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.
10 Below is a list of problems identified during the GDB 5.0 release
11 cycle. People hope to have these problems fixed in 5.1.
15 Hardware watchpint problems on x86 OSes, including Linux:
17 1. Delete/disable hardware watchpoints should free hardware debug
19 2. Watch for different values on a viariable with one hardware debug
22 According to Eli Zaretskii <eliz@delorie.com>:
24 These are not GDB/ia32 issues per se: the above features are all
25 implemented in the DJGPP port of GDB and work in v5.0. Every
26 x86-based target should be able to lift the relevant parts of
27 go32-nat.c and use them almost verbatim. You get debug register
28 sharing through reference counts, and the ability to watch large
29 regions (up to 16 bytes) using multiple registers. (The required
30 infrastructure in high-level GDB application code, mostly in
31 breakpoint.c, is also working since v5.0.)
35 RFD: infrun.c: No bpstat_stop_status call after proceed over break?
36 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00665.html
38 GDB misses watchpoint triggers after proceeding over a breakpoint on
43 x86 linux GDB and SIGALRM (???)
44 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00803.html
46 This problem has been fixed, but a regression test still needs to be
47 added to the testsuite:
48 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00309.html
54 Can't build IRIX -> arm GDB.
55 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00356.html
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.
66 Problem with weak functions
67 http://sourceware.cygnus.com/ml/gdb/2000-05/msg00060.html
69 Dan Nicolaescu writes:
70 > It seems that gdb-4.95.1 does not display correctly the function when
71 > stoping in weak functions.
73 > It stops in a function that is defined as weak, not in the function
74 > that is actually run...
78 GDB 5.0 doesn't work on Linux/SPARC
82 Thread support. Right now, as soon as a thread finishes and exits,
83 you're hosed. This problem is reported once a week or so.
87 Wow, three bug reports for the same problem in one day! We should
88 probably make fixing this a real priority :-).
90 Anyway, thanks for reporting.
92 The following patch will fix the problems with setting breakpoints in
93 dynamically loaded objects:
95 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00230.html
97 This patch isn't checked in yet (ping Michael/JimB), but I hope this
98 will be in the next GDB release.
100 There should really be a test in the testsuite for this problem, since
101 it keeps coming up :-(. Any volunteers?
108 http://sources.redhat.com/ml/gdb/2000-07/msg00038.html
110 Is the Solaris 8 x86 problem fixed? When you configure it, configure
111 incorrectly determines that I have no curses.h. This causes mucho
112 compilation errors later on.
114 Simply editing the config.h to define CURSES_H fixes the problem, and
115 then the build works fine.
117 The status for this problem:
119 Solaris 8 x86 (PIII-560)
122 I had the same problem with several of the snapshots shortly before
123 5.0 became official, and 5.0 has the same problem.
125 I sent some mail in about it long ago, and never saw a reply.
127 I haven't had time to figure it out myself, especially since I get all
128 confused trying to figure out what configure does, I was happy to find
135 GDB 5.1 - New features
136 ======================
138 The following new features should be included in 5.1.
142 Enable MI by default. Old code can be deleted after 5.1 is out.
146 Pascal (Pierre Muller, David Taylor)
148 Pierre Muller has contributed patches for adding Pascal Language
151 2 pascal language patches inserted in database
152 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00521.html
155 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00496.html
159 Java (Anthony Green, David Taylor)
161 Anthony Green has a number of Java patches that did not make it into
162 the 5.0 release. The first two are in cvs now, but the third needs
163 some fixing up before it can go in.
166 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00512.html
169 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00515.html
171 Patch: handle N_MAIN stab
172 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00527.html
178 Modify gdb to work correctly with Pascal.
182 Revised UDP support (was: Re: [Fwd: [patch] UDP transport support])
183 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00000.html
185 (Broken) support for GDB's remote protocol across UDP is to be
186 included in the follow-on release.
188 It should be noted that UDP can only work when the [Gg] packet fits in
191 There is also much debate over the merit of this.
198 The following code cleanups will hopefully be applied to GDB 5.1.
202 Change documentation to GFDL license.
204 ``It is time to make an effort to start using the GFDL more
205 thoroughly. Would all GNU maintainers please change the license to
206 the GFDL, for all manuals and other major documentation files?
