Commit | Line | Data |
---|---|---|
c906108c | 1 | If you find inaccuracies in this list, please send mail to |
2a00c9ce AC |
2 | gdb-patches@sourceware.cygnus.com. If you would like to work on any |
3 | of these, you should consider sending mail to the same address, to | |
4 | find out whether anyone else is working on it. | |
c906108c | 5 | |
138f88c0 | 6 | |
4afc966c AC |
7 | Known problems in GDB 5.0 |
8 | ========================= | |
138f88c0 | 9 | |
bc9e5bbf AC |
10 | Below is a list of problems identified during the GDB 5.0 release |
11 | cycle. People hope to have these problems fixed in a follow-on | |
12 | release. | |
138f88c0 | 13 | |
138f88c0 AC |
14 | -- |
15 | ||
4fd99b5a AC |
16 | The BFD directory requires bug-fixed AUTOMAKE et.al. |
17 | ||
18 | AUTOMAKE 1.4 incorrectly set the TEXINPUTS environment variable. It | |
19 | contained the full path to texinfo.tex when it should have only | |
20 | contained the directory. The bug has been fixed in the current | |
21 | AUTOMAKE sources. Automake snapshots can be found in: | |
22 | ftp://sourceware.cygnus.com/pub/gdb/snapshots | |
23 | and ftp://sourceware.cygnus.com/pub/binutils | |
24 | ||
25 | -- | |
26 | ||
bc9e5bbf | 27 | RFD: infrun.c: No bpstat_stop_status call after proceed over break? |
138f88c0 AC |
28 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00665.html |
29 | ||
bc9e5bbf AC |
30 | GDB misses watchpoint triggers after proceeding over a breakpoint on |
31 | x86 targets. | |
32 | ||
138f88c0 AC |
33 | -- |
34 | ||
67edb2c6 AC |
35 | x86 linux GDB and SIGALRM (???) |
36 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00803.html | |
37 | ||
37d4dc74 MK |
38 | This problem has been fixed, but a regression test still needs to be |
39 | added to the testsuite: | |
40 | http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00309.html | |
67edb2c6 | 41 | |
bc9e5bbf | 42 | Mark |
67edb2c6 AC |
43 | |
44 | -- | |
45 | ||
9d6d78f2 AC |
46 | Revised UDP support (was: Re: [Fwd: [patch] UDP transport support]) |
47 | http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00000.html | |
48 | ||
bc9e5bbf AC |
49 | (Broken) support for GDB's remote protocol across UDP is to be |
50 | included in the follow-on release. | |
138f88c0 | 51 | |
b2f4b24d AC |
52 | -- |
53 | ||
54 | Can't build IRIX -> arm GDB. | |
55 | http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00356.html | |
56 | ||
57 | David Whedon writes: | |
58 | > Now I'm building for an embedded arm target. If there is a way of turning | |
59 | > remote-rdi off, I couldn't find it. It looks like it gets built by default | |
60 | > in gdb/configure.tgt(line 58) Anyway, the build dies in | |
61 | > gdb/rdi-share/unixcomm.c. SERPORT1 et. al. never get defined because we | |
62 | > aren't one of the architectures supported. | |
63 | ||
6bc37a96 AC |
64 | -- |
65 | ||
66 | Problem with weak functions | |
67 | http://sourceware.cygnus.com/ml/gdb/2000-05/msg00060.html | |
68 | ||
69 | Dan Nicolaescu writes: | |
70 | > It seems that gdb-4.95.1 does not display correctly the function when | |
71 | > stoping in weak functions. | |
72 | > | |
73 | > It stops in a function that is defined as weak, not in the function | |
74 | > that is actualy run... | |
b2f4b24d | 75 | |
26099b4a AC |
76 | -- |
77 | ||
78 | GDB 5.0 doesn't work on Linux/SPARC | |
79 | ||
3fffcb5e AC |
80 | -- |
81 | ||
4afc966c AC |
82 | Code Cleanups: Next Release |
83 | =========================== | |
3fffcb5e | 84 | |
4afc966c AC |
85 | The following are small cleanups that will hopefully be completed by |
86 | the follow on to 5.0. | |
3fffcb5e | 87 | |
138f88c0 AC |
88 | -- |
89 | ||
4afc966c | 90 | Delete macro TARGET_BYTE_ORDER_SELECTABLE. |
bc9e5bbf | 91 | |
4afc966c AC |
92 | Patches in the database. |
93 | ||
94 | -- | |
95 | ||
26099b4a | 96 | Purge PARAMS. |
4afc966c AC |
97 | |
98 | Eliminate all uses of PARAMS in GDB's source code. | |
99 | ||
100 | -- | |
101 | ||
78566ebe AC |
102 | Fix copyright notices. |
103 | ||
104 | Turns out that ``1998-2000'' isn't considered valid :-( | |
105 | ||
106 | http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00467.html | |
107 | ||
1ba607ad AC |
108 | -- |
109 | ||
110 | Find something better than DEFAULT_BFD_ARCH, DEFAULT_BFD_VEC to | |
111 | determine the default isa/byte-order. | |
112 | ||
113 | -- | |
114 | ||
115 | Rely on BFD_BIG_ENDIAN and BFD_LITTLE_ENDIAN instead of host dependant | |
116 | BIG_ENDIAN and LITTLE_ENDIAN. | |
117 | ||
4afc966c AC |
118 | -- |
119 | ||
120 | Code Cleanups: General | |
121 | ====================== | |
122 | ||
123 | The following are more general cleanups and fixes. They are not tied | |
124 | to any specific release. | |
bc9e5bbf AC |
125 | |
126 | -- | |
127 | ||
d8038014 AC |
128 | Eliminate more compiler warnings. |
129 | ||
130 | Of course there also needs to be the usual debate over which warnings | |
131 | are valid and how to best go about this. | |
132 | ||
133 | One method: choose a single option; get agreement that it is | |
134 | reasonable; try it out to see if there isn't anything silly about it | |
135 | (-Wunused-parameters is an example of that) then incrementally hack | |
136 | away. | |
137 | ||
138 | The other method is to enable all warnings and eliminate them from one | |
139 | file at a time. | |
140 | ||
141 | -- | |
142 | ||
4afc966c | 143 | Elimination of ``(catch_errors_ftype *) func''. |
bc9e5bbf | 144 | |
4afc966c | 145 | Like make_cleanup_func it isn't portable. |
6ecce94d AC |
146 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00791.html |
147 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00814.html | |
148 | ||
4afc966c AC |
149 | |
150 | -- | |
151 | ||
152 | Nuke USG define. | |
153 | ||
154 | -- | |
155 | ||
156 | [PATCH/5] src/intl/Makefile.in:distclean additions | |
157 | http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00363.html | |
158 | ||
159 | Do not forget to merge the patch back into the trunk. | |
160 | ||
161 | -- | |
162 | ||
163 | Rationalize the host-endian code (grep for HOST_BYTE_ORDER). | |
164 | ||
165 | At present defs.h includes <endian.h> (which is linux specific) yet | |
166 | almost nothing depends on it. Suggest "gdb_endian.h" which can also | |
167 | handle <machine/endian.h> and only include that where it is really | |
168 | needed. | |
169 | ||
170 | -- | |
171 | ||
172 | Replace asprintf() calls with xasprintf() calls. | |
173 | ||
174 | As with things like strdup() most calls to asprintf() don't check the | |
175 | return value. | |
176 | ||
177 | -- | |
178 | ||
179 | Replace strsave() + mstrsave() with libiberty:xstrdup(). | |
180 | ||
181 | -- | |
182 | ||
183 | Replace savestring() with something from libiberty. | |
184 | ||
185 | An xstrldup()? but that would have different semantics. | |
186 | ||
187 | -- | |
188 | ||
189 | Rationalize use of floatformat_unknown in GDB sources. | |
190 | ||
191 | Instead of defaulting to floatformat_unknown, should hosts/targets | |
192 | specify the value explicitly? | |
193 | ||
194 | http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html | |
195 | ||
196 | -- | |
197 | ||
198 | Add a ``name'' member to include/floatformat.h:struct floatformat. | |
199 | Print that name in gdbarch.c. | |
200 | ||
201 | -- | |
202 | ||
203 | Sort out the harris mess in include/floatformat.h (it hardwires two | |
204 | different floating point formats). | |
205 | ||
206 | -- | |
207 | ||
208 | See of the GDB local floatformat_do_doublest() and libiberty's | |
209 | floatformat_to_double (which was once GDB's ...) can be merged some | |
210 | how. | |
211 | ||
212 | -- | |
213 | ||
214 | Eliminate mmalloc() from GDB. | |
215 | ||
216 | Also eliminate it from defs.h. | |
217 | ||
218 | -- | |
219 | ||
220 | Eliminate PTR. ISO-C allows ``void *''. | |
221 | ||
222 | -- | |
223 | ||
224 | Eliminate abort (). | |
