source: project/chicken/trunk/TODO @ 15270

Last change on this file since 15270 was 15270, checked in by felix winkelmann, 12 years ago

enabled gc-report shows reclaimed locative-table entries

File size: 6.4 KB
1TODO for chicken                                                        -*- Outline -*-
5* bugs
7** compiler
8*** pre-optimization
9**** changes call-sites and makes them invalid for later pre-optimization
10*** check profiling (-profile vs. profile-declaration)
11**** this doesn't always seem to work, in particular with library units
12**** ##sys#finish-profile should always be invoked
13     is it in `exit'?
14*** when re-defining intrinsics, the compiler should warn and disable re-writes
15    add declaration to keep re-writes enabled for core library files
16*** check in foreign.import.scm and compiler.import.scm whether the import
17    took place in the compiler
18*** -prologue, -epilogue, -prelude, -postlude should check for argument being directory
19    (reported by Eduardo Cavazos)
21** expander
22*** expansion of r5rs_pitfall 4.2 incorrect
23    possibly due to a non-aliased implicit "begin" somewhere
24*** dirty-macros.scm loops when using `defile'
25    possibly due to unrenamed special forms
26*** extended lambda-lists refer to `optional' and `let-optionals[*]'
27    this will break, when these macros are renamed on import or not imported
28    at all
29**** a possible solution is to use internal forms, provided by the "scheme" module.
30*** need way to force generating module-registration code for standalone executables.
32** modules
33*** `require-extension' fails in interpreter when extension is .so without import library
34    probably because load and import is in same toplevel form, and the
35    load is performed after the import (at runtime). A possibly fix
36    would be to load the extension immediately (at
37    compile-/expansion-time), when import is #t and o import lib
38    exists.
39*** mark import-source (module) on plist of imported symbols and use in re-import warning
40    doesn't work that easily, try to find another solution
41*** DSSSL lambda-lists need imports of `chicken' module ("optional", "let-optionals", ...)
42**** should probably use internal aliases
44** libraries
45*** library/runtime: cyclic list checks for assq/assv/assoc/memq/memv/member
46    and C_i_list_tail
47*** finalizers on lexically ref'd data not working in interpreter
48    reported by Jim Ursetto
49    reverted original patch, see patches/finalizer-closures.diff
51** tools
52*** chicken-bug: SMTP servers not accessible (greylisting)
55* tasks
57** branches
58*** try to improve performance in lazy-gensyms
60** module issues
61*** need a way to omit generating module registration without generating import lib
62    for example when compiling static version, where implib already exists
63*** code-duplication in compiler and evaluator for ##core#module
64*** "scheme" module does not include some special forms ("define-syntax", etc.)
65*** import-for-syntax imports value bindings into import-env
66    should actually be a distinct meta-import-env.
67    example: we need `(import-for-syntax chicken)' to have access to
68    `receive' in a procedural syntax definition.
69*** curried define performs expansion in empty se - problem?
70    (as comment in expand.scm indicated (##sys#register-export))
71*** check phase separation and module access
72**** see "expander" above
73*** consider adding support for unnamed modules
75** compiler
76*** (csc) generate object-files in /tmp (or TMPDIR)?
77*** try to improve inlining (single calls should be inlinable)
78    Currently only "simple" procedures are inlined, possibly because accessed variables
79    may not be lexically visible but access to globals should ok, or not?
80*** option ("-M") to compile file as unnamed module (importing scheme + chicken)
82** setup/install
83*** setup-download
84**** use "HTTP_PROXY"/"http_proxy" env. var
85**** handle redirects in http-fetch
86**** option in chicken-install to list available eggs
88** library units
89*** read-mark list should be stored in read-table
90*** ports should have indicator slot about exact type (tcp, file, ...)
91*** normalize-pathname: return short name on windows? (mingw/msvc)
92*** Use record-descriptors instead of symbols as 1st slot in structure objects?
93**** see Kon's proposal for new record-descriptors in "misc/Chicken Runtime Data Type Proposal"
94*** Weak locatives can probably be used to implement weak hash-tables
96** syntax-error
97*** if ##sys#current-module is set, add name to error message?
98*** fluidly keep track of expanded forms (extend meaning of culprit) to pprint pruned expr on error
100** build
101*** need script to process import libraries for generating indices for
102    then tell Toby Butzon about it
103*** test DESTDIR and proper linking
104*** using "touch" with WINDOWS_SHELL won't work (need alternative)
105*** extend scripts/ for more platforms
107** documentation
108*** document qs, normalize-pathname
110** scrutiny
111*** allow giving toplevel procedure names to `scrutinize' option?
112*** write test file to trigger every type of warning (diff with result file in test-suite)
113*** add support for keyword arguments and check even length and available keywords
115** libraries
116*** analyse usage statistics ("-debug v") of all core libraries
119* wiki
121** compatibility page
124* tests
126** optional longer run (env. var)
127*** syntax-rules-stress-test
128*** fully compiled ec-tests
129** 3-stage bootstrap with compiler-output comparison
132* optimizations
134** self-recursion optimization
135   what MacScheme called "benchmark-mode" (assume self-calls are recursion)
136*** needs declaration or option, >= -O2
138** when inlining, consing arg-list with "list" may make get-keyword possibly foldable
140** compiler-support for get-keyword ?
142** lambda-fusion / "fuse-and-dispatch" (suggested by Alex Shinn)
143   convert groups of local lambdas referenced to only in operator-position into
144   looping lambda + dispatch (static variable can be used), otherwise similar to
145   a conditional
147** lazy gensyms (see "lazy-gensyms" branch)
149** handle optional args primitively
150   for example, set to distinguished val (checking argc) on C level, core
151   primitive for checking this - should be quite simple, but llist processing
152   (decompose-lambda-list) will be more expensive and cumbersome.
154** pre-hashed symbols (extra symbol slot)
155   The memory usage should be acceptable, performance gain is hard to guess.
156   Some experiments indicate that hashing the string is cheaper than it appears,
157   but low-level hashtables should get the most of this (and thus speed up
158   the compiler)
160** Lisp2-hack by Dybvig
161*** add function-cell to symbol, call on direct call without checks
162*** initialized to trap function on set!/define, which does fixup
163*** would also allow inline-caching hacks
Note: See TracBrowser for help on using the repository browser.