208 The GFDL and some instructions for using it can be found in
209 http://www.gnu.org/copyleft/''
215 Delete macro TARGET_BYTE_ORDER_SELECTABLE.
217 Patches in the database.
221 Fix copyright notices.
223 Turns out that ``1998-2000'' isn't considered valid :-(
225 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00467.html
231 Eliminate all uses of PARAMS in GDB's source code.
235 printcmd.c (print_address_numeric):
237 NOTE: This assumes that the significant address information is kept in
238 the least significant bits of ADDR - the upper bits were either zero
239 or sign extended. Should ADDRESS_TO_POINTER() or some
240 ADDRESS_TO_PRINTABLE() be used to do the conversion?
246 Eliminate all warnings for at least one host/target for the flags:
247 -Wimplicit -Wreturn-type -Wcomment -Wtrigraphs -Wformat -Wparentheses
248 -Wpointer-arith -Wuninitialized
252 Follow through `make check' with --enable-shared.
254 When the srcware tree is configured with --enable-shared, the `expect'
255 program won't run properly. Jim Wilson found out gdb has a local hack
256 to set LD_LIBRARY_PATH, but, AFAIK, no other project has been hacked
259 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00845.html
268 Fix at least one thread bug.
272 GDB 5.2 - New features
273 ======================
277 Objective C/C++ Support. Bu hopefully sooner...
284 The following cleanups have been identified as part of GDB 5.2.
288 Eliminate more compiler warnings.
292 Restructure gdb directory tree so that it avoids any 8.3 and 14
297 Convert GDB build process to AUTOMAKE.
299 See also sub-directory configure below.
301 The current convention is (kind of) to use $(<header>_h) in all
302 dependency lists. It isn't done in a consistent way.
306 Code Cleanups: General
307 ======================
309 The following are more general cleanups and fixes. They are not tied
310 to any specific release.
314 The BFD directory requires bug-fixed AUTOMAKE et.al.
316 AUTOMAKE 1.4 incorrectly set the TEXINPUTS environment variable. It
317 contained the full path to texinfo.tex when it should have only
318 contained the directory. The bug has been fixed in the current
319 AUTOMAKE sources. Automake snapshots can be found in:
320 ftp://sourceware.cygnus.com/pub/gdb/snapshots
321 and ftp://sourceware.cygnus.com/pub/binutils
325 Find something better than DEFAULT_BFD_ARCH, DEFAULT_BFD_VEC to
326 determine the default isa/byte-order.
330 Rely on BFD_BIG_ENDIAN and BFD_LITTLE_ENDIAN instead of host dependent
331 BIG_ENDIAN and LITTLE_ENDIAN.
335 Eliminate more compiler warnings.
337 Of course there also needs to be the usual debate over which warnings
338 are valid and how to best go about this.
340 One method: choose a single option; get agreement that it is
341 reasonable; try it out to see if there isn't anything silly about it
342 (-Wunused-parameters is an example of that) then incrementally hack
345 The other method is to enable all warnings and eliminate them from one
350 Elimination of ``(catch_errors_ftype *) func''.
352 Like make_cleanup_func it isn't portable.
353 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00791.html
354 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00814.html
358 Nuke #define CONST_PTR.
366 [PATCH/5] src/intl/Makefile.in:distclean additions
367 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00363.html
369 Do not forget to merge the patch back into the trunk.
373 Rationalize the host-endian code (grep for HOST_BYTE_ORDER).
375 At present defs.h includes <endian.h> (which is linux specific) yet
376 almost nothing depends on it. Suggest "gdb_endian.h" which can also
377 handle <machine/endian.h> and only include that where it is really
382 Replace strsave() + mstrsave() with libiberty:xstrdup().
386 Replace savestring() with something from libiberty.
388 An xstrldup()? but that would have different semantics.
392 Rationalize use of floatformat_unknown in GDB sources.
394 Instead of defaulting to floatformat_unknown, should hosts/targets
395 specify the value explicitly?
397 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
401 Add a ``name'' member to include/floatformat.h:struct floatformat.
402 Print that name in gdbarch.c.
406 Sort out the harris mess in include/floatformat.h (it hardwires two
407 different floating point formats).
