Commit | Line | Data |
---|---|---|
c906108c SS |
1 | If you find inaccuracies in this list, please send mail to |
2 | bug-gdb@prep.ai.mit.edu. If you would like to work on any of these, | |
3 | you should consider sending mail to the same address, to find out | |
4 | whether anyone else is working on it. | |
5 | ||
138f88c0 AC |
6 | |
7 | TODO: GDB 5.0 | |
8 | ============= | |
9 | ||
10 | Here are _all_ the issues that have been raised vis-a-vis the 5.0 | |
11 | release. Also check the GDB, and other, mail archives | |
12 | (http://sourceware.cygnus.com/ml/gdb/). | |
13 | ||
14 | If, however, you fix something, then feel free to tweek this file | |
15 | (deleting the problem). Just send a note to gdb-patches so that I see | |
16 | the change. | |
17 | ||
18 | The names in paren are those that might know more about the problem. | |
19 | They don't necessarily indicate the people that will fix the problem. | |
20 | ||
21 | -- | |
22 | ||
23 | GDB 5.0: Must have | |
c906108c SS |
24 | ------------------ |
25 | ||
138f88c0 AC |
26 | These are things that have been identifed as must-have for this |
27 | release of GDB. | |
28 | ||
29 | -- | |
30 | ||
138f88c0 AC |
31 | Solaris/x86 - which? (Nick Duffek, Peter Schauer, Michael Snyder?) |
32 | ||
33 | Nick D's working through patches from Michael Snyder and Peter S. | |
34 | ||
35 | -- | |
36 | ||
138f88c0 | 37 | |
67edb2c6 AC |
38 | RFA: procfs.c: init_procfs_ops should set |
39 | procfs_ops.to_has_[all]_memory (Peter Schauer, Andrew Cagney?) | |
138f88c0 | 40 | |
67edb2c6 AC |
41 | I am pretty sure that this is caused by some accidental deletion, but |
42 | procfs.c:init_procfs_ops no longer sets procfs_ops.to_has_memory and | |
43 | procfs_ops.to_has_all_memory. | |
138f88c0 | 44 | |
67edb2c6 | 45 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg01057.html |
138f88c0 AC |
46 | |
47 | -- | |
48 | ||
49 | GDB 5.0: Nice to have | |
50 | --------------------- | |
51 | ||
52 | These are things that might make it in 5.0 but don't sit in the | |
53 | critical path. If they miss the 5.0 cut then they definitly should | |
54 | make the follow-on release. | |
55 | ||
56 | -- | |
57 | ||
58 | Generic: lin-thread cannot handle thread exit (Mark Kettenis, Michael Snyder) | |
59 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00525.html | |
60 | ||
61 | The thread_db assisted debugging code doesn't handle exiting threads | |
62 | properly, at least in combination with glibc 2.1.3 (the framework is | |
63 | there, just not the actual code). There are at least two problems | |
64 | that prevent this from working. | |
65 | ||
66 | As an additional reference point, the pre thread_db code didn't work | |
67 | either. | |
68 | ||
69 | -- | |
70 | ||
71 | Java (Anthony Green, David Taylor) | |
72 | ||
73 | Anthony Green has started contributing late breaking Java patches: | |
74 | ||
75 | Patch: java tests | |
76 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00512.html | |
77 | ||
78 | Patch: java booleans | |
79 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00515.html | |
80 | ||
81 | Patch: handle N_MAIN stab | |
82 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00527.html | |
83 | ||
84 | It should be able to squeeze these in. | |
85 | ||
86 | -- | |
87 | ||
88 | Pascal (Pierre Muller, David Taylor) | |
89 | ||
90 | The pascal support patches nave been added to the patch data base. I | |
91 | [cagney] strongly suspect that they are better suited for 5.1. | |
92 | ||
67edb2c6 AC |
93 | Indent -gnu ? |
94 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00496.html | |
95 | ||
96 | 2 pascal language patches inserted in database | |
97 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00521.html | |
98 | ||
138f88c0 AC |
99 | -- |
100 | ||
101 | Programs run under GDB have SIGCHLD masked. | |
102 | ||
103 | [I think this can be worked around by using the action command - | |
104 | cagney] | |
105 | ||
106 | -- | |
107 | ||
108 | GNU/Linux/x86 and random thread signals (and Solaris/SPARC but not | |
109 | Solaris/x86) | |
110 | ||
111 | Christopher Blizzard writes: | |
112 | ||
113 | So, I've done some more digging into this and it looks like Jim | |
114 | Kingdon has reported this problem in the past: | |
115 | ||
116 | http://sourceware.cygnus.com/ml/bug-gdb/1999-10/msg00058.html | |
117 | ||
118 | I can reproduce this problem both with and without Tom's patch. Has | |
119 | anyone seen this before? Maybe have a solution for it hanging around? | |
120 | :) | |
121 | ||
122 | There's a test case for this documented at: | |
123 | ||
124 | when debugging threaded applications you get extra SIGTRAPs | |
125 | http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=9565 | |
126 | ||
127 | [There should be a GDB testcase - cagney] | |
128 | ||
129 | -- | |
130 | ||
131 | IRIX? | |
132 | ||
133 | Benjamin Gamsa wrote: | |