225 | ||
226 | GDB should never abort. GDB should either throw ``error ()'' or | |
227 | ``internal_error ()''. Better still GDB should naturally unwind with | |
228 | an error status. | |
229 | ||
230 | -- | |
231 | ||
232 | GDB probably doesn't build on FreeBSD pre 2.2.x | |
233 | http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00378.html | |
234 | ||
235 | Fixes to get FreeBSD working on 2.2.x, 3.x and 4.x caused the code to | |
236 | suffer bit rot. | |
bc9e5bbf AC |
237 | |
238 | -- | |
239 | ||
78566ebe AC |
240 | Deprecate "fg". Apparently ``fg'' is actually continue. |
241 | ||
242 | http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00417.html | |
243 | ||
244 | -- | |
245 | ||
246 | Deprecate current use of ``floatformat_unknown''. | |
247 | ||
248 | Require all targets to explicitly provide their float format instead | |
249 | of defaulting to floatformat unknown. Doing the latter leads to nasty | |
250 | bugs. | |
251 | ||
252 | http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html | |
bc9e5bbf AC |
253 | |
254 | -- | |
255 | ||
78566ebe AC |
256 | Rationalize floatformat_to_double() vs floatformat_to_doublest(). |
257 | ||
258 | Looks like GDB migrated floatformat_to_double() to libiberty but then | |
259 | turned around and created a ..._to_doublest() the latter containing | |
260 | several bug fixes. | |
261 | ||
262 | http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00472.html | |
263 | ||
264 | -- | |
265 | ||
266 | Move floatformat_ia64_ext to libiberty/include floatformat.[ch]. | |
267 | ||
268 | http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00466.html | |
269 | ||
270 | -- | |
271 | ||
78566ebe AC |
272 | Follow through `make check' with --enable-shared. |
273 | ||
274 | When the srcware tree is configured with --enable-shared, the `expect' | |
275 | program won't run properly. Jim Wilson found out gdb has a local hack | |
276 | to set LD_LIBRARY_PATH, but, AFAIK, no other project has been hacked | |
277 | similarly. | |
278 | ||
279 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00845.html | |
280 | ||
b4a20239 AC |
281 | -- |
282 | ||
b4a20239 AC |
283 | The ``maintenance deprecate set endian big'' command doesn't notice |
284 | that it is deprecating ``set endian'' and not ``set endian big'' (big | |
285 | is implemented using an enum). Is anyone going to notice this? | |
286 | ||
287 | -- | |
288 | ||
289 | When tab expanding something like ``set arch<tab>'' ignore the | |
290 | deprecated ``set archdebug'' and expand to ``set architecture''. | |
291 | ||
78566ebe AC |
292 | -- |
293 | ||
53904c9e AC |
294 | Eliminate ``arm_register_names[j] = (char *) regnames[j]'' and the |
295 | like from arm-tdep.c. | |
296 | ||
297 | -- | |
298 | ||
299 | Fix uses of ->function.cfunc = set_function(). | |
300 | ||
301 | The command.c code calls sfunc() when a set command. Rather than | |
302 | change it suggest fixing the callback function so that it is more | |
303 | useful. See: | |
304 | ||
305 | http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html | |
306 | ||
307 | See also ``Fix implementation of ``target xxx''.'' below. | |
308 | ||
bf64bfd6 AC |
309 | -- |
310 | ||
311 | IRIX 3.x support is probably broken. | |
312 | ||
5d35f0ac AC |
313 | -- |
314 | ||
315 | Delete sim/SIM_HAVE_BREAKPOINTS and gdb/SIM_HAS_BREAKPOINTS. | |
316 | http://sourceware.cygnus.com/ml/gdb-patches/2000-07/msg00042.html | |
317 | ||
318 | Apart from the d30v, are there any sim/common simulators that make use | |
319 | of this? | |
320 | ||
321 | A brief summary of what happended is that sim/common/sim-break.c was | |
322 | created as a good idea. It turned out a better idea was to use | |
323 | SIM_SIGBREAK and have GDB pass back sim_resume (..., SIGBREAK). | |
324 | ||
325 | -- | |
326 | ||
327 | parse.c:build_parse() has a buffer overrun. | |
328 | ||
53904c9e | 329 | -- |
78566ebe | 330 | |
4afc966c AC |
331 | New Features and Fixes |
332 | ====================== | |
bc9e5bbf | 333 | |
4afc966c AC |
334 | These are harder than cleanups but easier than work involving |
335 | fundamental architectural change. | |
bc9e5bbf AC |
336 | |
337 | -- | |
338 | ||
4afc966c AC |
339 | Add built-by, build-date, tm, xm, nm and anything else into gdb binary |
340 | so that you can see how the GDB was created. | |
bc9e5bbf | 341 | |
4afc966c AC |
342 | Some of these (*m.h) would be added to the generated config.h. That |
343 | in turn would fix a long standing bug where by the build process many | |
344 | not notice a changed tm.h file. Since everything depends on config.h, | |
345 | a change to *m.h forces a change to config.h and, consequently forces | |
346 | a rebuild. | |
bc9e5bbf AC |
347 | |
348 | -- | |
349 | ||
4afc966c AC |
350 | Add an "info bfd" command that displays supported object formats, |
351 | similarly to objdump -i. | |
5683e87a | 352 | |
4afc966c | 353 | Is there a command already? |
5683e87a AC |
354 | |
355 | -- | |
356 | ||
4afc966c | 357 | Fix ``I'm sorry, Dave, I can't do that.'' from symfile.c. |
bc9e5bbf | 358 | |
4afc966c | 359 | This requires internationalization. |
bc9e5bbf | 360 | |
4afc966c | 361 | -- |
bc9e5bbf | 362 | |
4afc966c | 363 | Convert GDB build process to AUTOMAKE. |
bc9e5bbf | 364 | |
4afc966c | 365 | See also sub-directory configure below. |
138f88c0 | 366 | |
b4a20239 AC |
367 | The current convention is (kind of) to use $(<header>_h) in all |
368 | dependency lists. It isn't done in a consistent way. | |
369 | ||
67edb2c6 AC |
370 | -- |
371 | ||
4afc966c | 372 | Cleanup configury support for optional sub-directories. |
7ae38352 | 373 | |
4afc966c AC |
374 | Check how GCC handles multiple front ends for an example of how things |
375 | could work. A tentative first step is to rationalize things so that | |
376 | all sub directories are handled in a fashion similar to gdb/mi. | |
377 | ||
378 | See also automake above. | |
7ae38352 AC |
379 | |
380 | -- | |
381 | ||
4afc966c AC |
382 | Restructure gdb directory tree so that it avoids any 8.3 and 14 |
383 | filename problems. | |
67edb2c6 | 384 | |
4afc966c | 385 | -- |
67edb2c6 | 386 | |
4afc966c AC |
387 | Add a transcript mechanism to GDB. |
388 | ||
389 | Such a mechanism might log all gdb input and output to a file in a | |
390 | form that would allow it to be replayed. It could involve ``gdb | |
391 | --transcript=FILE'' or it could involve ``(gdb) transcript file''. | |
67edb2c6 AC |
392 | |
393 | -- | |
394 | ||
4afc966c | 395 | Can the xdep files be replaced by autoconf? |
bc9e5bbf | 396 | |
4afc966c | 397 | -- |
bc9e5bbf | 398 | |
4afc966c | 399 | Document trace machinery |
bc9e5bbf | 400 | |
4afc966c AC |
401 | -- |
402 | ||
78566ebe AC |
403 | Document ui-out and ui-file. |
404 | ||
405 | http://sourceware.cygnus.com/ml/gdb/2000-04/msg00121.html | |
406 | ||
407 | -- | |
408 | ||
409 | Update texinfo.tex to latest? | |
410 | ||
411 | ||
412 | ||
413 | -- | |
414 | ||
415 | Incorporate agentexpr.texi into gdb.texinfo | |
416 | ||
417 | agentexpr.texi mostly describes the details of the byte code used for | |
418 | tracepoints, not the internals of the support for this in GDB. So it | |
419 | looks like gdb.texinfo is a better place for this information. | |
420 | ||
421 | http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00566.html | |
422 | ||
423 | -- | |
424 | ||
4afc966c | 425 | Document overlay machinery. |
bc9e5bbf | 426 | |
7ae38352 AC |
427 | -- |
428 | ||
4afc966c | 429 | ``(gdb) catch signal SIGNAL'' |
7ae38352 | 430 | |
4afc966c AC |
431 | Overlaps with ``handle SIGNAL'' but the implied behavour is different. |
432 | You can attach commands to a catch but not a handle. A handle has a | |
433 | limited number of hardwired actions. | |
7ae38352 AC |
434 | |
435 | -- | |
436 | ||
4afc966c | 437 | Get the TUI working on all platforms. |
7ae38352 | 438 | |
bc9e5bbf AC |
