1 If you find inaccuracies in this list, please send mail to
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.
7 Known problems in GDB 5.0
8 =========================
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
14 (The names in paren indicate people that posted the original problem.)
18 GDB requires GCC to build under IRIX
20 IRIX, being more pedantic than GCC reports as errors certain
21 assignments that GCC treats as warnings.
23 This can be worked around by building GDB with the GCC compiler.
27 The BFD directory requires bug-fixed AUTOMAKE et.al.
29 AUTOMAKE 1.4 incorrectly set the TEXINPUTS environment variable. It
30 contained the full path to texinfo.tex when it should have only
31 contained the directory. The bug has been fixed in the current
32 AUTOMAKE sources. Automake snapshots can be found in:
33 ftp://sourceware.cygnus.com/pub/gdb/snapshots
34 and ftp://sourceware.cygnus.com/pub/binutils
38 Possible regressions with some devel GCCs.
39 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00475.html
41 gcc-2.95.2 outputs a line note *before* the prologue (and one for the
42 closing brace after the epilogue, instead of before it, as it used to
43 be). By disabling the RTL-style prologue generating mechanism
44 (undocumented GCC option -mno-schedule-prologue), you get back the
45 traditional behaviour.
46 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00510.html
48 This should now be fixed.
52 RFD: infrun.c: No bpstat_stop_status call after proceed over break?
54 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00665.html
56 GDB misses watchpoint triggers after proceeding over a breakpoint on
61 x86 linux GDB and SIGALRM (???)
62 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00803.html
64 I know there are problems with single stepping through signal
65 handlers. These problems were present in 4.18. They were just masked
66 because 4.18 failed to recognize signal handlers. Fixing it is not
67 easy, and will require changes to handle_inferior_event(), that I
68 prefer not to make before the 5.0 release.
74 Revised UDP support (was: Re: [Fwd: [patch] UDP transport support])
75 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00000.html
77 (Broken) support for GDB's remote protocol across UDP is to be
78 included in the follow-on release.
82 Can't build IRIX -> arm GDB.
83 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00356.html
86 > Now I'm building for an embedded arm target. If there is a way of turning
87 > remote-rdi off, I couldn't find it. It looks like it gets built by default
88 > in gdb/configure.tgt(line 58) Anyway, the build dies in
89 > gdb/rdi-share/unixcomm.c. SERPORT1 et. al. never get defined because we
90 > aren't one of the architectures supported.
94 Problem with weak functions
95 http://sourceware.cygnus.com/ml/gdb/2000-05/msg00060.html
97 Dan Nicolaescu writes:
98 > It seems that gdb-4.95.1 does not display correctly the function when
99 > stoping in weak functions.
101 > It stops in a function that is defined as weak, not in the function
102 > that is actualy run...
106 Code Cleanups: Next Release
107 ===========================
109 The following are small cleanups that will hopefully be completed by
110 the follow on to 5.0.
114 Delete macro TARGET_BYTE_ORDER_SELECTABLE.
116 Patches in the database.
122 Eliminate all uses of PARAMS in GDB's source code.
126 Elimination of make_cleanup_func. (Andrew Cagney)
128 make_cleanup_func elimination
129 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00791.html
130 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00814.html
134 Code Cleanups: General
135 ======================
137 The following are more general cleanups and fixes. They are not tied
138 to any specific release.
144 The need for this as almost been eliminated. The next version of GCC
145 (assuming cagney gets the relevant patch committed) will be able to
146 supress unused parameter warnings.
150 Eliminate more compiler warnings.
152 Of course there also needs to be the usual debate over which warnings
153 are valid and how to best go about this.
