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