439 | -- |
440 | ||
4afc966c AC |
441 | Add support for ``gdb --- PROGRAM ARGS ...''. |
442 | Add support for ``gdb -cmd=...'' | |
9debab2f | 443 | |
4afc966c AC |
444 | Along with many variations. Check: |
445 | ||
446 | ????? for a full discussion. | |
447 | ||
448 | for a discussion. | |
9debab2f AC |
449 | |
450 | -- | |
451 | ||
4afc966c | 452 | Implement ``(gdb) !ls''. |
e55e8cee | 453 | |
4afc966c AC |
454 | Which is very different from ``(gdb) ! ls''. Implementing the latter |
455 | is trivial. | |
456 | ||
457 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00034.html | |
e55e8cee AC |
458 | |
459 | -- | |
460 | ||
b4a20239 AC |
461 | Change the (char *list[]) to (const char (*)[]) so that dynamic lists can |
462 | be passed. | |
463 | ||
464 | -- | |
465 | ||
466 | When tab expanding something like ``set arch<tab>'' ignore the | |
467 | deprecated ``set archdebug'' and expand to ``set architecture''. | |
468 | ||
469 | -- | |
470 | ||
4afc966c AC |
471 | Replace the code that uses the host FPU with an emulator of the target |
472 | FPU. | |
7ae38352 | 473 | |
4afc966c AC |
474 | -- |
475 | ||
476 | Thread Support | |
477 | ============== | |
7ae38352 AC |
478 | |
479 | -- | |
480 | ||
4afc966c AC |
481 | Generic: lin-thread cannot handle thread exit (Mark Kettenis, Michael |
482 | Snyder) http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00525.html | |
7ae38352 | 483 | |
4afc966c AC |
484 | The thread_db assisted debugging code doesn't handle exiting threads |
485 | properly, at least in combination with glibc 2.1.3 (the framework is | |
486 | there, just not the actual code). There are at least two problems | |
487 | that prevent this from working. | |
488 | ||
489 | As an additional reference point, the pre thread_db code did not work | |
490 | either. | |
7ae38352 AC |
491 | |
492 | -- | |
493 | ||
4afc966c AC |
494 | GNU/Linux/x86 and random thread signals (and Solaris/SPARC but not |
495 | Solaris/x86). | |
496 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00336.html | |
7ae38352 | 497 | |
4afc966c AC |
498 | Christopher Blizzard writes: |
499 | ||
500 | So, I've done some more digging into this and it looks like Jim | |
501 | Kingdon has reported this problem in the past: | |
502 | ||
503 | http://sourceware.cygnus.com/ml/bug-gdb/1999-10/msg00058.html | |
504 | ||
505 | I can reproduce this problem both with and without Tom's patch. Has | |
506 | anyone seen this before? Maybe have a solution for it hanging around? | |
507 | :) | |
508 | ||
509 | There's a test case for this documented at: | |
510 | ||
511 | when debugging threaded applications you get extra SIGTRAPs | |
512 | http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=9565 | |
513 | ||
514 | [There should be a GDB testcase - cagney] | |
7ae38352 AC |
515 | |
516 | -- | |
517 | ||
4afc966c AC |
518 | GDB5 TOT on unixware 7 |
519 | http://sourceware.cygnus.com/ml/gdb/2000-04/msg00119.html | |
7ae38352 | 520 | |
4afc966c AC |
521 | Robert Lipe writes: |
522 | > I just spun the top of tree of the GDB5 branch on UnixWare 7. As a | |
523 | > practical matter, the current thread support is somewhat more annoying | |
524 | > than when GDB was thread-unaware. | |
7ae38352 AC |
525 | |
526 | -- | |
527 | ||
4afc966c | 528 | Migrate qfThreadInfo packet -> qThreadInfo. (Andrew Cagney) |
7ae38352 | 529 | |
4afc966c AC |
530 | Add support for packet enable/disable commands with these thread |
531 | packets. General cleanup. | |
532 | ||
533 | [PATCH] Document the ThreadInfo remote protocol queries | |
534 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00832.html | |
535 | ||
536 | [PATCH] "info threads" queries for remote.c | |
537 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00831.html | |
7ae38352 AC |
538 | |
539 | -- | |
540 | ||
4afc966c AC |
541 | Language Support |
542 | ================ | |
7ae38352 | 543 | |
4afc966c | 544 | New languages come onto the scene all the time. |
7ae38352 AC |
545 | |
546 | -- | |
547 | ||
4afc966c | 548 | Pascal (Pierre Muller, David Taylor) |
7ae38352 | 549 | |
4afc966c AC |
550 | Pierre Muller has contributed patches for adding Pascal Language |
551 | support to GDB. | |
552 | ||
553 | 2 pascal language patches inserted in database | |
554 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00521.html | |
555 | ||
556 | Indent -gnu ? | |
557 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00496.html | |
558 | ||
559 | -- | |
560 | ||
561 | Java (Anthony Green, David Taylor) | |
562 | ||
563 | Anthony Green has a number of Java patches that did not make it into | |
564 | the 5.0 release. | |
565 | ||
566 | Patch: java tests | |
567 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00512.html | |
568 | ||
569 | Patch: java booleans | |
570 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00515.html | |
571 | ||
572 | Patch: handle N_MAIN stab | |
573 | http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00527.html | |
574 | ||
575 | -- | |
576 | ||
577 | [Comming...] | |
578 | ||
579 | Modify gdb to work correctly with Pascal. | |
580 | ||
581 | -- | |
582 | ||
583 | Re: Various C++ things | |
584 | ||
26099b4a AC |
585 | value_headof/value_from_vtable_info are worthless, and should be |
586 | removed. The one place in printcmd.c that uses it should use the RTTI | |
587 | functions. | |
4afc966c | 588 | |
26099b4a AC |
589 | RTTI for g++ should be using the typeinfo functions rather than the |
590 | vtables. The typeinfo functions are always at offset 4 from the | |
591 | beginning of the vtable, and are always right. The vtables will have | |
592 | weird names like E::VB sometimes. The typeinfo function will always | |
593 | be "E type_info function", or somesuch. | |
4afc966c | 594 | |
26099b4a AC |
595 | value_virtual_fn_field needs to be fixed so there are no failures for |
596 | virtual functions for C++ using g++. | |
4afc966c | 597 | |
26099b4a AC |
598 | Testsuite cases are the major priority right now for C++ support, |
599 | since i have to make a lot of changes that could potentially break | |
600 | each other. | |
7ae38352 AC |
601 | |
602 | -- | |
603 | ||
4afc966c | 604 | Add support for Modula3 |
7ae38352 | 605 | |
4afc966c | 606 | Get DEC/Compaq to contribute their Modula-3 support. |
7ae38352 AC |
607 | |
608 | -- | |
609 | ||
4afc966c AC |
610 | Remote Protocol Support |
611 | ======================= | |
7ae38352 AC |
612 | |
613 | -- | |
614 | ||
4afc966c | 615 | set/show remote X-packet ... |
7ae38352 | 616 | |
4afc966c AC |
617 | ``(gdb) help set remote X-packet'' doesn't list the applicable |
618 | responses. The help message needs to be expanded. | |
7ae38352 AC |
619 | |
620 | -- | |
621 | ||
622 | Remote protocol doco feedback. | |
623 | ||
624 | Too much feedback to mention needs to be merged in (901660). Search | |
625 | for the word ``remote''. | |
626 | ||
4afc966c AC |
627 | |
628 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00023.html | |
629 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00056.html | |
630 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00382.html | |
631 | ||
7ae38352 AC |
632 | -- |
633 | ||
4afc966c | 634 | GDB doesn't recover gracefully from remote protocol errors. |
7ae38352 | 635 | |
4afc966c AC |
636 | GDB wasn't checking for NAKs from the remote target. Instead a NAK is |
637 | ignored and a timeout is required before GDB retries. A pre-cursor to | |
638 | fixing this this is making GDB's remote protocol packet more robust. | |
639 | ||
640 | While downloading to a remote protocol target, gdb ignores packet | |
641 | errors in so far as it will continue to edownload with chunk N+1 even | |
642 | if chunk N was not correctly sent. This causes gdb.base/remote.exp to | |
643 | take a painfully long time to run. As a PS that test needs to be | |
644 | fixed so that it builds on 16 bit machines. | |