155 One method: choose a single option; get agreement that it is
156 reasonable; try it out to see if there isn't anything silly about it
157 (-Wunused-parameters is an example of that) then incrementally hack
160 The other method is to enable all warnings and eliminate them from one
165 Elimination of ``(catch_errors_ftype *) func''.
167 Like make_cleanup_func it isn't portable.
175 [PATCH/5] src/intl/Makefile.in:distclean additions
176 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00363.html
178 Do not forget to merge the patch back into the trunk.
182 Rationalize the host-endian code (grep for HOST_BYTE_ORDER).
184 At present defs.h includes <endian.h> (which is linux specific) yet
185 almost nothing depends on it. Suggest "gdb_endian.h" which can also
186 handle <machine/endian.h> and only include that where it is really
191 Replace asprintf() calls with xasprintf() calls.
193 As with things like strdup() most calls to asprintf() don't check the
198 Replace strsave() + mstrsave() with libiberty:xstrdup().
202 Replace savestring() with something from libiberty.
204 An xstrldup()? but that would have different semantics.
208 Rationalize use of floatformat_unknown in GDB sources.
210 Instead of defaulting to floatformat_unknown, should hosts/targets
211 specify the value explicitly?
213 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
217 Add a ``name'' member to include/floatformat.h:struct floatformat.
218 Print that name in gdbarch.c.
222 Sort out the harris mess in include/floatformat.h (it hardwires two
223 different floating point formats).
227 See of the GDB local floatformat_do_doublest() and libiberty's
228 floatformat_to_double (which was once GDB's ...) can be merged some
233 Eliminate mmalloc() from GDB.
235 Also eliminate it from defs.h.
239 Eliminate PTR. ISO-C allows ``void *''.
245 GDB should never abort. GDB should either throw ``error ()'' or
246 ``internal_error ()''. Better still GDB should naturally unwind with
251 GDB probably doesn't build on FreeBSD pre 2.2.x
252 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00378.html
254 Fixes to get FreeBSD working on 2.2.x, 3.x and 4.x caused the code to
261 Readline 4.? is out. A merge wouldn't hurt.
265 New Features and Fixes
266 ======================
268 These are harder than cleanups but easier than work involving
269 fundamental architectural change.
273 Add built-by, build-date, tm, xm, nm and anything else into gdb binary
274 so that you can see how the GDB was created.
276 Some of these (*m.h) would be added to the generated config.h. That
277 in turn would fix a long standing bug where by the build process many
278 not notice a changed tm.h file. Since everything depends on config.h,
279 a change to *m.h forces a change to config.h and, consequently forces
284 Add an "info bfd" command that displays supported object formats,
285 similarly to objdump -i.
287 Is there a command already?
291 Fix ``I'm sorry, Dave, I can't do that.'' from symfile.c.
293 This requires internationalization.
297 Convert GDB build process to AUTOMAKE.
299 See also sub-directory configure below.
303 Cleanup configury support for optional sub-directories.
305 Check how GCC handles multiple front ends for an example of how things
306 could work. A tentative first step is to rationalize things so that
307 all sub directories are handled in a fashion similar to gdb/mi.
309 See also automake above.
313 Restructure gdb directory tree so that it avoids any 8.3 and 14
318 Add a transcript mechanism to GDB.
320 Such a mechanism might log all gdb input and output to a file in a
321 form that would allow it to be replayed. It could involve ``gdb
322 --transcript=FILE'' or it could involve ``(gdb) transcript file''.
326 Can the xdep files be replaced by autoconf?
330 Document trace machinery
334 Document overlay machinery.
338 ``(gdb) catch signal SIGNAL''
340 Overlaps with ``handle SIGNAL'' but the implied behavour is different.
341 You can attach commands to a catch but not a handle. A handle has a
342 limited number of hardwired actions.
346 Get the TUI working on all platforms.
350 Add support for ``gdb --- PROGRAM ARGS ...''.
351 Add support for ``gdb -cmd=...''
353 Along with many variations. Check:
355 ????? for a full discussion.
361 Implement ``(gdb) !ls''.
363 Which is very different from ``(gdb) ! ls''. Implementing the latter
366 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00034.html
370 Replace the code that uses the host FPU with an emulator of the target
380 Generic: lin-thread cannot handle thread exit (Mark Kettenis, Michael
381 Snyder) http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00525.html
383 The thread_db assisted debugging code doesn't handle exiting threads
384 properly, at least in combination with glibc 2.1.3 (the framework is
385 there, just not the actual code). There are at least two problems
386 that prevent this from working.