134 | ||
135 | Has anyone successfully built the latest (from cvs) gdb on IRIX6.4 or | |
136 | later? The first problem I hit is that proc-api.c includes | |
137 | sys/user.h, which no longer exists under IRIX6.4. If I comment out | |
138 | that include, the next problem I hit is that PIOCGETPR and PIOCGETU | |
139 | are no longer defined in IRIX6.4 (presumably related to the | |
140 | disappearance of user.h). | |
141 | ||
142 | -- | |
143 | ||
144 | Regressions (prologue) with devel GCC. | |
145 | ||
146 | The current head of the GCC branch doesn't co-operate well with GDB | |
147 | over debug information. | |
148 | ||
149 | Regressions problem (200 failures) | |
150 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00475.html | |
151 | ||
152 | -- | |
153 | ||
154 | RFA: infrun.c, breakpoint.c: Kludge for Solaris x86 hardware watchpoint support | |
155 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00664.html | |
156 | ||
157 | Unfortunately I'd need the following kludge to work around a Solaris | |
158 | x86 kernel problem with hardware watchpoint support. See the comment | |
159 | in the patches for a description of the problem. | |
160 | ||
161 | -- | |
162 | ||
163 | RFD: infrun.c: No bpstat_stop_status call after proceed over break ? | |
164 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00665.html | |
165 | ||
166 | I am currently trying to fix a GDB bug with missing watchpoint triggers | |
167 | after proceeding over a breakpoint on x86 targets. | |
168 | ||
169 | -- | |
170 | ||
67edb2c6 AC |
171 | x86 linux GDB and SIGALRM (???) |
172 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00803.html | |
173 | ||
174 | -- | |
175 | ||
176 | Migrate qfThreadInfo packet -> qThreadInfo. (Andrew Cagney) | |
177 | ||
178 | Add support for packet enable/disable commands with these thread | |
179 | packets. General cleanup. | |
180 | ||
181 | [PATCH] Document the ThreadInfo remote protocol queries | |
182 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00832.html | |
183 | ||
184 | [PATCH] "info threads" queries for remote.c | |
185 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00831.html | |
186 | ||
187 | -- | |
188 | ||
189 | MI documentation in GDB user guide. (Andrew Cagney, Elena Zannoni, | |
190 | Stan Shebs, anyone else?) | |
191 | ||
192 | > (Are there plans to make gdbmi.texi be part of the manual as well?) | |
193 | ||
194 | I'd like to see it go in there sooner rather than later too. Otherwise | |
195 | you're introducing discrepancies between the manual and the documentation, | |
196 | and everybody is confused - witness the lack of doc for the tracing | |
197 | commands still, some two years after they were added... | |
198 | ||
9d6d78f2 AC |
199 | Discussion on MI can be found on the thread: [PATCH] GDB command-line |
200 | switches and annotations docs | |
67edb2c6 AC |
201 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00639.html |
202 | ||
203 | -- | |
204 | ||
9d6d78f2 AC |
205 | Revised UDP support (was: Re: [Fwd: [patch] UDP transport support]) |
206 | http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00000.html | |
207 | ||
208 | -- | |
209 | ||
210 | problems loading shared libraries - with attached test case | |
211 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00820.html | |
212 | ||
213 | Hi, I'm having problems loading shared libraries. This is with a | |
214 | build of gdb out of cvs that I pulled and built on March 27th and has | |
215 | been there for at least a week. I haven't gone back further than | |
216 | that. This is with the gcc that is shipping with Red Hat 6.2: | |
217 | ||
218 | Reading specs from | |
219 | /usr/lib/gcc-lib/i386-redhat-linux/egcs-2.91.66/specs gcc version | |
220 | egcs-2.91.66 19990314/Linux (egcs-1.1.2 release) | |
221 | ||
222 | I'm using "set auto-solib-add 0" after main has been called. If I use | |
223 | "shar" to load a shared library manually once I can't use it again to | |
224 | load another shared library later. Please see the attached log for an | |
225 | example of how to reproduce the problem. | |
226 | ||
227 | -- | |
228 | ||
138f88c0 AC |
229 | GDB 5.0: Won't have |
230 | ------------------- | |
231 | ||
232 | The following are on hold until GDB 5.0 is branched. In general they | |
233 | won't go in as they unsettle the GDB sources. | |
234 | ||
235 | -- | |
236 | ||
237 | ATTRIBUTE_UNUSED | |
238 | ||
239 | The need for this as almost been eliminated. The next version of GCC | |
240 | (assuming cagney gets the revised patch approved) will be able to | |
241 | supress unused parameter warnings. | |
242 | ||
243 | -- | |
244 | ||
245 | Delete macro TARGET_BYTE_ORDER_SELECTABLE. | |
246 | ||
247 | Patches in the database. | |
248 | ||
249 | -- | |
250 | ||
251 | Updated readline | |
252 | ||
253 | Readline 4.? is out. A merge wouldn't hurt. | |
254 | ||
255 | -- | |