411 See of the GDB local floatformat_do_doublest() and libiberty's
412 floatformat_to_double (which was once GDB's ...) can be merged some
417 Eliminate mmalloc() from GDB.
419 Also eliminate it from defs.h.
423 Eliminate PTR. ISO-C allows ``void *''.
429 GDB should never abort. GDB should either throw ``error ()'' or
430 ``internal_error ()''. Better still GDB should naturally unwind with
435 Add __LINE__ and __FILE__ to internal_error().
439 GDB probably doesn't build on FreeBSD pre 2.2.x
440 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00378.html
442 Fixes to get FreeBSD working on 2.2.x, 3.x and 4.x caused the code to
447 Deprecate "fg". Apparently ``fg'' is actually continue.
449 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00417.html
453 Deprecate current use of ``floatformat_unknown''.
455 Require all targets to explicitly provide their float format instead
456 of defaulting to floatformat unknown. Doing the latter leads to nasty
459 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
463 Rationalize floatformat_to_double() vs floatformat_to_doublest().
465 Looks like GDB migrated floatformat_to_double() to libiberty but then
466 turned around and created a ..._to_doublest() the latter containing
469 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00472.html
473 Move floatformat_ia64_ext to libiberty/include floatformat.[ch].
475 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00466.html
479 The ``maintenance deprecate set endian big'' command doesn't notice
480 that it is deprecating ``set endian'' and not ``set endian big'' (big
481 is implemented using an enum). Is anyone going to notice this?
485 When tab expanding something like ``set arch<tab>'' ignore the
486 deprecated ``set archdebug'' and expand to ``set architecture''.
490 Eliminate ``arm_register_names[j] = (char *) regnames[j]'' and the
491 like from arm-tdep.c.
495 Fix uses of ->function.cfunc = set_function().
497 The command.c code calls sfunc() when a set command. Rather than
498 change it suggest fixing the callback function so that it is more
501 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
503 See also ``Fix implementation of ``target xxx''.'' below.
507 IRIX 3.x support is probably broken.
511 Delete sim/SIM_HAVE_BREAKPOINTS and gdb/SIM_HAS_BREAKPOINTS.
512 http://sourceware.cygnus.com/ml/gdb-patches/2000-07/msg00042.html
514 Apart from the d30v, are there any sim/common simulators that make use
517 A brief summary of what happened is that sim/common/sim-break.c was
518 created as a good idea. It turned out a better idea was to use
519 SIM_SIGBREAK and have GDB pass back sim_resume (..., SIGBREAK).
523 Move remote_remove_hw_breakpoint, remote_insert_hw_breakpoint,
524 remote_remove_watchpoint, remote_insert_watchpoint into target vector.
528 Eliminate ``extern'' from C files.
532 Replace ``STREQ()'' et.al. with ``strcmp() == 0'' et.al.
534 Extreme care is recommeded - perhaps only modify tests that are
535 exercised by the testsuite (as determined using some type of code
540 New Features and Fixes
541 ======================
543 These are harder than cleanups but easier than work involving
544 fundamental architectural change.
548 Add built-by, build-date, tm, xm, nm and anything else into gdb binary
549 so that you can see how the GDB was created.
553 Add an "info bfd" command that displays supported object formats,
554 similarly to objdump -i.
556 Is there a command already?
560 Fix ``I'm sorry, Dave, I can't do that.'' from symfile.c.
562 This requires internationalization.
568 (gdb) p fwprintf(stdout,L"%S\n", f)
569 No symbol "L" in current context.
573 Cleanup configury support for optional sub-directories.
575 Check how GCC handles multiple front ends for an example of how things
576 could work. A tentative first step is to rationalize things so that
577 all sub directories are handled in a fashion similar to gdb/mi.
579 See also automake above.
583 Add a transcript mechanism to GDB.
585 Such a mechanism might log all gdb input and output to a file in a
586 form that would allow it to be replayed. It could involve ``gdb
587 --transcript=FILE'' or it could involve ``(gdb) transcript file''.
591 Can the xdep files be replaced by autoconf?
595 Document trace machinery
599 Document ui-out and ui-file.
601 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00121.html
605 Update texinfo.tex to latest?
609 Incorporate agentexpr.texi into gdb.texinfo
611 agentexpr.texi mostly describes the details of the byte code used for
612 tracepoints, not the internals of the support for this in GDB. So it
613 looks like gdb.texinfo is a better place for this information.
615 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00566.html
619 Document overlay machinery.
623 ``(gdb) catch signal SIGNAL''
625 Overlaps with ``handle SIGNAL'' but the implied behavior is different.
626 You can attach commands to a catch but not a handle. A handle has a
627 limited number of hardwired actions.
631 Get the TUI working on all platforms.
635 Add support for ``gdb --- PROGRAM ARGS ...''.
636 Add support for ``gdb -cmd=...''
638 Along with many variations. Check:
640 ????? for a full discussion.
646 Implement ``(gdb) !ls''.
648 Which is very different from ``(gdb) ! ls''. Implementing the latter
651 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00034.html
655 Change the (char *list[]) to (const char (*)[]) so that dynamic lists can
660 When tab expanding something like ``set arch<tab>'' ignore the
661 deprecated ``set archdebug'' and expand to ``set architecture''.