7ae38352 AC |
645 | |
646 | -- | |
647 | ||
4afc966c | 648 | Add the cycle step command. |
7ae38352 | 649 | |
4afc966c | 650 | http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00237.html |
7ae38352 | 651 | |
26099b4a AC |
652 | -- |
653 | ||
654 | Resolve how to scale things to support very large packets. | |
655 | ||
656 | -- | |
657 | ||
658 | Resolve how to handle a target that changes things like its endianess | |
659 | on the fly - should it be returned in the ``T'' packet? | |
660 | ||
661 | Underlying problem is that the register file is target endian. If the | |
662 | target endianess changes gdb doesn't know. | |
663 | ||
7ae38352 AC |
664 | -- |
665 | ||
4afc966c AC |
666 | Symbol Support |
667 | ============== | |
7ae38352 | 668 | |
4afc966c AC |
669 | If / when GDB starts to support the debugging of multi-processor |
670 | (rather than multi-thread) applications the symtab code will need to | |
671 | be updated a little so that several independant symbol tables are | |
672 | active at a given time. | |
673 | ||
674 | The other interesting change is a clarification of the exact meaning | |
675 | of CORE_ADDR and that has had consequences for a few targets (that | |
676 | were abusing that data type). | |
7ae38352 | 677 | |
d8038014 AC |
678 | -- |
679 | ||
4afc966c | 680 | Investiagate ways of reducing memory. |
d8038014 AC |
681 | |
682 | -- | |
683 | ||
4afc966c | 684 | Investigate ways of improving load time. |
d8038014 | 685 | |
4afc966c AC |
686 | -- |
687 | ||
688 | Get the d10v to use POINTER_TO_ADDRESS and ADDRESS_TO_POINTER. | |
689 | ||
690 | Consequence of recent symtab clarification. No marks for figuring out | |
691 | who maintains the d10v. | |
d8038014 | 692 | |
0aaf65d7 AC |
693 | -- |
694 | ||
4afc966c AC |
695 | Get the MIPS to correctly sign extend all address <-> pointer |
696 | conversions. | |
0aaf65d7 | 697 | |
4afc966c AC |
698 | Consequence of recent symtab clarification. No marks for figuring out |
699 | who maintains the MIPS. | |
0aaf65d7 | 700 | |
5d35f0ac AC |
701 | -- |
702 | ||
703 | GDB truncates 64 bit enums. | |
704 | ||
705 | http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00290.html | |
706 | ||
26099b4a AC |
707 | -- |
708 | ||
709 | Testsuite Support | |
710 | ================= | |
711 | ||
712 | There are never to many testcases. | |
713 | ||
714 | -- | |
715 | ||
716 | Better thread testsuite. | |
717 | ||
718 | -- | |
719 | ||
720 | Better C++ testsuite. | |
721 | ||
722 | -- | |
723 | ||
724 | Look at adding a GDB specific testsuite directory so that white box | |
725 | tests of key internals can be added (eg ui_file). | |
726 | ||
727 | -- | |
728 | ||
729 | Separate out tests that involve the floating point (FP). | |
730 | ||
731 | (Something for people brining up new targets). FP and non-fp tests | |
732 | are combined. I think there should be set of basic tests that | |
733 | exercise pure integer support and then a more expanded set that | |
734 | exercise FP and FP/integer interactions. | |
735 | ||
736 | As an example, the MIPS, for n32 as problems with passing FP's and | |
737 | structs. Since most inferior call tests include FP it is difficult to | |
738 | determine of the integer tests are ok. | |
739 | ||
7ae38352 AC |
740 | -- |
741 | ||
4afc966c AC |
742 | Architectural Changes: General |
743 | ============================== | |
7ae38352 AC |
744 | |
745 | These are harder than simple cleanups / fixes and, consequently | |
746 | involve more work. Typically an Architectural Change will be broken | |
747 | down into a more digestible set of cleanups and fixes. | |
748 | ||
749 | -- | |
750 | ||
4afc966c AC |
751 | Cleanup software single step. |
752 | ||
753 | At present many targets implement software single step by directly | |
754 | blatting memory (see rs6000-tdep.c). Those targets should register | |
755 | the applicable breakpoints using the breakpoint framework. Perhaphs a | |
756 | new internal breakpoint class ``step'' is needed. | |
757 | ||
758 | -- | |
759 | ||
7ae38352 AC |
760 | Replace READ_FP() with FRAME_HANDLE(). |
761 | ||
762 | READ_FP() is a hangover from the days of the vax when the ABI really | |
763 | did have a frame pointer register. Modern architectures typically | |
764 | construct a virtual frame-handle from the stack pointer and various | |
765 | other bits of string. | |
766 | ||
767 | Unfortunatly GDB still treats this synthetic FP register as though it | |
768 | is real. That in turn really confuses users (arm and ``print $fp'' VS | |
769 | ``info registers fp''). The synthetic FP should be separated out of | |
770 | the true register set presented to the user. | |
771 | ||
772 | -- | |
773 | ||
2a00c9ce AC |
774 | Register Cache Cleanup (below from Andrew Cagney) |
775 | ||
776 | I would depict the current register architecture as something like: | |
777 | ||
778 | High GDB --> Low GDB | |
779 | | | | |
780 | \|/ \|/ | |
781 | --- REG NR ----- | |
782 | | | |
783 | register + REGISTER_BYTE(reg_nr) | |
784 | | | |
785 | \|/ | |
786 | ------------------------- | |
787 | | extern register[] | | |
788 | ------------------------- | |
789 | ||
790 | where neither the high (valops.c et.al.) or low gdb (*-tdep.c) are | |
791 | really clear on what mechanisms they should be using to manipulate that | |
792 | buffer. Further, much code assumes, dangerously, that registers are | |
793 | contigious. Having got mips-tdep.c to support multiple ABIs, believe | |
794 | me, that is a bad assumption. Finally, that register cache layout is | |
795 | determined by the current remote/local target and _not_ the less | |
796 | specific target ISA. In fact, in many cases it is determined by the | |
797 | somewhat arbitrary layout of the [gG] packets! | |
798 | ||
799 | ||
800 | How I would like the register file to work is more like: | |
801 | ||
802 | ||
803 | High GDB | |
804 | | | |
805 | \|/ | |
806 | pseudo reg-nr | |
807 | | | |
808 | map pseudo <-> | |
809 | random cache | |
810 | bytes | |
811 | | | |
812 | \|/ | |
813 | ------------ | |
814 | | register | | |
815 | | cache | | |
816 | ------------ | |
817 | /|\ | |
818 | | | |
819 | map random cache | |
820 | bytes to target | |
821 | dependant i-face | |
822 | /|\ | |
823 | | | |
824 | target dependant | |
825 | such as [gG] packet | |
826 | or ptrace buffer | |
827 | ||
828 | The main objectives being: | |
829 | ||
830 | o a clear separation between the low | |
831 | level target and the high level GDB | |
832 | ||
833 | o a mechanism that solves the general | |
834 | problem of register aliases, overlaps | |
835 | etc instead of treating them as optional | |
836 | extras that can be wedged in as an after | |
837 | thought (that is a reasonable description | |
838 | of the current code). | |
839 | ||
840 | Identify then solve the hard case and the | |
841 | rest just falls out. GDB solved the easy | |
842 | case and then tried to ignore the real | |
843 | world :-) | |
844 | ||
845 | o a removal of the assumption that the | |
846 | mapping between the register cache | |
847 | and virtual registers is largely static. | |
848 | If you flip the USR/SSR stack register | |
849 | select bit in the status-register then | |
850 | the corresponding stack registers should | |
851 | reflect the change. | |
852 | ||
853 | o a mechanism that clearly separates the | |
854 | gdb internal register cache from any | |
855 | target (not architecture) dependant | |
856 | specifics such as [gG] packets. | |
857 | ||
858 | Of course, like anything, it sounds good in theory. In reality, it | |
859 | would have to contend with many<->many relationships at both the | |
860 | virt<->cache and cache<->target level. For instance: | |
861 | ||
862 | virt<->cache | |
863 | Modifying an mmx register may involve | |
864 | scattering values across both FP and | |
865 | mmpx specific parts of a buffer | |