256 | ||
257 | Purge PARAMS | |
258 | ||
259 | Something to do post 5.0 branch | |
260 | ||
261 | -- | |
262 | ||
67edb2c6 AC |
263 | Elimination of make_cleanup_func. (Andrew Cagney) |
264 | ||
265 | make_cleanup_func elimination | |
266 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00791.html | |
267 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00814.html | |
268 | ||
269 | -- | |
270 | ||
67edb2c6 AC |
271 | ChangeLog.mi vs ChangeLog-mi (Andrew Cagney) |
272 | Needs further debate. | |
273 | ||
274 | Re: [PATCH] Add change-log variables to more MI files | |
275 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00811.html | |
276 | ||
277 | -- | |
278 | ||
9d6d78f2 AC |
279 | Re: gdb-cvs fails on freebsd-elf |
280 | http://sourceware.cygnus.com/ml/gdb/2000-04/msg00004.html | |
281 | ||
282 | FreeBSD haven't contributed their local GDB changes back to the master | |
283 | sources (they would at least need an FSF assignment by all | |
284 | individuales that contributed to the work). Given the strong | |
285 | likelhood that this will never happen, I'd suggest that a better | |
286 | strategy would be for someone (with an FSF/GDD assignment) to do a new | |
287 | (clean-room) implementation. That can then be accepted in time for | |
288 | GDB 5.1. | |
289 | ||
290 | -- | |
291 | ||
eb3d1d19 DB |
292 | Re: Various C++ things |
293 | ||
294 | value_headof/value_from_vtable_info are worthless, and should be removed. | |
295 | The one place in printcmd.c that uses it should use the RTTI functions. | |
296 | ||
297 | RTTI for g++ should be using the typeinfo functions rather than the vtables. | |
298 | The typeinfo functions are always at offset 4 from the beginning of the vtable, | |
299 | and are always right. The vtables will have weird names like E::VB sometimes. | |
300 | The typeinfo function will always be "E type_info function", or somesuch. | |
301 | ||
302 | value_virtual_fn_field needs to be fixed so there are no failures for virtual | |
303 | functions for C++ using g++. | |
304 | ||
305 | Testsuite cases are the major priority right now for C++ support, since i have | |
306 | to make a lot of changes that could potentially break each other. | |
307 | ||
308 | ||
309 | -- | |
138f88c0 AC |
310 | GDB 5.0: Test results |
311 | --------------------- | |
312 | ||
313 | Please include: | |
314 | ||
315 | o the output of `config.guess` | |
316 | o the date | |
317 | o the compiler | |
318 | o a note mentioning the reason | |
319 | for any serious failures. | |
320 | ||
321 | -- | |
322 | ||
323 | alpha-dec-osf4.0a, vendor compiler, 2000-03-04 | |
324 | ||
325 | Still has many compile warnings (mostly relating back to PTR vs void*) | |
326 | but it did compile using: | |
327 | ||
328 | CC=cc .../configure | |
329 | make | |
330 | ||
331 | Test results are: | |
332 | ||
333 | # of expected passes 6223 | |
334 | # of unexpected failures 103 | |
335 | # of unexpected successes 2 | |
336 | # of expected failures 196 | |
337 | # of unresolved testcases 6 | |
338 | # of unsupported tests 1 | |
339 | ||
340 | Looking at the output it would appear that GDB is stepping into some | |
341 | functions instead of ``next'' ing over them: | |
342 | ||
343 | 35 dummy(); | |
344 | (gdb) next | |
345 | dummy () at /home/cagney/GDB-DEJAGNU/src/gdb/testsuite/gdb.base/all-types.c:41 | |
346 | 41 { | |
347 | ||
348 | Since there is no active maintainer, I'd consider this sufficient for | |
349 | 5.0 :-/ | |
350 | ||
351 | -- | |
352 | ||
353 | sparc-sun-solaris2.6, egcs-2.91.66, 2000-02-10 | |
354 | http://sourceware.cygnus.com/ml/gdb-testers/2000-q1/msg00030.html | |
355 | ||
356 | There is a SIGTRAP problem that occures in ptrace.exp (Cagney to | |
357 | expand on). | |
358 | ||
359 | # of expected passes 6420 | |
360 | # of unexpected failures 7 | |
361 | # of expected failures 199 | |
362 | ||
363 | -- | |
364 | ||
365 | solaris 2.5.1 sparc?, 2.9-gnupro-99r1, 2000-02-10 | |
366 | http://sourceware.cygnus.com/ml/gdb-testers/2000-q1/msg00032.html | |
367 | ||
368 | # of expected passes 6420 | |
369 | # of unexpected failures 6 | |
370 | # of expected failures 199 | |
371 | ||
372 | -- | |
373 | ||
374 | sparc-unknown-netbsdelf1.4P, egcs-1.1.2+, 2000-03-01 | |
375 | ||
376 | This is with a very recent kernel. | |
377 | ||
378 | # of expected passes 6055 | |
379 | # of unexpected failures 88 | |
380 | # of unexpected successes 1 | |
381 | # of expected failures 190 | |
382 | # of unresolved testcases 59 | |
383 | ||
384 | -- | |
385 | ||
386 | GNU/Linux PPC | |
387 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00185.html | |
388 | ||
389 | Kevins merged it all in. | |
390 | ||
391 | -- | |
392 | ||
393 | Unixware | |
394 | ||
395 | Builds ok. Problems with some of the thread code. Unfortunate but | |
396 | not a show stopper. Nick D's still looking at it. | |
397 | ||
