1 | TODO -*- org -*- |
---|
2 | ==== |
---|
3 | |
---|
4 | |
---|
5 | * bugs |
---|
6 | |
---|
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 | |
---|
19 | ** expander |
---|
20 | *** TODO unstripped quoted constants :BUG: |
---|
21 | if macro expands into the name of an exported macro, ##sys#strip-syntax will |
---|
22 | not strip the symbol (the se contains a macro-def). Possibly just use the |
---|
23 | macro name, then. |
---|
24 | *** expansion of r5rs_pitfall 4.2 incorrect |
---|
25 | possibly due to a non-aliased implicit "begin" somewhere |
---|
26 | *** dirty-macros.scm loops when using `defile' |
---|
27 | possibly due to unrenamed special forms |
---|
28 | *** extended lambda-lists refer to `optional' and `let-optionals[*]' |
---|
29 | this will break, when these macros are renamed on import |
---|
30 | |
---|
31 | ** modules |
---|
32 | *** `require-extension' fails in interpreter when extension is .so without import library |
---|
33 | probably because load and import is in same toplevel form, and the |
---|
34 | load is performed after the import (at runtime). A possibly fix |
---|
35 | would be to load the extension immediately (at |
---|
36 | compile-/expansion-time), when import is #t and o import lib |
---|
37 | exists. |
---|
38 | *** mark import-source (module) on plist of imported symbols and use in re-import warning |
---|
39 | doesn't work that easily, try to find another solution |
---|
40 | |
---|
41 | ** libraries |
---|
42 | *** library/runtime: cyclic list checks for assq/assv/assoc/memq/memv/member |
---|
43 | and C_i_list_tail |
---|
44 | *** finalizers on lexically ref'd data not working in interpreter (reported |
---|
45 | by Jim Ursetto) |
---|
46 | |
---|
47 | |
---|
48 | * tasks |
---|
49 | |
---|
50 | ** Convert this file to org mode |
---|
51 | |
---|
52 | ** branches |
---|
53 | *** try to improve performance in lazy-gensyms |
---|
54 | |
---|
55 | ** expander |
---|
56 | *** test new implementation of `define-for-syntax' |
---|
57 | **** test "numbers" egg |
---|
58 | **** is s48-modules still working? |
---|
59 | *** at some stage remove debug-output in expand.scm |
---|
60 | |
---|
61 | ** module issues |
---|
62 | *** need a way to omit generating module registration without generating import lib |
---|
63 | for example when compiling static version, where implib already exists |
---|
64 | *** code-duplication in compiler and evaluator for ##core#module |
---|
65 | *** "scheme" module does not include some special forms ("define-syntax", etc.) |
---|
66 | *** import-for-syntax imports value bindings into import-env |
---|
67 | should actually be a distinct meta-import-env. |
---|
68 | (does it work at all?) |
---|
69 | example: we need `(import-for-syntax chicken)' to have access to |
---|
70 | `receive' in a procedural syntax definition. |
---|
71 | **** could be tested with xlib egg |
---|
72 | *** curried define performs expansion in empty se - problem? |
---|
73 | (as comment in expand.scm indicated (##sys#register-export)) |
---|
74 | *** check phase separation and module access |
---|
75 | **** see "expander" above |
---|
76 | **** interaction for define-for-syntax, begin-for-syntax, and macros |
---|
77 | ***** check for different execution modes (compile-time, run-time, csi, ...) |
---|
78 | ***** figure out visibility |
---|
79 | |
---|
80 | ** compiler |
---|
81 | *** test define-rewrite-rule |
---|
82 | *** (csc) generate object-files in /tmp (or TMPDIR)? |
---|
83 | |
---|
84 | ** setup/install |
---|
85 | *** setup-download |
---|
86 | **** use "HTTP_PROXY"/"http_proxy" env. var |
---|
87 | **** handle redirects in http-fetch |
---|
88 | *** automatically update db after extension installation? |
---|
89 | *** test sudo |
---|
90 | |
---|
91 | ** library units |
---|
92 | *** read-mark list should be stored in read-table |
---|
93 | *** ports should have indicator slot about exact type (tcp, file, ...) |
---|
94 | *** normalize-pathname: return short name on windows? (mingw/msvc) |
---|
95 | |
---|
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 |
---|
99 | |
---|
100 | ** build |
---|
101 | *** need script to process import libraries for generating indices for doc.callcc.org |
---|
102 | then tell Toby Butzon about it |
---|
103 | *** test DESTDIR and proper linking |
---|
104 | *** test special status of setup-api.import.scm and setup-download.import.scm |
---|
105 | touching .c files caused chicken-install to be rebuilt; did using |
---|
106 | .SECONDARY help? |
---|
107 | *** using "touch" with WINDOWS_SHELL won't work (need alternative) |
---|
108 | |
---|
109 | ** documentation |
---|
110 | *** document qs, normalize-pathname |
---|
111 | |
---|
112 | |
---|
113 | * tests |
---|
114 | |
---|
115 | ** optional longer run (env. var) |
---|
116 | *** syntax-rules-stress-test |
---|
117 | *** fully compiled ec-tests |
---|
118 | ** 3-stage bootstrap with compiler-output comparison |
---|
119 | |
---|
120 | |
---|
121 | * documentation |
---|
122 | |
---|
123 | ** document new .meta entries in tutorials on wiki |
---|
124 | depends, test-depends |
---|
125 | ** chicken.texi needs to be regenerated |
---|
126 | |
---|
127 | |
---|
128 | * optimizations |
---|
129 | |
---|
130 | ** self-recursion optimization |
---|
131 | what MacScheme called "benchmark-mode" (assume self-calls are recursion) |
---|
132 | *** needs declaration or option, >= -O2 |
---|
133 | |
---|
134 | ** when inlining, consing arg-list with "list" may make get-keyword possible foldable |
---|
135 | |
---|
136 | ** compiler-support for get-keyword ? |
---|
137 | |
---|
138 | ** lambda-fusion / "fuse-and-dispatch" (suggested by Alex Shinn) |
---|
139 | convert groups of local lambdas referenced to only in operator-position into |
---|
140 | looping lambda + dispatch (static variable can be used), otherwise similar to |
---|
141 | a conditional |
---|
142 | *** new forms (after optimization, prepared language) |
---|
143 | [##core#dispatch LAMBDABODY1 ... BODY] |
---|
144 | [##core#goto {INDEX} ARGUMENT1 ...} |
---|
145 | |
---|
146 | ** lazy gensyms (see "lazy-gensyms" branch) |
---|
147 | |
---|
148 | ** handle optional args primitively |
---|
149 | for example, set to distinguished val (checking argc) on C level, core |
---|
150 | primitive for checking this - should be quite simple, but llist processing |
---|
151 | (decompose-lambda-list) will be more expensive and cumbersome. |
---|
152 | |
---|
153 | ** pre-hashed symbols (extra symbol slot) |
---|
154 | The memory usage should be acceptable, performance gain is hard to guess. |
---|
155 | Some experiments indicate that hashing the string is cheaper than it appears, |
---|
156 | but low-level hashtables should get the most of this (and thus speed up |
---|
157 | the compiler) |
---|
158 | |
---|
159 | ** Lisp2-hack by Dybvig |
---|
160 | *** add function-cell to symbol, call on direct call without checks |
---|
161 | *** initialized to trap function on set!/define, which does fixup |
---|
162 | *** would also allow inline-caching hacks |
---|