866 | ||
867 | cache<->target | |
868 | When writing back a SP it may need to | |
869 | both be written to both SP and USP. | |
870 | ||
871 | ||
872 | Hmm, | |
873 | ||
874 | Rather than let this like the last time it was discussed, just slip, I'm | |
875 | first going to add this e-mail (+ references) to TODO. I'd then like to | |
876 | sketch out a broad strategy I think could get us there. | |
877 | ||
878 | ||
879 | First thing I'd suggest is separating out the ``extern registers[]'' | |
880 | code so that we can at least identify what is using it. At present | |
881 | things are scattered across many files. That way we can at least | |
882 | pretend that there is a cache instead of a global array :-) | |
883 | ||
884 | I'd then suggest someone putting up a proposal for the pseudo-reg / | |
885 | high-level side interface so that code can be adopted to it. For old | |
886 | code, initially a blanket rename of write_register_bytes() to | |
887 | deprecated_write_register_bytes() would help. | |
888 | ||
889 | Following that would, finaly be the corresponding changes to the target. | |
bc9e5bbf AC |
890 | |
891 | -- | |
892 | ||
67edb2c6 AC |
893 | Check that GDB can handle all BFD architectures (Andrew Cagney) |
894 | ||
895 | There should be a test that checks that BFD/GDB are in sync with | |
896 | regard to architecture changes. Something like a test that first | |
897 | queries GDB for all supported architectures and then feeds each back | |
898 | to GDB.. Anyone interested in learning how to write tests? :-) | |
899 | ||
900 | -- | |
901 | ||
4afc966c AC |
902 | Architectural Change: Multi-arch et al. |
903 | ======================================= | |
2a00c9ce | 904 | |
4afc966c AC |
905 | The long term objective is to remove all assumptions that there is a |
906 | single target with a single address space with a single instruction | |
907 | set architecture and single application binary interface. | |
2a00c9ce | 908 | |
4afc966c AC |
909 | This is an ongoing effort. The first milestone is to enable |
910 | ``multi-arch'' where by all architectural decisions are made at | |
911 | runtime. | |
7ae38352 | 912 | |
4afc966c AC |
913 | It should be noted that ``gdbarch'' is really ``gdbabi'' and |
914 | ``gdbisa''. Once things are multi-arched breaking that down correctly | |
915 | will become much easier. | |
7ae38352 AC |
916 | |
917 | -- | |
918 | ||
4afc966c | 919 | GDBARCH cleanup (Andrew Cagney) |
7ae38352 | 920 | |
4afc966c AC |
921 | The non-generated parts of gdbarch.{sh,h,c} should be separated out |
922 | into arch-utils.[hc]. | |
923 | ||
924 | Document that gdbarch_init_ftype could easily fail because it didn't | |
925 | identify an architecture. | |
ed952ac5 AC |
926 | |
927 | -- | |
928 | ||
4afc966c | 929 | Fix BELIEVE_PPC_PROMOTION. Change it to BELIEVE_PPC_PROMOTION_P? |
ed952ac5 | 930 | |
4afc966c AC |
931 | At present there is still #ifdef BELIEVE_PPC_PROMOTION code in the |
932 | symtab file. | |
ed952ac5 | 933 | |
4afc966c AC |
934 | -- |
935 | ||
8e6a3c35 AC |
936 | Fix target_signal_from_host() etc. |
937 | ||
938 | The name is wrong for starters. ``target_signal'' should probably be | |
939 | ``gdb_signal''. ``from_host'' should be ``from_target_signal''. | |
940 | After that it needs to be multi-arched and made independant of any | |
941 | host signal numbering. | |
942 | ||
943 | -- | |
944 | ||
4afc966c AC |
945 | Update ALPHA so that it uses ``struct frame_extra_info'' instead of |
946 | EXTRA_FRAME_INFO. | |
947 | ||
948 | This is a barrier to replacing mips_extra_func_info with something | |
949 | that works with multi-arch. | |
7ae38352 AC |
950 | |
951 | -- | |
952 | ||
4afc966c AC |
953 | Multi-arch mips_extra_func_info. |
954 | ||
955 | This first needs the alpha to be updated so that it uses ``struct | |
956 | frame_extra_info''. | |
7ae38352 AC |
957 | |
958 | -- | |
959 | ||
4afc966c | 960 | Rationalize TARGET_SINGLE_FORMAT and TARGET_SINGLE_BIT et al. |
7ae38352 | 961 | |
4afc966c | 962 | Surely one of them is redundant. |
7ae38352 AC |
963 | |
964 | -- | |
965 | ||
4afc966c | 966 | Convert ALL architectures to MULTI-ARCH. |
7ae38352 AC |
967 | |
968 | -- | |
969 | ||
970 | Select the initial multi-arch ISA / ABI based on --target or similar. | |
971 | ||
972 | At present the default is based on what ever is first in the BFD | |
973 | archures table. It should be determined based on the ``--target=...'' | |
974 | name. | |
975 | ||
976 | -- | |
977 | ||
bf64bfd6 AC |
978 | Make MIPS pure multi-arch. |
979 | ||
980 | It is only at the multi-arch enabled stage. | |
981 | ||
982 | -- | |
983 | ||
7ae38352 AC |
984 | Truly multi-arch. |
985 | ||
986 | Enable the code to recognize --enable-targets=.... like BINUTILS does. | |
987 | ||
4afc966c AC |
988 | Can the tm.h and nm.h files be eliminated by multi-arch. |
989 | ||
7ae38352 AC |
990 | -- |
991 | ||
4afc966c AC |
992 | Architectural Change: MI, LIBGDB and scripting languages |
993 | ======================================================== | |
7ae38352 | 994 | |
4afc966c AC |
995 | See also architectural changes related to the event loop. LIBGDB |
996 | can't be finished until there is a generic event loop being used by | |
997 | all targets. | |
998 | ||
999 | The long term objective is it to be possible to integrate GDB into | |
1000 | scripting languages. | |
7ae38352 AC |
1001 | |
1002 | -- | |
1003 | ||
4afc966c | 1004 | Implement generic ``(gdb) commmand > file'' |
7ae38352 | 1005 | |
4afc966c AC |
1006 | Once everything is going through ui_file it should be come fairly |
1007 | easy. | |
1008 | ||
1009 | http://sourceware.cygnus.com/ml/gdb/2000-04/msg00104.html | |
1010 | ||
1011 | -- | |
1012 | ||
1013 | Replace gdb_stdtarg with gdb_targout (and possibly gdb_targerr). | |
1014 | ||
1015 | gdb_stdtarg is easily confused with gdb_stdarg. | |
1016 | ||
1017 | -- | |
1018 | ||
1019 | Extra ui_file methods - dump. | |
1020 | ||
26099b4a | 1021 | Very useful for whitebox testing. |
4afc966c AC |
1022 | |
1023 | -- | |
1024 | ||
1025 | Eliminate error_begin(). | |
1026 | ||
1027 | With ui_file, there is no need for the statefull error_begin () | |
1028 | function. | |
1029 | ||
1030 | -- | |
1031 | ||
1032 | Send normal output to gdb_stdout. | |
1033 | Send error messages to gdb_stderror. | |
1034 | Send debug and log output log gdb_stdlog. | |
1035 | ||
1036 | GDB still contains many cases where (f)printf or printf_filtered () is | |
1037 | used when it should be sending the messages to gdb_stderror or | |
1038 | gdb_stdlog. The thought of #defining printf to something has crossed | |
1039 | peoples minds ;-) | |
7ae38352 AC |
1040 | |
1041 | -- | |
1042 | ||
1043 | Re-do GDB's output pager. | |
1044 | ||
1045 | GDB's output pager still relies on people correctly using *_filtered | |
1046 | for gdb_stdout and *_unfiltered for gdb_stdlog / gdb_stderr. | |
1047 | Hopefully, with all normal output going to gdb_stdout, the pager can | |
1048 | just look at the ui_file that the output is on and then use that to | |
1049 | decide what to do about paging. Sounds good in theory. | |
1050 | ||
1051 | -- | |
1052 | ||
4afc966c AC |
1053 | Check/cleanup MI documentation. |
1054 | ||
1055 | The list of commands specified in the documentation needs to be | |
1056 | checked against the mi-cmds.c table in a mechanical way (so that they | |
1057 | two can be kept up-to-date). | |
1058 | ||
1059 | -- | |
1060 | ||
1061 | Convert MI into libgdb | |
1062 | ||
1063 | MI provides a text interface into what should be many of the libgdb | |
1064 | functions. The implementation of those functions should be separated | |
1065 | into the MI interface and the functions proper. Those functions being | |
1066 | moved to gdb/lib say. | |
1067 | ||
1068 | -- | |
1069 | ||
1070 | Create libgdb.h | |
1071 | ||
1072 | The first part can already be found in defs.h. | |