398 | Re: uw-threads issues | |
399 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00025.html | |
400 | ||
401 | ||
402 | ------------------------------------------------ | |
403 | ||
404 | ||
405 | General Wish List | |
406 | ================= | |
407 | ||
67edb2c6 AC |
408 | -- |
409 | ||
410 | GDBARCH cleanup (Andrew Cagney) | |
411 | ||
412 | The non-generated parts of gdbarch.{sh,h,c} should be separated out | |
413 | into gdbarch-utils.[hc] (Name ok). | |
414 | ||
415 | The ``info architecture'' command should be replaced with a fixed | |
416 | ``set architecture'' (implemented using the command.c enum code). | |
417 | ||
418 | Document that gdbarch_init_ftype could easily fail because it didn't | |
419 | identify an architecture. | |
420 | ||
421 | -- | |
422 | ||
423 | Check that GDB can handle all BFD architectures (Andrew Cagney) | |
424 | ||
425 | There should be a test that checks that BFD/GDB are in sync with | |
426 | regard to architecture changes. Something like a test that first | |
427 | queries GDB for all supported architectures and then feeds each back | |
428 | to GDB.. Anyone interested in learning how to write tests? :-) | |
429 | ||
430 | -- | |
431 | ||
c906108c SS |
432 | This list is probably not up to date, and opinions vary about the |
433 | importance or even desirability of some of the items. | |
434 | ||
b83266a0 SS |
435 | Document trace machinery. |
436 | ||
437 | Document overlay machinery. | |
438 | ||
439 | Extend .gdbinit mechanism to specify name on command line, allow for | |
440 | lists of files to load, include function of --tclcommand. | |
441 | ||
442 | @c This does not work (yet if ever). FIXME. | |
443 | @c @item --parse=@var{lang} @dots{} | |
444 | @c Configure the @value{GDBN} expression parser to parse the listed languages. | |
445 | @c @samp{all} configures @value{GDBN} for all supported languages. To get a | |
446 | @c list of all supported languages, omit the argument. Without this | |
447 | @c option, @value{GDBN} is configured to parse all supported languages. | |
448 | ||
c906108c SS |
449 | Add an "info bfd" command that displays supported object formats, |
450 | similarly to objdump -i. | |
451 | ||
452 | START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that | |
453 | is its default value. Clean this up. | |
454 | ||
455 | It should be possible to use symbols from shared libraries before we know | |
456 | exactly where the libraries will be loaded. E.g. "b perror" before running | |
457 | the program. This could maybe be done as an extension of the "breakpoint | |
458 | re-evaluation" after new symbols are loaded. | |
459 | ||
460 | Make single_step() insert and remove breakpoints in one operation. | |
461 | ||
462 | Speed up single stepping by avoiding extraneous ptrace calls. | |
463 | ||
464 | Speed up single stepping by not inserting and removing breakpoints | |
465 | each time the inferior starts and stops. | |
466 | ||
467 | Breakpoints should not be inserted and deleted all the time. Only the | |
468 | one(s) there should be removed when we have to step over one. Support | |
469 | breakpoints that don't have to be removed to step over them. | |
470 | ||
471 | Update gdbint.texinfo to include doc on the directory structure and | |
472 | the various tricks of building gdb. | |
473 | ||
474 | Do a tutorial in gdb.texinfo on how to do simple things in gdb. | |
475 | E.g. how to set a breakpoint that just prints something and continues. | |
476 | How to break on aborts. Etc. | |
477 | ||
478 | Provide "voodoo" debugging of core files. This creates a zombie | |
479 | process as a child of the debugger, and loads it up with the data, | |
480 | stack, and regs of the core file. This allows you to call functions | |
481 | in the executable, to manipulate the data in the core file. | |
482 | ||
483 | GDB reopens the source file on every line, as you "next" through it. | |
484 | ||
485 | Referencing the vtbl member of a struct doesn't work. It prints OK | |
486 | if you print the struct, but it gets 0 if you try to deref it. | |
487 | ||
488 | Persistent command history: A feature where you could save off a list | |
489 | of the commands you did, so you can edit it into something that will bring | |
490 | the target to the same place every time you source it. | |
491 | This would also be useful for automated fast watchpointing; if you go | |
492 | past the place where it watchpoints, you just start it over again and | |
493 | do it more carefully. | |
494 | ||
495 | Deal with the SunOS 4.0 and 4.1.1 ptrace bug that loses the registers if | |
496 | the stack is paged out. | |
497 | ||
498 | Finish the C++ exception handling stub routines. Lint points them out | |
499 | as unused statics functions. | |
500 | ||
501 | Perhaps "i source" should take an argument like that of "list". | |
502 | ||
