Mercurial > pub > dyncall > dyncall
annotate ToDo @ 132:b0cca984c740
- manual update, started splitting platform overview in EP and EL for some platforms
author | cslag |
---|---|
date | Sat, 09 Jul 2016 16:39:33 +0200 |
parents | a130676d0ceb |
children | fcbec299b736 |
rev | line source |
---|---|
0 | 1 1.0: |
2 ---- | |
86 | 3 - the CallVM-free functions are per VTable, however the latter can be changed out |
4 at runtime... which basically changes potentially the deallocator... | |
5 same goes for functions setting the mode. think about how to make this safer - currently | |
6 we have a higher flexibility and we need to potentially clean up from the mode we are in, | |
7 however, this is open to misuse | |
78 | 8 - check all vararg calling conv implementations if they promote correctly |
9 (e.g. float to double); we used to not care as it was the user's job to | |
10 think of that, but we allow users to use ArgFloat, etc. on "..." now, | |
11 taking care of the promotion; so it should be consistent all across the | |
12 implementations -> also write testcode specifically for that | |
0 | 13 - autovar was never really pushed all the way through and still plays second |
9 | 14 fiddle to dyncall_macros.h, clean this all up, eventually (Tassilo: for now, I |
15 removed out of sync files to avoid confusion; but still probably duped code) | |
0 | 16 - ppc32 callback fixes on bsd |
69 | 17 - add portable "make install" |
18 * make install should also install manpages | |
0 | 19 - mailing list announcements (html email?) |
20 - fix for NDS: | |
69 | 21 * ARM: |
104
dbca6763f2be
- complete, working mips o32 callback (using hardware fp); fixes error from last commit, which ignored first 2 float args
cslag
parents:
100
diff
changeset
|
22 - callback_plain retval wrong (not only platform), area on stack where |
dbca6763f2be
- complete, working mips o32 callback (using hardware fp); fixes error from last commit, which ignored first 2 float args
cslag
parents:
100
diff
changeset
|
23 it's stored is probably not zeroed, so half-words, etc., come with garbage |
69 | 24 * THUMB: |
25 - ellipsis (might be test itself, not respecting fixed part of args) | |
104
dbca6763f2be
- complete, working mips o32 callback (using hardware fp); fixes error from last commit, which ignored first 2 float args
cslag
parents:
100
diff
changeset
|
26 - callback_plain retval, see above under ARM |
69 | 27 - callf |
0 | 28 - fix Minix/x86 callbacks (see callback_suite) |
54 | 29 - armhf ellipsis: fix weirdness with long long as first ... arg |
24 | 30 - support Solaris11 CFLAGS with -D_FILE_OFFSET_BITS=64 |
31 - make sure selinux works (esp. regarding NX bits, asm might need (or similar): .section .note.GNU-stack,"",@progbits) | |
54 | 32 Daniel has a patch for it, if needed, implementing allocwx with two memory blocks, one for w^x, mmaped to one for r|x |
68 | 33 - support /SAFESEH on cl/win32 |
83
54930a037e8a
- PPC64 single-precision float fixes for more than 13 float args (thanks Masanori!)
cslag
parents:
78
diff
changeset
|
34 - test code functions should be in .so files, optionally, so the suites can directly |
54930a037e8a
- PPC64 single-precision float fixes for more than 13 float args (thanks Masanori!)
cslag
parents:
78
diff
changeset
|
35 be used to test dycnall bindings |
54930a037e8a
- PPC64 single-precision float fixes for more than 13 float args (thanks Masanori!)