1073 | ||
1074 | -- | |
1075 | ||
1076 | MI's input does not use buffering. | |
1077 | ||
1078 | At present the MI interface reads raw characters of from an unbuffered | |
1079 | FD. This is to avoid several nasty buffer/race conditions. That code | |
1080 | should be changed so that it registers its self with the event loop | |
1081 | (on the input FD) and then push commands up to MI as they arrive. | |
1082 | ||
1083 | The serial code already does this. | |
1084 | ||
1085 | -- | |
1086 | ||
1087 | Make MI interface accessable from existing CLI. | |
1088 | ||
1089 | -- | |
1090 | ||
1091 | Add a breakpoint-edit command to MI. | |
1092 | ||
1093 | It would be similar to MI's breakpoint create but would apply to an | |
1094 | existing breakpoint. It saves the need to delete/create breakpoints | |
1095 | when ever they are changed. | |
1096 | ||
1097 | -- | |
1098 | ||
1099 | Add directory path to MI breakpoint. | |
1100 | ||
1101 | That way the GUI's task of finding the file within which the | |
1102 | breakpoint was set is simplified. | |
1103 | ||
1104 | -- | |
1105 | ||
1106 | Add a mechanism to reject certain expression classes to MI | |
7ae38352 AC |
1107 | |
1108 | There are situtations where you don't want GDB's expression | |
1109 | parser/evaluator to perform inferior function calls or variable | |
4afc966c AC |
1110 | assignments. A way of restricting the expression parser so that such |
1111 | operations are not accepted would be very helpful. | |
7ae38352 AC |
1112 | |
1113 | -- | |
1114 | ||
1115 | Remove sideffects from libgdb breakpoint create function. | |
1116 | ||
1117 | The user can use the CLI to create a breakpoint with partial | |
1118 | information - no file (gdb would use the file from the last | |
1119 | breakpoint). | |
1120 | ||
1121 | The libgdb interface currently affects that environment which can lead | |
1122 | to confusion when a user is setting breakpoints via both the MI and | |
1123 | the CLI. | |
1124 | ||
1125 | This is also a good example of how getting the CLI ``right'' will be | |
1126 | hard. | |
1127 | ||
1128 | -- | |
1129 | ||
4afc966c | 1130 | Move gdb_lasterr to ui_out? |
7ae38352 | 1131 | |
4afc966c AC |
1132 | The way GDB throws errors and records them needs a re-think. ui_out |
1133 | handles the correct output well. It doesn't resolve what to do with | |
1134 | output / error-messages when things go wrong. | |
7ae38352 | 1135 | |
97c3646f AC |
1136 | -- |
1137 | ||
1138 | do_setshow_command contains a 1024 byte buffer. | |
1139 | ||
1140 | The function assumes that there will never be any more than 1024 bytes | |
1141 | of enum. It should use mem_file. | |
1142 | ||
1143 | -- | |
1144 | ||
1145 | Should struct cmd_list_element . completer take the command as an | |
1146 | argument? | |
1147 | ||
1148 | -- | |
1149 | ||
1150 | Should the bulk of top.c:line_completion_function() be moved to | |
1151 | command.[hc]? complete_on_cmdlist() and complete_on_enums() could | |
1152 | then be made private. | |
1153 | ||
1154 | -- | |
1155 | ||
1156 | top.c (execute_command): Should a command being valid when the target | |
1157 | is running be made an attribute (predicate) to the command rather than | |
1158 | an explicit set of tests. | |
1159 | ||
1160 | -- | |
1161 | ||
1162 | top.c (execute_command): Should the bulk of this function be moved | |
1163 | into command.[hc] so that top.c doesn't grub around in the command | |
1164 | internals? | |
1165 | ||
4afc966c AC |
1166 | -- |
1167 | ||
1168 | Architectural Change: Async | |
1169 | =========================== | |
1170 | ||
1171 | While GDB uses an event loop when prompting the user for input. That | |
1172 | event loop is not exploited by targets when they allow the target | |
1173 | program to continue. Typically targets still block in (target_wait()) | |
1174 | until the program again halts. | |
1175 | ||
1176 | The closest a target comes to supporting full asynchronous mode are | |
1177 | the remote targets ``async'' and ``extended-async''. | |
7ae38352 AC |
1178 | |
1179 | -- | |
1180 | ||
4afc966c | 1181 | Asynchronous expression evaluator |
7ae38352 | 1182 | |
4afc966c | 1183 | Inferior function calls hang GDB. |
7ae38352 AC |
1184 | |
1185 | -- | |
1186 | ||
1187 | Fix implementation of ``target xxx''. | |
1188 | ||
1189 | At present when the user specifies ``target xxxx'', the CLI maps that | |
1190 | directly onto a target open method. It is then assumed that the | |
1191 | target open method should do all sorts of complicated things as this | |
1192 | is the only chance it has. Check how the various remote targets | |
1193 | duplicate the target operations. Check also how the various targets | |
1194 | behave differently for purely arbitrary reasons. | |
1195 | ||
1196 | What should happen is that ``target xxxx'' should call a generic | |
1197 | ``target'' function and that should then co-ordinate the opening of | |
1198 | ``xxxx''. This becomes especially important when you're trying to | |
1199 | open an asynchronous target that may need to perform background tasks | |
1200 | as part of the ``attach'' phase. | |
1201 | ||
1202 | Unfortunatly, due to limitations in the old/creaking command.h | |
1203 | interface, that isn't possible. The function being called isn't told | |
1204 | of the ``xxx'' or any other context information. | |
1205 | ||
1206 | Consequently a precursor to fixing ``target xxxx'' is to clean up the | |
1207 | CLI code so that it passes to the callback function (attatched to a | |
1208 | command) useful information such as the actual command and a context | |
1209 | for that command. Other changes such as making ``struct command'' | |
1210 | opaque may also help. | |
1211 | ||
53904c9e AC |
1212 | See also: |
1213 | http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html | |
1214 | ||
7ae38352 AC |
1215 | -- |
1216 | ||
4afc966c AC |
1217 | Make "target xxx" command interruptible. |
1218 | ||
1219 | As things become async this becomes possible. A target would start | |
1220 | the connect and then return control to the event loop. A cntrl-c | |
1221 | would notify the target that the operation is to be abandoned and the | |
1222 | target code could respond. | |
7ae38352 AC |
1223 | |
1224 | -- | |
1225 | ||
4afc966c AC |
1226 | Add a "suspend" subcommand of the "continue" command to suspend gdb |
1227 | while continuing execution of the subprocess. Useful when you are | |
1228 | debugging servers and you want to dodge out and initiate a connection | |
1229 | to a server running under gdb. | |
1230 | ||
1231 | [hey async!!] | |
7ae38352 | 1232 | |
2a00c9ce AC |
1233 | -- |
1234 | ||
26099b4a AC |
1235 | TODO FAQ |
1236 | ======== | |
1237 | ||
1238 | Frequently requested but not approved requests. | |
1239 | ||
1240 | -- | |
1241 | ||
1242 | Eliminate unused argument warnings using ATTRIBUTE_UNUSED. | |
1243 | ||
1244 | The benefits on this one are thought to be marginal - GDBs design | |
1245 | means that unused parameters are very common. GCC 3.0 will also | |
1246 | include the option -Wno-unused-parameter which means that ``-Wall | |
1247 | -Wno-unused-parameters -Werror'' can be specified. | |
1248 | ||
1249 | -- | |
1250 | ||
1251 | ||
1252 | ||
2a00c9ce AC |
1253 | Legacy Wish List |
1254 | ================ | |
1255 | ||
1256 | This list is not up to date, and opinions vary about the importance or | |
1257 | even desirability of some of the items. If you do fix something, it | |
1258 | always pays to check the below. | |
1259 | ||
1260 | -- | |
c906108c | 1261 | |
b83266a0 SS |
1262 | @c This does not work (yet if ever). FIXME. |
1263 | @c @item --parse=@var{lang} @dots{} | |
1264 | @c Configure the @value{GDBN} expression parser to parse the listed languages. | |
1265 | @c @samp{all} configures @value{GDBN} for all supported languages. To get a | |