503 | See if core-aout.c's fetch_core_registers can be used on more machines. | |
504 | E.g. MIPS (mips-xdep.c). | |
505 | ||
506 | unpack_double() does not handle IEEE float on the target unless the host | |
507 | is also IEEE. Death on a vax. | |
508 | ||
509 | Set up interface between GDB and INFO so that you can hop into interactive | |
510 | INFO and back out again. When running under Emacs, should use Emacs | |
511 | info, else fork the info program. Installation of GDB should install | |
512 | its texinfo files into the info tree automagically, including the readline | |
513 | texinfo files. | |
514 | ||
515 | "help address" ought to find the "help set print address" entry. | |
516 | ||
517 | Remove the VTBL internal guts from printouts of C++ structs, unless | |
518 | vtblprint is set. | |
519 | ||
520 | Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if | |
521 | it matches the source line indicated. | |
522 | ||
523 | The prompt at end of screen should accept space as well as CR. | |
524 | ||
525 | Check STORE_RETURN_VALUE on all architectures. Check near it in tm-sparc.h | |
526 | for other bogosities. | |
527 | ||
528 | Check for storage leaks in GDB, I'm sure there are a lot! | |
529 | ||
530 | vtblprint of a vtbl should demangle the names it's printing. | |
531 | ||
532 | Backtrace should point out what the currently selected frame is, in | |
533 | its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar, | |
534 | ...)" rather than "#3 foo (bar, ...)". | |
535 | ||
536 | "i program" should work for core files, and display more info, like what | |
537 | actually caused it to die. | |
538 | ||
539 | "x/10i" should shorten the long name, if any, on subsequent lines. | |
540 | ||
541 | Check through the code for FIXME comments and fix them. dbxread.c, | |
542 | blockframe.c, and plenty more. (I count 634 as of 940621 - sts) | |
543 | ||
544 | "next" over a function that longjumps, never stops until next time you happen | |
545 | to get to that spot by accident. E.g. "n" over execute_command which has | |
546 | an error. | |
547 | ||
548 | "set zeroprint off", don't bother printing members of structs which | |
549 | are entirely zero. Useful for those big structs with few useful | |
550 | members. | |
551 | ||
552 | GDB does four ioctl's for every command, probably switching terminal modes | |
553 | to/from inferior or for readline or something. | |
554 | ||
555 | terminal_ours versus terminal_inferior: cache state. Switch should be a noop | |
556 | if the state is the same, too. | |
557 | ||
558 | ptype $i6 = void??! | |
559 | ||
560 | Clean up invalid_float handling so gdb doesn't coredump when it tries to | |
561 | access a NaN. While this might work on SPARC, other machines are not | |
562 | configured right. | |
563 | ||
564 | "b value_at ; commands ; continue ; end" stops EVERY OTHER TIME! | |
565 | Then once you enter a command, it does the command, runs two more | |
566 | times, and then stops again! Bizarre... (This behaviour has been | |
567 | modified, but it is not yet 100% predictable when e.g. the commands | |
568 | call functions in the child, and while there, the child is interrupted | |
569 | with a signal, or hits a breakpoint.) | |
570 | ||
571 | help completion, help history should work. | |
572 | ||
573 | Check that we can handle stack trace through varargs AND alloca in same | |
574 | function, on 29K. | |
575 | ||
576 | wait_for_inferior loops forever if wait() gives it an error. | |
577 | ||
578 | "i frame" shows wrong "arglist at" location, doesn't show where the args | |
579 | should be found, only their actual values. | |
580 | ||
581 | There should be a way for "set" commands to validate the new setting | |
582 | before it takes effect. | |
583 | ||
584 | A mess of floating point opcodes are missing from sparc-opcode.h. | |
585 | Also, a little program should test the table for bits that are | |
586 | overspecified or underspecified. E.g. if the must-be-ones bits | |
587 | and the must-be-zeroes bits leave some fields unexamined, and the format | |
588 | string leaves them unprinted, then point this out. If multiple | |
589 | non-alias patterns match, point this out too. Finally, there should | |
590 | be a sparc-optest.s file that tries each pattern out. This file | |
591 | should end up coming back the same (modulo transformation comments) | |
592 | if fed to "gas" then the .o is fed to gdb for disassembly. | |
593 | ||
594 | Eliminate all the core_file_command's in all the xdep files. | |
595 | Eliminate separate declarations of registers[] everywhere. | |
596 | ||
597 | "ena d" is ambiguous, why? "ena delete" seems to think it is a command! | |
598 | ||
599 | Perhaps move the tdep, xdep, and nat files, into the config | |
600 | subdirectories. If not, at least straighten out their names so that | |
601 | they all start with the machine name. | |
602 | ||
