary with the new egg system?) and the include path when looking for import libraries; and then there's {{{##sys#resolve-include-filename}}}, which is being used for include, extend, and inline files. This one is pretty complex and has a somewhat confusing signature. It would also be worth investigating whether {{{##sys#canonicalize-extension-path}}} is really necessary wherever it appears." evhan Milestone someday 1439 Add switch to csi to set initial module medium core tools 5.0.0 enhancement new 2018-01-03T06:42:39+01:00 2018-01-24T21:43:09+01:00 "Add a command line option to csi to specify the starting module, similar to Chibi's ""-x"" switch. This would be equivalent to running csi interactively and immediately using "",m "", but could be used with non-interactive scripts, for example." evhan Milestone someday 1445 Support parameterised hash table type declarations medium scrutinizer 5.0.0 enhancement new 2018-02-19T06:40:05+01:00 2018-02-22T20:46:21+01:00 "As suggested by megane on [https://lists.nongnu.org/archive/html/chicken-hackers/2018-02/msg00008.html chicken-hackers]. Support specifying the types of hash table keys and values: {{{ (define t (the (hash-table fixnum symbol) ...)) (hash-table-ref t 'foo) ; warning ""expected argument #2 of type `fixnum' but was given an argument of type `symbol'"" (hash-table-ref t 42) ; result would be known to be a symbol }}} And so on." evhan Milestone someday 1446 Implement contravariant type checking for procedure types scrutinizer 5.0.0 enhancement new 2018-02-20T16:40:45+01:00 2023-11-06T23:09:25+01:00 "Consider list types `A = (list foo)` and `B = (list bar)`. In functions where list of type `A` is expected you can give list of type `B` if `bar` is more specific (or of same type) than `foo`. For example `foo` could be `(or fixnum string)` and `bar` be `fixnum`. Now consider procedure types `F = (t1 -> *)` and `H = (t2 -> *)`. In functions where a function parameter of type `F` is expected you can give function of type `H` if `t2` is more '''general''' than `t1`. For example, if `t2` is `(or fixnum string)` and `t1` is `string`. Here is an example where the scrutinizer can't tell anything is wrong: {{{ (define-type dog (vector string)) (define-type cat (vector fixnum string)) (define-type animal (or cat dog)) (: make-dog (-> dog)) (define (make-dog) (vector ""woof"")) (define (dog? a) (= 1 (vector-length a))) (: make-cat (-> cat)) (define (make-cat) (vector 0 ""meow"")) (define (cat? a) (= 2 (vector-length a))) (: animal-print (animal -> *)) (define (animal-print a) (cond [(dog? a) (dog-print a)] [(cat? a) (cat-print a)])) (: dog-print (dog -> *)) (define (dog-print d) (print ""dog says "" (vector-ref d 0))) (: cat-print (cat -> *)) (define (cat-print c) (print ""cat says "" (vector-ref c 1))) (: apply-printer (forall (T) ((T -> *) T -> *))) (define (apply-printer pr e) (pr e)) (: print-animals ((animal -> *) (list-of animal) -> *)) (define (print-animals pr animals) (for-each pr animals)) ;; (apply-printer cat-print (make-dog)) ; warns ok ;; (print-animals animal-print (list (make-dog) (make-cat))) ; works ok ;; doesn't warn, but fails at runtime (print-animals cat-print (list (make-cat) (make-dog))) ;; $ csc -O3 vector-contravariant.scm && ./vector-contravariant ;; cat says meow ;; ;; Error: (vector-ref) out of range ;; #(""woof"") ;; 1 ;; ;; Call history: ;; ;; vector-contravariant.scm:44: print-animals ;; vector-contravariant.scm:36: g30 ;; vector-contravariant.scm:29: print ;; vector-contravariant.scm:36: g30 }}}" megane Milestone someday 1471 [types] Enforce parameter types on specialization medium scrutinizer 5.0.0 enhancement new 2018-06-06T09:02:43+02:00 2018-06-06T09:02:43+02:00 "Currently no enforcing happens if specialization takes place. Here's a case from scrutiny-tests. Note that scheme#+ has a specialization for parameters (* *). When using `-A` flag no specialization takes place, so the enforcing happens. In this example you'd like to have the type for `x` to be enforced to `number` after the `if`. {{{ ;; noreturn conditional branch enforces ""number"" on x (define (foo2 x) (if (string? x) (error ""foo"") (+ x 3)) (string-append x ""abc"")) ;; $ csc -O3 enforce-fail.scm ;; ;; Warning: in toplevel procedure `foo2': ;; (enforce-fail.scm:7) in procedure call to `scheme#string-append', expected argument #1 of type `string' but was given an argument of type `(not string)' ;; $ csc -A enforce-fail.scm ;; ;; Warning: in toplevel procedure `foo2': ;; (enforce-fail.scm:7) in procedure call to `scheme#string-append', expected argument #1 of type `string' but was given an argument of type `number' }}}" megane Milestone someday 1657 "It would be nice if glob ""/*/*"" worked" easy core libraries 5.1.0 enhancement new 2019-11-27T23:21:38+01:00 2021-08-29T21:19:25+02:00 Per [https://lists.nongnu.org/archive/html/chicken-users/2019-11/msg00013.html an email] from Matt Welland. evhan Milestone someday 1749 Parallelize CHICKEN core tests medium build system enhancement new 2021-04-18T00:02:43+02:00 2021-04-18T00:02:43+02:00 "= What? This ticket is meant to be the place to