388 As an additional reference point, the pre thread_db code did not work
393 GNU/Linux/x86 and random thread signals (and Solaris/SPARC but not
395 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00336.html
397 Christopher Blizzard writes:
399 So, I've done some more digging into this and it looks like Jim
400 Kingdon has reported this problem in the past:
402 http://sourceware.cygnus.com/ml/bug-gdb/1999-10/msg00058.html
404 I can reproduce this problem both with and without Tom's patch. Has
405 anyone seen this before? Maybe have a solution for it hanging around?
408 There's a test case for this documented at:
410 when debugging threaded applications you get extra SIGTRAPs
411 http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=9565
413 [There should be a GDB testcase - cagney]
417 GDB5 TOT on unixware 7
418 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00119.html
421 > I just spun the top of tree of the GDB5 branch on UnixWare 7. As a
422 > practical matter, the current thread support is somewhat more annoying
423 > than when GDB was thread-unaware.
427 Migrate qfThreadInfo packet -> qThreadInfo. (Andrew Cagney)
429 Add support for packet enable/disable commands with these thread
430 packets. General cleanup.
432 [PATCH] Document the ThreadInfo remote protocol queries
433 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00832.html
435 [PATCH] "info threads" queries for remote.c
436 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00831.html
443 New languages come onto the scene all the time.
447 Pascal (Pierre Muller, David Taylor)
449 Pierre Muller has contributed patches for adding Pascal Language
452 2 pascal language patches inserted in database
453 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00521.html
456 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00496.html
460 Java (Anthony Green, David Taylor)
462 Anthony Green has a number of Java patches that did not make it into
466 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00512.html
469 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00515.html
471 Patch: handle N_MAIN stab
472 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00527.html
478 Modify gdb to work correctly with Pascal.
482 Re: Various C++ things
484 value_headof/value_from_vtable_info are worthless, and should be removed.
485 The one place in printcmd.c that uses it should use the RTTI functions.
487 RTTI for g++ should be using the typeinfo functions rather than the vtables.
488 The typeinfo functions are always at offset 4 from the beginning of the vtable,
489 and are always right. The vtables will have weird names like E::VB sometimes.
490 The typeinfo function will always be "E type_info function", or somesuch.
492 value_virtual_fn_field needs to be fixed so there are no failures for virtual
493 functions for C++ using g++.
495 Testsuite cases are the major priority right now for C++ support, since i have
496 to make a lot of changes that could potentially break each other.
500 Add support for Modula3
502 Get DEC/Compaq to contribute their Modula-3 support.
506 Remote Protocol Support
507 =======================
511 set/show remote X-packet ...
513 ``(gdb) help set remote X-packet'' doesn't list the applicable
514 responses. The help message needs to be expanded.
518 Remote protocol doco feedback.
520 Too much feedback to mention needs to be merged in (901660). Search
521 for the word ``remote''.
524 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00023.html
525 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00056.html
526 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00382.html
530 GDB doesn't recover gracefully from remote protocol errors.
532 GDB wasn't checking for NAKs from the remote target. Instead a NAK is
533 ignored and a timeout is required before GDB retries. A pre-cursor to
534 fixing this this is making GDB's remote protocol packet more robust.
536 While downloading to a remote protocol target, gdb ignores packet
537 errors in so far as it will continue to edownload with chunk N+1 even
538 if chunk N was not correctly sent. This causes gdb.base/remote.exp to
539 take a painfully long time to run. As a PS that test needs to be
540 fixed so that it builds on 16 bit machines.
544 Add the cycle step command.
546 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00237.html
553 If / when GDB starts to support the debugging of multi-processor
554 (rather than multi-thread) applications the symtab code will need to
555 be updated a little so that several independant symbol tables are
556 active at a given time.
558 The other interesting change is a clarification of the exact meaning
559 of CORE_ADDR and that has had consequences for a few targets (that
560 were abusing that data type).
564 Investiagate ways of reducing memory.
568 Investigate ways of improving load time.
572 Get the d10v to use POINTER_TO_ADDRESS and ADDRESS_TO_POINTER.
574 Consequence of recent symtab clarification. No marks for figuring out
575 who maintains the d10v.
579 Get the MIPS to correctly sign extend all address <-> pointer
582 Consequence of recent symtab clarification. No marks for figuring out
583 who maintains the MIPS.