603 | inferior_status should include stop_print_frame. It won't need to be | |
604 | reset in wait_for_inferior after bpstat_stop_status call, then. | |
605 | ||
606 | i line VAR produces "Line number not known for symbol ``var''.". I | |
607 | thought we were stashing that info now! | |
608 | ||
609 | We should be able to write to random files at hex offsets like adb. | |
610 | ||
611 | Make "target xxx" command interruptible. | |
612 | ||
613 | Handle add_file with separate text, data, and bss addresses. Maybe | |
614 | handle separate addresses for each segment in the object file? | |
615 | ||
616 | Handle free_named_symtab to cope with multiply-loaded object files | |
617 | in a dynamic linking environment. Should remember the last copy loaded, | |
618 | but not get too snowed if it finds references to the older copy. | |
619 | ||
620 | Generalize and Standardize the RPC interface to a target program, | |
621 | improve it beyond the "ptrace" interface, and see if it can become a | |
622 | standard for remote debugging. (This is talking about the vxworks | |
623 | interface. Seems unlikely to me that there will be "a standard" for | |
624 | remote debugging anytime soon --kingdon, 8 Nov 1994). | |
625 | ||
626 | Remove all references to: | |
627 | text_offset | |
628 | data_offset | |
629 | text_data_start | |
630 | text_end | |
631 | exec_data_offset | |
632 | ... | |
633 | now that we have BFD. All remaining are in machine dependent files. | |
634 | ||
635 | When quitting with a running program, if a core file was previously | |
636 | examined, you get "Couldn't read float regs from core file"...if | |
637 | indeed it can't. generic_mourn_inferior... | |
638 | ||
639 | Have remote targets give a warning on a signal argument to | |
640 | target_resume. Or better yet, extend the protocols so that it works | |
641 | like it does on the Unix-like systems. | |
642 | ||
643 | Sort help and info output. | |
644 | ||
645 | Re-organize help categories into things that tend to fit on a screen | |
646 | and hang together. | |
647 | ||
648 | renote-nindy.c handles interrupts poorly; it error()s out of badly | |
649 | chosen places, e.g. leaving current_frame zero, which causes core dumps | |
650 | on the next command. | |
651 | ||
652 | Add in commands like ADB's for searching for patterns, etc. We should | |
653 | be able to examine and patch raw unsymboled binaries as well in gdb as | |
654 | we can in adb. (E.g. increase the timeout in /bin/login without source). | |
655 | ||
656 | Those xdep files that call register_addr without defining it are | |
657 | probably simply broken. When reconfiguring this part of gdb, I could | |
658 | only make guesses about how to redo some of those files, and I | |
659 | probably guessed wrong, or left them "for later" when I have a | |
660 | machine that can attempt to build them. | |
661 | ||
662 | When doing "step" or "next", if a few lines of source are skipped between | |
663 | the previous line and the current one, print those lines, not just the | |
664 | last line of a multiline statement. | |
665 | ||
666 | When searching for C++ superclasses in value_cast in valops.c, we must | |
667 | not search the "fields", only the "superclasses". There might be a | |
668 | struct with a field name that matches the superclass name. This can | |
669 | happen when the struct was defined before the superclass (before the | |
670 | name became a typedef). | |
671 | ||
672 | Handling of "&" address-of operator needs some serious overhaul | |
673 | for ANSI C and consistency on arrays and functions. | |
674 | For "float point[15];": | |
675 | ptype &point[4] ==> Attempt to take address of non-lvalue. | |
676 | For "char *malloc();": | |
677 | ptype malloc ==> "char *()"; should be same as | |
678 | ptype &malloc ==> "char *(*)()" | |
679 | call printf ("%x\n", malloc) ==> weird value, should be same as | |
680 | call printf ("%x\n", &malloc) ==> correct value | |
681 | ||
682 | Fix dbxread.c symbol reading in the presence of interrupts. It | |
683 | currently leaves a cleanup to blow away the entire symbol table when a | |
684 | QUIT occurs. (What's wrong with that? -kingdon, 28 Oct 1993). | |
685 | ||
686 | Mipsread.c reads include files depth-first, because the dependencies | |
687 | in the psymtabs are way too inclusive (it seems to me). Figure out what | |
688 | really depends on what, to avoid recursing 20 or 30 times while reading | |
689 | real symtabs. | |
690 | ||
691 | value_add() should be subtracting the lower bound of arrays, if known, | |
692 | and possibly checking against the upper bound for error reporting. | |
693 | ||
694 | mipsread.c symbol table allocation and deallocation should be checked. | |
695 | My suspicion is that it's full of memory leaks. | |
696 | ||
697 | SunOS should have a target_lookup_symbol() for common'd things allocated | |