1266 | @c list of all supported languages, omit the argument. Without this | |
1267 | @c option, @value{GDBN} is configured to parse all supported languages. | |
1268 | ||
7ae38352 | 1269 | -- |
c906108c SS |
1270 | |
1271 | START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that | |
1272 | is its default value. Clean this up. | |
1273 | ||
7ae38352 AC |
1274 | -- |
1275 | ||
c906108c SS |
1276 | It should be possible to use symbols from shared libraries before we know |
1277 | exactly where the libraries will be loaded. E.g. "b perror" before running | |
1278 | the program. This could maybe be done as an extension of the "breakpoint | |
1279 | re-evaluation" after new symbols are loaded. | |
1280 | ||
7ae38352 AC |
1281 | -- |
1282 | ||
c906108c SS |
1283 | Make single_step() insert and remove breakpoints in one operation. |
1284 | ||
26099b4a AC |
1285 | [If this is talking about having single_step() insert the breakpoints, |
1286 | run the target then pull the breakpoints then it is wrong. The | |
1287 | function has to return as control has to eventually be passed back to | |
1288 | the main event loop.] | |
1289 | ||
7ae38352 AC |
1290 | -- |
1291 | ||
c906108c SS |
1292 | Speed up single stepping by avoiding extraneous ptrace calls. |
1293 | ||
7ae38352 AC |
1294 | -- |
1295 | ||
c906108c SS |
1296 | Speed up single stepping by not inserting and removing breakpoints |
1297 | each time the inferior starts and stops. | |
1298 | ||
1299 | Breakpoints should not be inserted and deleted all the time. Only the | |
1300 | one(s) there should be removed when we have to step over one. Support | |
1301 | breakpoints that don't have to be removed to step over them. | |
1302 | ||
7ae38352 | 1303 | [this has resulted in numerous debates. The issue isn't clear cut] |
c906108c | 1304 | |
7ae38352 | 1305 | -- |
c906108c SS |
1306 | |
1307 | Provide "voodoo" debugging of core files. This creates a zombie | |
1308 | process as a child of the debugger, and loads it up with the data, | |
1309 | stack, and regs of the core file. This allows you to call functions | |
1310 | in the executable, to manipulate the data in the core file. | |
1311 | ||
7ae38352 | 1312 | [you wish] |
c906108c | 1313 | |
7ae38352 | 1314 | -- |
c906108c | 1315 | |
7ae38352 | 1316 | GDB reopens the source file on every line, as you "next" through it. |
c906108c | 1317 | |
7ae38352 | 1318 | [still true? I've a memory of this being fixed] |
c906108c | 1319 | |
7ae38352 | 1320 | -- |
c906108c SS |
1321 | |
1322 | Perhaps "i source" should take an argument like that of "list". | |
1323 | ||
7ae38352 | 1324 | -- |
c906108c SS |
1325 | |
1326 | Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if | |
1327 | it matches the source line indicated. | |
1328 | ||
7ae38352 | 1329 | -- |
c906108c | 1330 | |
7ae38352 | 1331 | The prompt at end of screen should accept space as well as CR. |
c906108c | 1332 | |
7ae38352 | 1333 | -- |
c906108c SS |
1334 | |
1335 | Backtrace should point out what the currently selected frame is, in | |
1336 | its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar, | |
1337 | ...)" rather than "#3 foo (bar, ...)". | |
1338 | ||
7ae38352 AC |
1339 | -- |
1340 | ||
c906108c SS |
1341 | "i program" should work for core files, and display more info, like what |
1342 | actually caused it to die. | |
1343 | ||
7ae38352 AC |
1344 | -- |
1345 | ||
c906108c SS |
1346 | "x/10i" should shorten the long name, if any, on subsequent lines. |
1347 | ||
7ae38352 | 1348 | -- |
c906108c SS |
1349 | |
1350 | "next" over a function that longjumps, never stops until next time you happen | |
1351 | to get to that spot by accident. E.g. "n" over execute_command which has | |
1352 | an error. | |
1353 | ||
7ae38352 AC |
1354 | -- |
1355 | ||
c906108c SS |
1356 | "set zeroprint off", don't bother printing members of structs which |
1357 | are entirely zero. Useful for those big structs with few useful | |
1358 | members. | |
1359 | ||
7ae38352 AC |
1360 | -- |
1361 | ||
c906108c SS |
1362 | GDB does four ioctl's for every command, probably switching terminal modes |
1363 | to/from inferior or for readline or something. | |
1364 | ||
7ae38352 AC |
1365 | -- |
1366 | ||
c906108c SS |
1367 | terminal_ours versus terminal_inferior: cache state. Switch should be a noop |
1368 | if the state is the same, too. | |
1369 | ||
7ae38352 | 1370 | -- |
c906108c SS |
1371 | |
1372 | "i frame" shows wrong "arglist at" location, doesn't show where the args | |
1373 | should be found, only their actual values. | |
1374 | ||
7ae38352 AC |
1375 | -- |
1376 | ||
c906108c SS |
1377 | There should be a way for "set" commands to validate the new setting |
1378 | before it takes effect. | |
1379 | ||
7ae38352 | 1380 | -- |
c906108c SS |
1381 | |
1382 | "ena d" is ambiguous, why? "ena delete" seems to think it is a command! | |
1383 | ||
7ae38352 | 1384 | -- |
c906108c SS |
1385 | |
1386 | i line VAR produces "Line number not known for symbol ``var''.". I | |
1387 | thought we were stashing that info now! | |
1388 | ||
7ae38352 AC |
1389 | -- |
1390 | ||
c906108c SS |
1391 | We should be able to write to random files at hex offsets like adb. |
1392 | ||
7ae38352 AC |
1393 | -- |
1394 | ||
7ae38352 AC |
1395 | [elena - delete this] |
1396 | ||
c906108c SS |
1397 | Handle add_file with separate text, data, and bss addresses. Maybe |
1398 | handle separate addresses for each segment in the object file? | |
1399 | ||
7ae38352 AC |
1400 | -- |
1401 | ||
1402 | [Jimb/Elena delete this one] | |
1403 | ||
c906108c SS |
1404 | Handle free_named_symtab to cope with multiply-loaded object files |
1405 | in a dynamic linking environment. Should remember the last copy loaded, | |
1406 | but not get too snowed if it finds references to the older copy. | |
1407 | ||
7ae38352 AC |
1408 | -- |
1409 | ||
1410 | [elena delete this also] | |
c906108c SS |
1411 | |
1412 | Remove all references to: | |
1413 | text_offset | |
1414 | data_offset | |
1415 | text_data_start | |
1416 | text_end | |
1417 | exec_data_offset | |
1418 | ... | |
1419 | now that we have BFD. All remaining are in machine dependent files. | |
1420 | ||
7ae38352 | 1421 | -- |
c906108c SS |
1422 | |
1423 | Re-organize help categories into things that tend to fit on a screen | |
1424 | and hang together. | |
1425 | ||
7ae38352 | 1426 | -- |
c906108c SS |
1427 | |
1428 | Add in commands like ADB's for searching for patterns, etc. We should | |
1429 | be able to examine and patch raw unsymboled binaries as well in gdb as | |
1430 | we can in adb. (E.g. increase the timeout in /bin/login without source). | |
1431 | ||
7ae38352 AC |
1432 | [actually, add ADB interface :-] |
1433 | ||
1434 | -- | |
c906108c SS |
1435 | |
1436 | When doing "step" or "next", if a few lines of source are skipped between | |
1437 | the previous line and the current one, print those lines, not just the | |
1438 | last line of a multiline statement. | |
1439 | ||
7ae38352 | 1440 | -- |
c906108c SS |
1441 | |
1442 | Handling of "&" address-of operator needs some serious overhaul | |
1443 | for ANSI C and consistency on arrays and functions. | |
1444 | For "float point[15];": | |
1445 | ptype &point[4] ==> Attempt to take address of non-lvalue. | |
1446 | For "char *malloc();": | |
1447 | ptype malloc ==> "char *()"; should be same as | |
1448 | ptype &malloc ==> "char *(*)()" | |
1449 | call printf ("%x\n", malloc) ==> weird value, should be same as | |
1450 | call printf ("%x\n", &malloc) ==> correct value | |
1451 | ||
7ae38352 AC |
1452 | -- |
1453 | ||
c906108c SS |
1454 | Fix dbxread.c symbol reading in the presence of interrupts. It |
1455 | currently leaves a cleanup to blow away the entire symbol table when a | |
1456 | QUIT occurs. (What's wrong with that? -kingdon, 28 Oct 1993). | |
1457 | ||
7ae38352 AC |
1458 | [I suspect that the grype was that, on a slow system, you might want |