cslag
parents:
78
diff
changeset
|
36 - pkg-config support? |
90 | 37 - quadmath support (long double) |
117 | 38 - add tests using alloca(), as this might depend on the frame pointer being correctly |
39 set; we don't check for this, currently, at all | |
0 | 40 |
41 portasm: | |
42 -------- | |
43 - add solaris support for x64 | |
44 | |
45 dyncall: | |
46 -------- | |
47 - support for return values: aggregate return values | |
48 - support for argument values: structures, half-precision, vector types, long double | |
128
9a554d1c4da6
- mips n64 vararg call fixes if non-variadic arguments are floats
cslag
parents:
127
diff
changeset
|
49 - varargs for mips (might exist, but test all ABIs (mipseb n64 already confirmed)) |
0 | 50 - consider automatic type-promotion for arguments passed through DC_CALL_C_ELLIPSIS_VARARGS; |
51 this would make it easier to just pass arguments instead of having to know about the C | |
88 | 52 type promotions (this is partly done, see todo-item above under 1.0) |
0 | 53 - syscalls for x64 |
54 * bsd | |
55 * linux | |
56 - other syscalls | |
110 | 57 - MIPS32 softfloat support (-msoft-float) |
127
645443fcfb47
- doc and todo update to reflect working mips64 n64 big endian support
cslag
parents:
117
diff
changeset
|
58 - MIPS64 little-endian needs testing (current port works on openbsd/octeon, which is big endian) |
132
b0cca984c740
- manual update, started splitting platform overview in EP and EL for some platforms
cslag
parents:
129
diff
changeset
|
59 - implement MIPS64 N32 (gcc -mabi=n32); both, little and big-endian (looks like NetNBSD on and |
b0cca984c740
- manual update, started splitting platform overview in EP and EL for some platforms
cslag
parents:
129
diff
changeset
|
60 EdgeRouter lite uses this: https://blog.netbsd.org/tnf/entry/hands_on_experience_with_edgerouter) |
129 | 61 - implement MIPS64 EABI (gcc -mabi=eabi); both, little and big-endian |
0 | 62 |
63 dynload: | |
64 -------- | |
65 - bug: test/nm crashes on qemu mips linux o32 debian etch | |
66 - test/resolve_self crashes/asserts on windows and bsds | |
67 - rename SymsInit to InitSyms and similar for cleanup function to reflect naming | |
68 in dyncall | |
69 - get rid of Dl_info and friends (_GNU_SOURCE) in dynload/dynload_syms_elf.c | |
70 for implementation of dlSymsNameFromValue(..) | |
71 - check if PSP can support dynload (missing elf.h, dlfcn.h, etc.); fixup readme if successful | |
84 | 72 - some functionality like java's System.mapLibraryName('name') might be helpful? |
0 | 73 |
74 dyncallback: | |
75 ------------ | |
89 | 76 - add SPARC and SPARC64 callback support |
104
dbca6763f2be
- complete, working mips o32 callback (using hardware fp); fixes error from last commit, which ignored first 2 float args
cslag
parents:
100
diff
changeset
|
77 - callback_plain's return value not correct anymore on NDS (maybe just broken testcode?), |
dbca6763f2be
- complete, working mips o32 callback (using hardware fp); fixes error from last commit, which ignored first 2 float args
cslag
parents:
100
diff
changeset
|
78 see above under 1.0 items |
dbca6763f2be
- complete, working mips o32 callback (using hardware fp); fixes error from last commit, which ignored first 2 float args
cslag
parents:
100
diff
changeset
|
79 * check other platforms also, if asm stub initializes retval space, correctly |
21 | 80 - finish PPC32 callbacks (see bugs section, below, BSD not working) |
110 | 81 - MIPS32 softfloat support (-msoft-float) |
116 | 82 - test MIPS o32 big endian (current port works on netbsd/pmax, which is little endian, and gcc -EB doesn't work on there) |
83 - test MIPS n64 little endian (current port works on openbsd/octeon, which is big endian) | |
132
b0cca984c740
- manual update, started splitting platform overview in EP and EL for some platforms
cslag
parents:
129
diff
changeset
|
84 - implement MIPS64 N32 (gcc -mabi=n32); both, little and big-endian (looks like NetNBSD on and |
b0cca984c740
- manual update, started splitting platform overview in EP and EL for some platforms
cslag
parents:
129
diff
changeset
|
85 EdgeRouter lite uses this: https://blog.netbsd.org/tnf/entry/hands_on_experience_with_edgerouter) |
129 | 86 - implement MIPS64 EABI (gcc -mabi=eabi); both, little and big-endian |
0 | 87 |
88 bindings: | |
89 --------- | |
90 - release bindings as standalone packages (already done for rbdc as a gem and rdyncall on cran) | |
91 * add note to documentation, where bindings are and that some bindings are svn-only ATM | |
92 - add rdoc documentation to ruby gem | |
93 - add pydoc for python | |
94 - add manpage for shdc | |
95 - expose callf-convenience functions to go | |
96 - add godoc for godc | |
97 - add javadoc to java | |
98 - add more bindings | |
99 - add callback support for erlang, go, java, lua, python and ruby (and shell? how?) | |
100 - add call-conv-mode-support for rbdc through signature | |
101 | |
102 manual: | |
103 ------- | |
100
dcb7157392f1
- manual: fixed html output of tables rendered as .png, using tex multirow
cslag
parents:
98
diff
changeset
|
104 - callconv diagrams: |
98 | 105 * good idea to review it all, anyways, if the calling conventions are actually |
100
dcb7157392f1
- manual: fixed html output of tables rendered as .png, using tex multirow
cslag
parents:
98
diff
changeset
|
106 represented correctly (and if they are in sync with the .pdf, we used to have |
dcb7157392f1
- manual: fixed html output of tables rendered as .png, using tex multirow
cslag
parents:
98
diff
changeset
|
107 problems with multirow.4ht) |
0 | 108 - provide more examples on how to use dyncall, dynload and dyncallback |
109 * e.g. enhance manual with a couple of examples (e.g. calling MessageBoxA on windows, etc.) | |
110 - manuals for bindings (as stated in bindings-section, above) | |
117 | 111 - under "dyncall support", in each calling convention appendix, list crucial details of e.g. big/little endian support, etc. |
112 - specify integer models in use, use common ILP32, LLP64, etc. terms | |
0 | 113 |
88 | 114 regressions: |
115 ------------ | |
0 | 116 - finish SPARC/SPARC64 calling conventions in doc's callconv-appendix !!! |
117 - Haiku/gmake build doesn't build dynload due to the fact, that elf.h is not found | |
118 on default installations (one has to install the system headers) | |
119 * add to doc | |
120 - Solaris/gmake build doesn't build dynload and tests anymore (maybe not even dyncallback) | |
121 - test arm32/ATPCS/THUMB (and then change Status in doc Appendix) | |
122 - look over code and fix TODO or @@@ marks | |
123 - add ppc-support to doc (both, 32 and 64, as well as little and big endian remarks) | |
124 | |
125 nice to have: | |
126 ------------- | |
127 - update: microsoft visual c++ build files - integrate all test suites in solution file | |
128 - consistency: on x64 platforms: make both ABIs available ('win64','sysv') for all x64 OS ports | |
129 - microsoft build enhancement: autodetect assembly via C preprocessor (like with .S gcc files) | |
130 - test/thunk: some platforms do not allow for allocation of executable code on heap and stack -- | |
131 currently this test will crash, maybe we should collect information which platforms allow and which not. | |
132 | |
133 bugs: | |
134 ----- | |
88 | 135 - setting callconv, specifically for platforms that have only one (default)) excepts on some |
6 | 136 platforms, eg. setting X64_sysv on amd64/Linux - keeping cdecl which is the same in |
88 | 137 this case, works (this should work now...) |
0 | 138 - vararg function calls don't work for ARM THUMB mode (Daniel's new interface for modesetting for varag |
139 args should be used); adapt test/ellipsis and/or retest | |
140 | |
141 missing implementations (descending priority): | |
142 ---------------------------------------------- | |
143 - riscv | |
144 - itanium | |
88 | 145 - superh/j-core |
0 | 146 - 68k, 88k, 6502, etc. |
88 | 147 - arc |
0 | 148 - alpha |
149 - cell | |
150 - blackfin | |
88 | 151 - seaforth |
21 | 152 - s/390 |
0 | 153 - z80 |
154 - hppa | |
155 - vax | |
156 - atmel avr | |
157 - propeller | |
158 - amd29k | |
159 |