discuss everything related to the parallelization of the (current) CHICKEN core tests (`make check`). It's not a ""please do this"" ticket! I'm willing to work on it. I'd like to know if there's any interest in it, so please share your opinions, thoughts, and ideas! = Why? The tests take the majority of the time of the whole build process. On my desktop: * Build: ~3m32s (1 CPU); ~2m25s (2 CPUs); ~2m12s (3 CPUs); ~2m8s (4 CPUs) * Test: ~7m38s * Install: ~2m19s On my RBPi2 the whole process takes hours -- I usually leave it running over night. On my RBPi4 it takes around 45min, which is not so bad, I'm still impressed with it, but ""I want it all, and I want it now""! This is actually my biggest incentive, because of the salmonella tests. megane said on IRC that there's been some talk about rewriting the tests in Scheme, but that it's not for the near future, which I can understand. However, I think parallelizing the current tests should be pretty maneagable, and quicker than rewritting them in Scheme (a guess before actually putting my hands to work, of course), so we could take advantage of those in the meantime. = How? Assuming that there's interest, I'd like to know what others think -- ideas on how to implement, and how not to implement the tests. Related, I'd also like to know what's OK, and what isn't -- file structure, file names, number of new files in the repository, files created by the tests, etc. Now, my current idea/plan: Keep everything contained in the `tests/` directory, as it is now. Have a script for each test -- one for ""repository search path"", one for ""types.db consistency"", etc. Have an ""environment"" script that sets up the environment variables (`CHICKEN`, `CHICKEN_PROFILE`, ...), commands (`compile`, `compile_r`, ...), etc, to be used by each of the previously mentioned scripts. Have a Make target for each test, so the actual parallelization is managed by Make. Each test script should return the usual 0 for OK, and !0 for !OK, so that Make works correctly. The first obvious problem is the tests' output. Right now, since they're sequential, everything is printed to stdout, and all is fine. But running in parallel that's not an option. My first idea to solve this was to save the output of each test script to a test-specific output file. And the last thing I'm wondering about is whether having a test-specific directory inside `tests/` is worth it or not. This is mainly to unclutter the `tests/` directory, due to the tests' output files. That's all I have for now! Hopefully I didn't miss anything stupidly obvious... " siiky Milestone someday 1762 Maybe/Just Int checker in Silex compilation scrutinizer 5.2.0 enhancement reopened 2021-06-06T07:09:14+02:00 2021-06-06T18:04:50+02:00 "When chicken-install silex I get this Warning: Invalid argument In file `silex.scm:472', In module `silex', In procedure `digraph', In procedure `store-final338', In procedure `loop', In procedure call: (scheme#vector-set! prio voisin infinity) Argument #2 to procedure `vector-set!' has an invalid type: false The expected type is: fixnum This is the expression: voisin Procedure `vector-set!' from module `scheme' has this type: (vector fixnum * -> undefined) In the code, the index value ""voisin"" is derived from a vector-ref call on another vector that's initialized with #f's but populated by fixnums (hopefully? I didn't fully analyze the code here…) before being referenced. So probably one of two things are going on: The silex code is bugged and the variable won't be a fixnum in time (in which case this is is a serious bug in silex), or (what I believe is more likely) The type-checking code is mislabeling the situation here. It doesn't know that it'll be an int in time. This is just an, uh, ""cosmetic"" bug in core, which is how I'm labeling it. Or I'm wrong and it's instead a serious bug in exts (silex)." Idiomdrottning Milestone someday 1767 patch to port the netstring egg to CHICKEN 5 trivial extensions 5.2.0 enhancement Moritz Heidkamp assigned 2021-06-22T12:29:11+02:00 2022-10-21T16:00:33+02:00 "I needed to be able to talk to something that speaks netstrings, so I ported the netstring egg to CHICKEN 5. The patch is attached." Christopher Brannon Milestone someday 1848 cairo: quality of life improvements for matrix operations easy extensions 5.4.0 enhancement new 2025-01-02T19:34:44+01:00 2025-01-02T19:34:44+01:00 I've tried to manipulate a font matrix with cairo and failed. This seems to be due to that part of the interface being half-finished (text/font extents use records, matrices use f64vectors for some reason?), so I rewrote it to use records, exposed accessors and {{{cairo_get_font_matrix}}}. I've attached a small test program as well which exercises this feature set. Vasilij Schneidermann Milestone someday 1860 overwrite-error in chicken-install should identify egg medium core tools 6.0.0 enhancement new 2025-08-20T10:07:58+02:00 2025-08-20T10:07:58+02:00 When an egg installation would overwrite already installed files from another egg, it should show the offending egg in the error message. felix winkelmann