587 Architectural Changes: General
588 ==============================
590 These are harder than simple cleanups / fixes and, consequently
591 involve more work. Typically an Architectural Change will be broken
592 down into a more digestible set of cleanups and fixes.
596 Cleanup software single step.
598 At present many targets implement software single step by directly
599 blatting memory (see rs6000-tdep.c). Those targets should register
600 the applicable breakpoints using the breakpoint framework. Perhaphs a
601 new internal breakpoint class ``step'' is needed.
605 Replace READ_FP() with FRAME_HANDLE().
607 READ_FP() is a hangover from the days of the vax when the ABI really
608 did have a frame pointer register. Modern architectures typically
609 construct a virtual frame-handle from the stack pointer and various
610 other bits of string.
612 Unfortunatly GDB still treats this synthetic FP register as though it
613 is real. That in turn really confuses users (arm and ``print $fp'' VS
614 ``info registers fp''). The synthetic FP should be separated out of
615 the true register set presented to the user.
619 Register Cache Cleanup (below from Andrew Cagney)
621 I would depict the current register architecture as something like:
628 register + REGISTER_BYTE(reg_nr)
631 -------------------------
632 | extern register[] |
633 -------------------------
635 where neither the high (valops.c et.al.) or low gdb (*-tdep.c) are
636 really clear on what mechanisms they should be using to manipulate that
637 buffer. Further, much code assumes, dangerously, that registers are
638 contigious. Having got mips-tdep.c to support multiple ABIs, believe
639 me, that is a bad assumption. Finally, that register cache layout is
640 determined by the current remote/local target and _not_ the less
641 specific target ISA. In fact, in many cases it is determined by the
642 somewhat arbitrary layout of the [gG] packets!
645 How I would like the register file to work is more like:
673 The main objectives being:
675 o a clear separation between the low
676 level target and the high level GDB
678 o a mechanism that solves the general
679 problem of register aliases, overlaps
680 etc instead of treating them as optional
681 extras that can be wedged in as an after
682 thought (that is a reasonable description
683 of the current code).
685 Identify then solve the hard case and the
686 rest just falls out. GDB solved the easy
687 case and then tried to ignore the real
690 o a removal of the assumption that the
691 mapping between the register cache
692 and virtual registers is largely static.
693 If you flip the USR/SSR stack register
694 select bit in the status-register then
695 the corresponding stack registers should
698 o a mechanism that clearly separates the
699 gdb internal register cache from any
700 target (not architecture) dependant
701 specifics such as [gG] packets.
703 Of course, like anything, it sounds good in theory. In reality, it
704 would have to contend with many<->many relationships at both the
705 virt<->cache and cache<->target level. For instance:
708 Modifying an mmx register may involve
709 scattering values across both FP and
710 mmpx specific parts of a buffer
713 When writing back a SP it may need to
714 both be written to both SP and USP.
719 Rather than let this like the last time it was discussed, just slip, I'm
720 first going to add this e-mail (+ references) to TODO. I'd then like to
721 sketch out a broad strategy I think could get us there.
724 First thing I'd suggest is separating out the ``extern registers[]''
725 code so that we can at least identify what is using it. At present
726 things are scattered across many files. That way we can at least
727 pretend that there is a cache instead of a global array :-)
729 I'd then suggest someone putting up a proposal for the pseudo-reg /
730 high-level side interface so that code can be adopted to it. For old
731 code, initially a blanket rename of write_register_bytes() to
732 deprecated_write_register_bytes() would help.
734 Following that would, finaly be the corresponding changes to the target.
738 Check that GDB can handle all BFD architectures (Andrew Cagney)
740 There should be a test that checks that BFD/GDB are in sync with
741 regard to architecture changes. Something like a test that first
742 queries GDB for all supported architectures and then feeds each back
743 to GDB.. Anyone interested in learning how to write tests? :-)
747 Architectural Change: Multi-arch et al.
748 =======================================
750 The long term objective is to remove all assumptions that there is a
751 single target with a single address space with a single instruction
752 set architecture and single application binary interface.