698 | by the shared library linker ld.so. | |
699 | ||
700 | When listing source lines, check for a preceding \n, to verify that | |
701 | the file hasn't changed out from under us. | |
702 | ||
703 | When listing source lines, eat leading whitespace corresponding to the | |
704 | line-number prefix we print. This avoids long lines wrapping. | |
705 | ||
706 | mipsread.c needs to check for old symtabs and psymtabs for the same | |
707 | files, the way it happens for dbxread.c and coffread.c, for VxWorks | |
708 | incremental symbol table reloading. | |
709 | ||
710 | Get all the remote systems (where the protocol allows it) to be able to | |
711 | stop the remote system when the GDB user types ^C (like remote.c | |
712 | does). For ebmon, use ^Ak. | |
713 | ||
714 | Possible feature: A version of the "disassemble" command which shows | |
715 | both source and assembly code ("set symbol-filename on" is a partial | |
716 | solution). | |
717 | ||
718 | investigate "x/s 0" (right now stops early) (I think maybe GDB is | |
719 | using a 0 address for bad purposes internally). | |
720 | ||
721 | Make "info path" and path_command work again (but independent of the | |
722 | environment either of gdb or that we'll pass to the inferior). | |
723 | ||
724 | Make GDB understand the GCC feature for putting octal constants in | |
725 | enums. Make it so overflow on an enum constant does not error_type | |
726 | the whole type. Allow arbitrarily large enums with type attributes. | |
727 | Put all this stuff in the testsuite. | |
728 | ||
729 | Make TYPE_CODE_ERROR with a non-zero TYPE_LENGTH more useful (print | |
730 | the value in hex; process type attributes). Add this to the | |
731 | testsuite. This way future compilers can add new types and old | |
732 | versions of GDB can do something halfway reasonable. | |
733 | ||
734 | Clean up formatting of "info registers" on MIPS and 88k. See if it | |
735 | is possible to do this generically across all target architectures. | |
736 | ||
737 | GDB gets bfd/corefile.c and gdb/corefile.c confused (this should be easy to | |
738 | repeat even with something more recent than GDB 4.9). | |
739 | ||
740 | Check that unmatched RBRAC doesn't abort(). | |
741 | ||
742 | Fix mdebugread.c:parse_type to do fundamental types right (see | |
743 | rs6000_builtin_type in stabsread.c for what "right" is--the point is | |
744 | that the debug format fixes the sizes of these things and it shouldn't | |
745 | depend on stuff like TARGET_PTR_BIT and so on. For mdebug, there seem | |
746 | to be separate bt* codes for 64 bit and 32 bit things, and GDB should | |
747 | be aware of that). Also use a switch statement for clarity and speed. | |
748 | ||
749 | Investigate adding symbols in target_load--some targets do, some | |
750 | don't. | |
751 | ||
752 | Put dirname in psymtabs and change lookup*symtab to use dirname (so | |
753 | /foo/bar.c works whether compiled by cc /foo/bar.c, or cd /foo; cc | |
754 | bar.c). | |
755 | ||
756 | Merge xcoffread.c and coffread.c. Use breakpoint_re_set instead of | |
757 | fixup_breakpoints. | |
758 | ||
759 | Fix byte order and int size sins in tm-a29k.h | |
760 | (EXTRACT_RETURN_VALUE). Perhaps should reproduce bug and verify fix | |
761 | (or perhaps should just fix it...). | |
762 | ||
763 | Make a watchpoint on a constant expression an error (or warning | |
764 | perhaps) | |
765 | ||
766 | Make a watchpoint which contains a function call an error (it is | |
767 | broken now, making it work is probably not worth the effort). | |
768 | ||
769 | Re-do calls to signal() in remote.c, and inflow.c (set_sigint_trap and | |
770 | so on) to be independent of the debugging target, using target_stop to | |
771 | stop the inferior. Probably the part which is now handled by | |
772 | interrupt_query in remote.c can be done without any new features in | |
773 | the debugging target. | |
774 | ||
775 | New test case based on weird.exp but in which type numbers are not | |
776 | renumbered (thus multiply defining a type). This currently causes an | |
777 | infinite loop on "p v_comb". | |
778 | ||
779 | Nuke baseclass_addr. | |
780 | ||
781 | Nuke USG define. | |
782 | ||
783 | "source file more recent" loses on re-read | |
784 | ||
785 | Fix 386 floating point so that floating point registers are real | |
786 | registers (but code can deal at run-time if they are missing, like | |
787 | mips and 68k). This would clean up "info float" and related stuff. | |
788 | ||
789 | Look at Solaris bug in interrupt.exp. Can get out of syscall with | |
790 | PRSABORT (syscall will return EINTR) but merely doing that leads to a | |
791 | "can't read memory" error. | |
792 | ||
793 | gcc -g -c enummask.c then gdb enummask.o, then "p v". GDB complains | |
794 | about not being able to access memory location 0. | |
795 | ||
796 | -------------------- enummask.c | |
797 | enum mask | |