1459 | to cntrl-c and get just half the symbols and then load the rest later | |
1460 | - scary to be honest] | |
1461 | ||
1462 | -- | |
1463 | ||
c906108c SS |
1464 | Mipsread.c reads include files depth-first, because the dependencies |
1465 | in the psymtabs are way too inclusive (it seems to me). Figure out what | |
1466 | really depends on what, to avoid recursing 20 or 30 times while reading | |
1467 | real symtabs. | |
1468 | ||
7ae38352 AC |
1469 | -- |
1470 | ||
c906108c SS |
1471 | value_add() should be subtracting the lower bound of arrays, if known, |
1472 | and possibly checking against the upper bound for error reporting. | |
1473 | ||
7ae38352 | 1474 | -- |
c906108c SS |
1475 | |
1476 | When listing source lines, check for a preceding \n, to verify that | |
1477 | the file hasn't changed out from under us. | |
1478 | ||
7ae38352 AC |
1479 | [fixed by some other means I think. That hack wouldn't actually work |
1480 | reliably - the file might move such that another \n appears. ] | |
c906108c | 1481 | |
7ae38352 | 1482 | -- |
c906108c SS |
1483 | |
1484 | Get all the remote systems (where the protocol allows it) to be able to | |
1485 | stop the remote system when the GDB user types ^C (like remote.c | |
1486 | does). For ebmon, use ^Ak. | |
1487 | ||
7ae38352 AC |
1488 | -- |
1489 | ||
c906108c SS |
1490 | Possible feature: A version of the "disassemble" command which shows |
1491 | both source and assembly code ("set symbol-filename on" is a partial | |
1492 | solution). | |
1493 | ||
7ae38352 AC |
1494 | [has this been done? It was certainly done for MI and GDBtk] |
1495 | ||
1496 | -- | |
1497 | ||
c906108c SS |
1498 | investigate "x/s 0" (right now stops early) (I think maybe GDB is |
1499 | using a 0 address for bad purposes internally). | |
1500 | ||
7ae38352 AC |
1501 | -- |
1502 | ||
c906108c SS |
1503 | Make "info path" and path_command work again (but independent of the |
1504 | environment either of gdb or that we'll pass to the inferior). | |
1505 | ||
7ae38352 AC |
1506 | -- |
1507 | ||
c906108c SS |
1508 | Make GDB understand the GCC feature for putting octal constants in |
1509 | enums. Make it so overflow on an enum constant does not error_type | |
1510 | the whole type. Allow arbitrarily large enums with type attributes. | |
1511 | Put all this stuff in the testsuite. | |
1512 | ||
7ae38352 AC |
1513 | -- |
1514 | ||
c906108c SS |
1515 | Make TYPE_CODE_ERROR with a non-zero TYPE_LENGTH more useful (print |
1516 | the value in hex; process type attributes). Add this to the | |
1517 | testsuite. This way future compilers can add new types and old | |
1518 | versions of GDB can do something halfway reasonable. | |
1519 | ||
7ae38352 | 1520 | -- |
c906108c SS |
1521 | |
1522 | Fix mdebugread.c:parse_type to do fundamental types right (see | |
1523 | rs6000_builtin_type in stabsread.c for what "right" is--the point is | |
1524 | that the debug format fixes the sizes of these things and it shouldn't | |
1525 | depend on stuff like TARGET_PTR_BIT and so on. For mdebug, there seem | |
1526 | to be separate bt* codes for 64 bit and 32 bit things, and GDB should | |
1527 | be aware of that). Also use a switch statement for clarity and speed. | |
1528 | ||
7ae38352 AC |
1529 | -- |
1530 | ||
c906108c SS |
1531 | Investigate adding symbols in target_load--some targets do, some |
1532 | don't. | |
1533 | ||
7ae38352 AC |
1534 | -- |
1535 | ||
c906108c SS |
1536 | Put dirname in psymtabs and change lookup*symtab to use dirname (so |
1537 | /foo/bar.c works whether compiled by cc /foo/bar.c, or cd /foo; cc | |
1538 | bar.c). | |
1539 | ||
7ae38352 AC |
1540 | -- |
1541 | ||
c906108c SS |
1542 | Merge xcoffread.c and coffread.c. Use breakpoint_re_set instead of |
1543 | fixup_breakpoints. | |
1544 | ||
7ae38352 | 1545 | -- |
c906108c SS |
1546 | |
1547 | Make a watchpoint which contains a function call an error (it is | |
1548 | broken now, making it work is probably not worth the effort). | |
1549 | ||
7ae38352 | 1550 | -- |
c906108c SS |
1551 | |
1552 | New test case based on weird.exp but in which type numbers are not | |
1553 | renumbered (thus multiply defining a type). This currently causes an | |
1554 | infinite loop on "p v_comb". | |
1555 | ||
7ae38352 | 1556 | -- |
c906108c | 1557 | |
7ae38352 | 1558 | [Hey! Hint Hint Delete Delete!!!] |
c906108c SS |
1559 | |
1560 | Fix 386 floating point so that floating point registers are real | |
1561 | registers (but code can deal at run-time if they are missing, like | |
1562 | mips and 68k). This would clean up "info float" and related stuff. | |
1563 | ||
7ae38352 | 1564 | -- |
c906108c SS |
1565 | |
1566 | gcc -g -c enummask.c then gdb enummask.o, then "p v". GDB complains | |
1567 | about not being able to access memory location 0. | |
1568 | ||
1569 | -------------------- enummask.c | |
1570 | enum mask | |
1571 | { | |
1572 | ANIMAL = 0, | |
1573 | VEGETABLE = 1, | |
1574 | MINERAL = 2, | |
1575 | BASIC_CATEGORY = 3, | |
1576 | ||
1577 | WHITE = 0, | |
1578 | BLUE = 4, | |
1579 | GREEN = 8, | |
1580 | BLACK = 0xc, | |
1581 | COLOR = 0xc, | |
1582 | ||
1583 | ALIVE = 0x10, | |
1584 | ||
1585 | LARGE = 0x20 | |
1586 | } v; | |
1587 | ||
7ae38352 AC |
1588 | -- |
1589 | ||
c906108c SS |
1590 | If try to modify value in file with "set write off" should give |
1591 | appropriate error not "cannot access memory at address 0x65e0". | |
1592 | ||
7ae38352 | 1593 | -- |
c906108c | 1594 | |
c906108c SS |
1595 | Allow core file without exec file on RS/6000. |
1596 | ||
7ae38352 AC |
1597 | -- |
1598 | ||
c906108c SS |
1599 | Make sure "shell" with no arguments works right on DOS. |
1600 | ||
7ae38352 AC |
1601 | -- |
1602 | ||
c906108c SS |
1603 | Make gdb.ini (as well as .gdbinit) be checked on all platforms, so |
1604 | the same directory can be NFS-mounted on unix or DOS, and work the | |
1605 | same way. | |
1606 | ||
7ae38352 AC |
1607 | -- |
1608 | ||
1609 | [Is this another delete???] | |
c906108c SS |
1610 | |
1611 | Get SECT_OFF_TEXT stuff out of objfile_relocate (might be needed to | |
1612 | get RS/6000 to work right, might not be immediately relevant). | |
1613 | ||
7ae38352 | 1614 | -- |
c906108c SS |
1615 | |
1616 | Work out some kind of way to allow running the inferior to be done as | |
1617 | a sub-execution of, eg. breakpoint command lists. Currently running | |
1618 | the inferior interupts any command list execution. This would require | |
1619 | some rewriting of wait_for_inferior & friends, and hence should | |
1620 | probably be done in concert with the above. | |
1621 | ||
7ae38352 AC |
1622 | -- |
1623 | ||
c906108c SS |
1624 | Add function arguments to gdb user defined functions. |
1625 | ||
7ae38352 AC |
1626 | -- |
1627 | ||
c906108c SS |
1628 | Add convenience variables that refer to exec file, symbol file, |
1629 | selected frame source file, selected frame function, selected frame | |
1630 | line number, etc. | |
1631 | ||
7ae38352 AC |
1632 | -- |
1633 | ||
c906108c SS |
1634 | Modify the handling of symbols grouped through BINCL/EINCL stabs to |
1635 | allocate a partial symtab for each BINCL/EINCL grouping. This will | |
1636 | seriously decrease the size of inter-psymtab dependencies and hence | |
1637 | lessen the amount that needs to be read in when a new source file is | |
1638 | accessed. | |
1639 | ||
7ae38352 | 1640 | -- |
c906108c | 1641 | |
c906108c SS |
1642 | Add a command for searching memory, a la adb. It specifies size, |
1643 | mask, value, start address. ADB searches until it finds it or hits | |
1644 | an error (or is interrupted). | |
1645 | ||
7ae38352 AC |
1646 | -- |
1647 | ||
b83266a0 SS |
1648 | Remove the range and type checking code and documentation, if not |
1649 | going to implement. | |
1650 | ||
c906108c SS |
1651 | # Local Variables: |
1652 | # mode: text | |
1653 | # End: |