754 This is an ongoing effort. The first milestone is to enable
755 ``multi-arch'' where by all architectural decisions are made at
758 It should be noted that ``gdbarch'' is really ``gdbabi'' and
759 ``gdbisa''. Once things are multi-arched breaking that down correctly
760 will become much easier.
764 GDBARCH cleanup (Andrew Cagney)
766 The non-generated parts of gdbarch.{sh,h,c} should be separated out
767 into arch-utils.[hc].
769 Document that gdbarch_init_ftype could easily fail because it didn't
770 identify an architecture.
774 Fix BELIEVE_PPC_PROMOTION. Change it to BELIEVE_PPC_PROMOTION_P?
776 At present there is still #ifdef BELIEVE_PPC_PROMOTION code in the
781 Fix ``set architecture <tab>''
783 This command should expand to a list of all supported architectures.
784 At present ``info architecture'' needs to be used. That is simply
785 wrong. It involves the use of add_set_enum_cmd().
789 Update ALPHA so that it uses ``struct frame_extra_info'' instead of
792 This is a barrier to replacing mips_extra_func_info with something
793 that works with multi-arch.
797 Multi-arch mips_extra_func_info.
799 This first needs the alpha to be updated so that it uses ``struct
804 Rationalize TARGET_SINGLE_FORMAT and TARGET_SINGLE_BIT et al.
806 Surely one of them is redundant.
810 Convert ALL architectures to MULTI-ARCH.
814 Select the initial multi-arch ISA / ABI based on --target or similar.
816 At present the default is based on what ever is first in the BFD
817 archures table. It should be determined based on the ``--target=...''
824 Enable the code to recognize --enable-targets=.... like BINUTILS does.
826 Can the tm.h and nm.h files be eliminated by multi-arch.
830 Architectural Change: MI, LIBGDB and scripting languages
831 ========================================================
833 See also architectural changes related to the event loop. LIBGDB
834 can't be finished until there is a generic event loop being used by
837 The long term objective is it to be possible to integrate GDB into
842 Implement generic ``(gdb) commmand > file''
844 Once everything is going through ui_file it should be come fairly
847 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00104.html
851 Replace gdb_stdtarg with gdb_targout (and possibly gdb_targerr).
853 gdb_stdtarg is easily confused with gdb_stdarg.
857 Extra ui_file methods - dump.
859 These are for debugging / testing. An aside is to set up a whitebox
860 testsuite for key internals such as ui_file.
864 Eliminate error_begin().
866 With ui_file, there is no need for the statefull error_begin ()
871 Send normal output to gdb_stdout.
872 Send error messages to gdb_stderror.
873 Send debug and log output log gdb_stdlog.
875 GDB still contains many cases where (f)printf or printf_filtered () is
876 used when it should be sending the messages to gdb_stderror or
877 gdb_stdlog. The thought of #defining printf to something has crossed
882 Re-do GDB's output pager.
884 GDB's output pager still relies on people correctly using *_filtered
885 for gdb_stdout and *_unfiltered for gdb_stdlog / gdb_stderr.
886 Hopefully, with all normal output going to gdb_stdout, the pager can
887 just look at the ui_file that the output is on and then use that to
888 decide what to do about paging. Sounds good in theory.
892 Check/cleanup MI documentation.
894 The list of commands specified in the documentation needs to be
895 checked against the mi-cmds.c table in a mechanical way (so that they
896 two can be kept up-to-date).
900 Convert MI into libgdb
902 MI provides a text interface into what should be many of the libgdb
903 functions. The implementation of those functions should be separated
904 into the MI interface and the functions proper. Those functions being
905 moved to gdb/lib say.
911 The first part can already be found in defs.h.
915 MI's input does not use buffering.
917 At present the MI interface reads raw characters of from an unbuffered
918 FD. This is to avoid several nasty buffer/race conditions. That code
919 should be changed so that it registers its self with the event loop
920 (on the input FD) and then push commands up to MI as they arrive.
922 The serial code already does this.
926 Make MI interface accessable from existing CLI.
930 Add a breakpoint-edit command to MI.
932 It would be similar to MI's breakpoint create but would apply to an
933 existing breakpoint. It saves the need to delete/create breakpoints
934 when ever they are changed.