798 | { | |
799 | ANIMAL = 0, | |
800 | VEGETABLE = 1, | |
801 | MINERAL = 2, | |
802 | BASIC_CATEGORY = 3, | |
803 | ||
804 | WHITE = 0, | |
805 | BLUE = 4, | |
806 | GREEN = 8, | |
807 | BLACK = 0xc, | |
808 | COLOR = 0xc, | |
809 | ||
810 | ALIVE = 0x10, | |
811 | ||
812 | LARGE = 0x20 | |
813 | } v; | |
814 | ||
815 | If try to modify value in file with "set write off" should give | |
816 | appropriate error not "cannot access memory at address 0x65e0". | |
817 | ||
818 | Why do we allow a target to omit standard register names (NO_STD_REGS | |
819 | in tm-z8k.h)? I thought the standard register names were supposed to | |
820 | be just that, standard. | |
821 | ||
c906108c SS |
822 | Allow core file without exec file on RS/6000. |
823 | ||
824 | Make sure "shell" with no arguments works right on DOS. | |
825 | ||
826 | Make gdb.ini (as well as .gdbinit) be checked on all platforms, so | |
827 | the same directory can be NFS-mounted on unix or DOS, and work the | |
828 | same way. | |
829 | ||
830 | cd ~/tmp/<M-?> causes infinite loop (where ~/tmp is a directory). | |
831 | ||
832 | Get SECT_OFF_TEXT stuff out of objfile_relocate (might be needed to | |
833 | get RS/6000 to work right, might not be immediately relevant). | |
834 | ||
835 | Clean up add_toc_to_loadinfo | |
836 | ||
837 | Think about attached processes and sharing terminal. | |
838 | ||
839 | John sez in reference to ignoring errors from tcsegpgrp if attach_flag: | |
840 | set_tty_state should not have any trouble with attached processes. | |
841 | Instead, the tty handling should leave the pgrp of the tty alone when | |
842 | attaching to processes (perhaps pass terminal_init_inferior a flag | |
843 | saying whether we're attaching). | |
844 | ||
845 | PAGE_SIZE redefined warnings on AIX. Probably should be using | |
846 | BFD_PAGE_SIZE throughout BFD. | |
847 | ||
848 | Rewrite proceed, wait_for_inferior, and normal_stop to clean them up. | |
849 | Suggestions: | |
850 | ||
851 | 1) Make each test in wait_for_inferior a seperate subroutine | |
852 | call. | |
853 | 2) Combine wait_for_inferior and normal_stop to clean up | |
854 | communication via global variables. | |
855 | 3) See if you can find some way to clean up the global | |
856 | variables that are used; possibly group them by data flow | |
857 | and information content? | |
858 | ||
859 | Work out some kind of way to allow running the inferior to be done as | |
860 | a sub-execution of, eg. breakpoint command lists. Currently running | |
861 | the inferior interupts any command list execution. This would require | |
862 | some rewriting of wait_for_inferior & friends, and hence should | |
863 | probably be done in concert with the above. | |
864 | ||
865 | Add function arguments to gdb user defined functions. | |
866 | ||
867 | Add convenience variables that refer to exec file, symbol file, | |
868 | selected frame source file, selected frame function, selected frame | |
869 | line number, etc. | |
870 | ||
871 | Add a "suspend" subcommand of the "continue" command to suspend gdb | |
872 | while continuing execution of the subprocess. Useful when you are | |
873 | debugging servers and you want to dodge out and initiate a connection | |
874 | to a server running under gdb. | |
875 | ||
876 | Add stab information to allow reasonable debugging of inline functions | |
877 | (possibly they should show up on a stack backtrace? With a note | |
878 | indicating that they weren't "real"?). | |
879 | ||
880 | Modify the naked "until" command to step until past the current source | |
881 | line, rather than past the current pc value. This is tricky simply | |
882 | because the low level routines have no way of specifying a multi-line | |
883 | step range, and there is no way of saying "don't print stuff when we | |
884 | stop" from above (otherwise could just call step many times). | |
885 | ||
886 | Modify the handling of symbols grouped through BINCL/EINCL stabs to | |
887 | allocate a partial symtab for each BINCL/EINCL grouping. This will | |
888 | seriously decrease the size of inter-psymtab dependencies and hence | |
889 | lessen the amount that needs to be read in when a new source file is | |
890 | accessed. | |
891 | ||
892 | Do an "x/i $pc" after each stepi or nexti. | |
893 | ||
894 | Modify all of the disassemblers to use printf_filtered to get correct | |
895 | more filtering. | |
896 | ||
897 | Modify gdb to work correctly with Pascal. | |
898 | ||
899 | Add a command for searching memory, a la adb. It specifies size, | |
900 | mask, value, start address. ADB searches until it finds it or hits | |
901 | an error (or is interrupted). | |
902 | ||
b83266a0 SS |
903 | Remove the range and type checking code and documentation, if not |
904 | going to implement. | |
905 | ||
c906108c SS |
906 | # Local Variables: |
907 | # mode: text | |
908 | # End: |