938 Add directory path to MI breakpoint.
940 That way the GUI's task of finding the file within which the
941 breakpoint was set is simplified.
945 Add a mechanism to reject certain expression classes to MI
947 There are situtations where you don't want GDB's expression
948 parser/evaluator to perform inferior function calls or variable
949 assignments. A way of restricting the expression parser so that such
950 operations are not accepted would be very helpful.
954 Remove sideffects from libgdb breakpoint create function.
956 The user can use the CLI to create a breakpoint with partial
957 information - no file (gdb would use the file from the last
960 The libgdb interface currently affects that environment which can lead
961 to confusion when a user is setting breakpoints via both the MI and
964 This is also a good example of how getting the CLI ``right'' will be
969 Move gdb_lasterr to ui_out?
971 The way GDB throws errors and records them needs a re-think. ui_out
972 handles the correct output well. It doesn't resolve what to do with
973 output / error-messages when things go wrong.
977 Architectural Change: Async
978 ===========================
980 While GDB uses an event loop when prompting the user for input. That
981 event loop is not exploited by targets when they allow the target
982 program to continue. Typically targets still block in (target_wait())
983 until the program again halts.
985 The closest a target comes to supporting full asynchronous mode are
986 the remote targets ``async'' and ``extended-async''.
990 Asynchronous expression evaluator
992 Inferior function calls hang GDB.
996 Fix implementation of ``target xxx''.
998 At present when the user specifies ``target xxxx'', the CLI maps that
999 directly onto a target open method. It is then assumed that the
1000 target open method should do all sorts of complicated things as this
1001 is the only chance it has. Check how the various remote targets
1002 duplicate the target operations. Check also how the various targets
1003 behave differently for purely arbitrary reasons.
1005 What should happen is that ``target xxxx'' should call a generic
1006 ``target'' function and that should then co-ordinate the opening of
1007 ``xxxx''. This becomes especially important when you're trying to
1008 open an asynchronous target that may need to perform background tasks
1009 as part of the ``attach'' phase.
1011 Unfortunatly, due to limitations in the old/creaking command.h
1012 interface, that isn't possible. The function being called isn't told
1013 of the ``xxx'' or any other context information.
1015 Consequently a precursor to fixing ``target xxxx'' is to clean up the
1016 CLI code so that it passes to the callback function (attatched to a
1017 command) useful information such as the actual command and a context
1018 for that command. Other changes such as making ``struct command''
1019 opaque may also help.
1023 Make "target xxx" command interruptible.
1025 As things become async this becomes possible. A target would start
1026 the connect and then return control to the event loop. A cntrl-c
1027 would notify the target that the operation is to be abandoned and the
1028 target code could respond.
1032 Add a "suspend" subcommand of the "continue" command to suspend gdb
1033 while continuing execution of the subprocess. Useful when you are
1034 debugging servers and you want to dodge out and initiate a connection
1035 to a server running under gdb.
1044 This list is not up to date, and opinions vary about the importance or
1045 even desirability of some of the items. If you do fix something, it
1046 always pays to check the below.
1050 @c This does not work (yet if ever). FIXME.
1051 @c @item --parse=@var{lang} @dots{}
1052 @c Configure the @value{GDBN} expression parser to parse the listed languages.
1053 @c @samp{all} configures @value{GDBN} for all supported languages. To get a
1054 @c list of all supported languages, omit the argument. Without this
1055 @c option, @value{GDBN} is configured to parse all supported languages.
1059 START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that
1060 is its default value. Clean this up.
1064 It should be possible to use symbols from shared libraries before we know
1065 exactly where the libraries will be loaded. E.g. "b perror" before running
1066 the program. This could maybe be done as an extension of the "breakpoint
1067 re-evaluation" after new symbols are loaded.
1071 Make single_step() insert and remove breakpoints in one operation.
1075 Speed up single stepping by avoiding extraneous ptrace calls.
1079 Speed up single stepping by not inserting and removing breakpoints
1080 each time the inferior starts and stops.
1082 Breakpoints should not be inserted and deleted all the time. Only the
1083 one(s) there should be removed when we have to step over one. Support
1084 breakpoints that don't have to be removed to step over them.
1086 [this has resulted in numerous debates. The issue isn't clear cut]
1090 Provide "voodoo" debugging of core files. This creates a zombie
1091 process as a child of the debugger, and loads it up with the data,
1092 stack, and regs of the core file. This allows you to call functions
1093 in the executable, to manipulate the data in the core file.
1099 GDB reopens the source file on every line, as you "next" through it.
1101 [still true? I've a memory of this being fixed]
1105 Perhaps "i source" should take an argument like that of "list".
1109 Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if
1110 it matches the source line indicated.
1114 The prompt at end of screen should accept space as well as CR.
1118 Backtrace should point out what the currently selected frame is, in
1119 its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar,
1120 ...)" rather than "#3 foo (bar, ...)".
1124 "i program" should work for core files, and display more info, like what
1125 actually caused it to die.
1129 "x/10i" should shorten the long name, if any, on subsequent lines.
1133 "next" over a function that longjumps, never stops until next time you happen
1134 to get to that spot by accident. E.g. "n" over execute_command which has
1139 "set zeroprint off", don't bother printing members of structs which
1140 are entirely zero. Useful for those big structs with few useful
1145 GDB does four ioctl's for every command, probably switching terminal modes
1146 to/from inferior or for readline or something.
1150 terminal_ours versus terminal_inferior: cache state. Switch should be a noop
1151 if the state is the same, too.
1155 "i frame" shows wrong "arglist at" location, doesn't show where the args
1156 should be found, only their actual values.
1160 There should be a way for "set" commands to validate the new setting
1161 before it takes effect.
1165 "ena d" is ambiguous, why? "ena delete" seems to think it is a command!
1169 i line VAR produces "Line number not known for symbol ``var''.". I
1170 thought we were stashing that info now!
1174 We should be able to write to random files at hex offsets like adb.
1178 [elena - delete this]
1180 Handle add_file with separate text, data, and bss addresses. Maybe
1181 handle separate addresses for each segment in the object file?
1185 [Jimb/Elena delete this one]
1187 Handle free_named_symtab to cope with multiply-loaded object files
1188 in a dynamic linking environment. Should remember the last copy loaded,
1189 but not get too snowed if it finds references to the older copy.
1193 [elena delete this also]
1195 Remove all references to:
1202 now that we have BFD. All remaining are in machine dependent files.
1206 Re-organize help categories into things that tend to fit on a screen
1211 Add in commands like ADB's for searching for patterns, etc. We should
1212 be able to examine and patch raw unsymboled binaries as well in gdb as
1213 we can in adb. (E.g. increase the timeout in /bin/login without source).
1215 [actually, add ADB interface :-]
1219 When doing "step" or "next", if a few lines of source are skipped between
1220 the previous line and the current one, print those lines, not just the
1221 last line of a multiline statement.
1225 Handling of "&" address-of operator needs some serious overhaul
1226 for ANSI C and consistency on arrays and functions.
1227 For "float point[15];":
1228 ptype &point[4] ==> Attempt to take address of non-lvalue.
1229 For "char *malloc();":
1230 ptype malloc ==> "char *()"; should be same as
1231 ptype &malloc ==> "char *(*)()"
1232 call printf ("%x\n", malloc) ==> weird value, should be same as
1233 call printf ("%x\n", &malloc) ==> correct value
1237 Fix dbxread.c symbol reading in the presence of interrupts. It
1238 currently leaves a cleanup to blow away the entire symbol table when a
1239 QUIT occurs. (What's wrong with that? -kingdon, 28 Oct 1993).
1241 [I suspect that the grype was that, on a slow system, you might want
1242 to cntrl-c and get just half the symbols and then load the rest later
1243 - scary to be honest]
1247 Mipsread.c reads include files depth-first, because the dependencies
1248 in the psymtabs are way too inclusive (it seems to me). Figure out what
1249 really depends on what, to avoid recursing 20 or 30 times while reading
1254 value_add() should be subtracting the lower bound of arrays, if known,
1255 and possibly checking against the upper bound for error reporting.
1259 When listing source lines, check for a preceding \n, to verify that
1260 the file hasn't changed out from under us.
1262 [fixed by some other means I think. That hack wouldn't actually work
1263 reliably - the file might move such that another \n appears. ]
1267 Get all the remote systems (where the protocol allows it) to be able to
1268 stop the remote system when the GDB user types ^C (like remote.c
1269 does). For ebmon, use ^Ak.
1273 Possible feature: A version of the "disassemble" command which shows
1274 both source and assembly code ("set symbol-filename on" is a partial
1277 [has this been done? It was certainly done for MI and GDBtk]
1281 investigate "x/s 0" (right now stops early) (I think maybe GDB is
1282 using a 0 address for bad purposes internally).
1286 Make "info path" and path_command work again (but independent of the
1287 environment either of gdb or that we'll pass to the inferior).
1291 Make GDB understand the GCC feature for putting octal constants in
1292 enums. Make it so overflow on an enum constant does not error_type
1293 the whole type. Allow arbitrarily large enums with type attributes.
1294 Put all this stuff in the testsuite.
1298 Make TYPE_CODE_ERROR with a non-zero TYPE_LENGTH more useful (print
1299 the value in hex; process type attributes). Add this to the
1300 testsuite. This way future compilers can add new types and old
1301 versions of GDB can do something halfway reasonable.
1305 Fix mdebugread.c:parse_type to do fundamental types right (see
1306 rs6000_builtin_type in stabsread.c for what "right" is--the point is
1307 that the debug format fixes the sizes of these things and it shouldn't
1308 depend on stuff like TARGET_PTR_BIT and so on. For mdebug, there seem
1309 to be separate bt* codes for 64 bit and 32 bit things, and GDB should
1310 be aware of that). Also use a switch statement for clarity and speed.
1314 Investigate adding symbols in target_load--some targets do, some
1319 Put dirname in psymtabs and change lookup*symtab to use dirname (so
1320 /foo/bar.c works whether compiled by cc /foo/bar.c, or cd /foo; cc
1325 Merge xcoffread.c and coffread.c. Use breakpoint_re_set instead of
1330 Make a watchpoint which contains a function call an error (it is
1331 broken now, making it work is probably not worth the effort).
1335 New test case based on weird.exp but in which type numbers are not
1336 renumbered (thus multiply defining a type). This currently causes an
1337 infinite loop on "p v_comb".
1341 [Hey! Hint Hint Delete Delete!!!]
1343 Fix 386 floating point so that floating point registers are real
1344 registers (but code can deal at run-time if they are missing, like
1345 mips and 68k). This would clean up "info float" and related stuff.
1349 gcc -g -c enummask.c then gdb enummask.o, then "p v". GDB complains
1350 about not being able to access memory location 0.
1352 -------------------- enummask.c
1373 If try to modify value in file with "set write off" should give
1374 appropriate error not "cannot access memory at address 0x65e0".
1378 Allow core file without exec file on RS/6000.
1382 Make sure "shell" with no arguments works right on DOS.
1386 Make gdb.ini (as well as .gdbinit) be checked on all platforms, so
1387 the same directory can be NFS-mounted on unix or DOS, and work the
1392 [Is this another delete???]
1394 Get SECT_OFF_TEXT stuff out of objfile_relocate (might be needed to
1395 get RS/6000 to work right, might not be immediately relevant).
1399 Work out some kind of way to allow running the inferior to be done as
1400 a sub-execution of, eg. breakpoint command lists. Currently running
1401 the inferior interupts any command list execution. This would require
1402 some rewriting of wait_for_inferior & friends, and hence should
1403 probably be done in concert with the above.
1407 Add function arguments to gdb user defined functions.
1411 Add convenience variables that refer to exec file, symbol file,
1412 selected frame source file, selected frame function, selected frame
1417 Modify the handling of symbols grouped through BINCL/EINCL stabs to
1418 allocate a partial symtab for each BINCL/EINCL grouping. This will
1419 seriously decrease the size of inter-psymtab dependencies and hence
1420 lessen the amount that needs to be read in when a new source file is
1425 Add a command for searching memory, a la adb. It specifies size,
1426 mask, value, start address. ADB searches until it finds it or hits
1427 an error (or is interrupted).
1431 Remove the range and type checking code and documentation, if not