665 Replace the code that uses the host FPU with an emulator of the target
670 The "ocd reset" command needs to flush the dcache, which requires breaking
671 the abstraction layer between the target independent and target code. One
672 way to address this is provide a generic "reset" command and target vector.
674 http://sources.redhat.com/ml/gdb-patches/2000-10/msg00011.html
683 Generic: lin-thread cannot handle thread exit (Mark Kettenis, Michael
684 Snyder) http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00525.html
686 The thread_db assisted debugging code doesn't handle exiting threads
687 properly, at least in combination with glibc 2.1.3 (the framework is
688 there, just not the actual code). There are at least two problems
689 that prevent this from working.
691 As an additional reference point, the pre thread_db code did not work
696 GNU/Linux/x86 and random thread signals (and Solaris/SPARC but not
698 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00336.html
700 Christopher Blizzard writes:
702 So, I've done some more digging into this and it looks like Jim
703 Kingdon has reported this problem in the past:
705 http://sourceware.cygnus.com/ml/bug-gdb/1999-10/msg00058.html
707 I can reproduce this problem both with and without Tom's patch. Has
708 anyone seen this before? Maybe have a solution for it hanging around?
711 There's a test case for this documented at:
713 when debugging threaded applications you get extra SIGTRAPs
714 http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=9565
716 [There should be a GDB testcase - cagney]
720 GDB5 TOT on unixware 7
721 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00119.html
724 > I just spun the top of tree of the GDB5 branch on UnixWare 7. As a
725 > practical matter, the current thread support is somewhat more annoying
726 > than when GDB was thread-unaware.
730 Migrate qfThreadInfo packet -> qThreadInfo. (Andrew Cagney)
732 Add support for packet enable/disable commands with these thread
733 packets. General cleanup.
735 [PATCH] Document the ThreadInfo remote protocol queries
736 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00832.html
738 [PATCH] "info threads" queries for remote.c
739 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00831.html
746 New languages come onto the scene all the time.
750 Re: Various C++ things
752 value_headof/value_from_vtable_info are worthless, and should be
753 removed. The one place in printcmd.c that uses it should use the RTTI
756 RTTI for g++ should be using the typeinfo functions rather than the
757 vtables. The typeinfo functions are always at offset 4 from the
758 beginning of the vtable, and are always right. The vtables will have
759 weird names like E::VB sometimes. The typeinfo function will always
760 be "E type_info function", or somesuch.
762 value_virtual_fn_field needs to be fixed so there are no failures for
763 virtual functions for C++ using g++.
765 Testsuite cases are the major priority right now for C++ support,
766 since i have to make a lot of changes that could potentially break
771 Add support for Modula3
773 Get DEC/Compaq to contribute their Modula-3 support.
777 Remote Protocol Support
778 =======================
782 Remote protocol doco feedback.
784 Too much feedback to mention needs to be merged in (901660). Search
785 for the word ``remote''.
788 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00023.html
789 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00056.html
790 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00382.html
794 GDB doesn't recover gracefully from remote protocol errors.
796 GDB wasn't checking for NAKs from the remote target. Instead a NAK is
797 ignored and a timeout is required before GDB retries. A pre-cursor to
798 fixing this this is making GDB's remote protocol packet more robust.
800 While downloading to a remote protocol target, gdb ignores packet
801 errors in so far as it will continue to download with chunk N+1 even
802 if chunk N was not correctly sent. This causes gdb.base/remote.exp to
803 take a painfully long time to run. As a PS that test needs to be
804 fixed so that it builds on 16 bit machines.
808 Add the cycle step command.
810 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00237.html
814 Resolve how to scale things to support very large packets.
818 Resolve how to handle a target that changes things like its endianess
819 on the fly - should it be returned in the ``T'' packet?
821 Underlying problem is that the register file is target endian. If the
822 target endianess changes gdb doesn't know.
826 Rename read_register{,_pid}() to read_unsigned_register{,_pid}().
833 If / when GDB starts to support the debugging of multi-processor
834 (rather than multi-thread) applications the symtab code will need to
835 be updated a little so that several independent symbol tables are
836 active at a given time.
838 The other interesting change is a clarification of the exact meaning
839 of CORE_ADDR and that has had consequences for a few targets (that
840 were abusing that data type).
844 Investiagate ways of reducing memory.
848 Investigate ways of improving load time.
852 Get the d10v to use POINTER_TO_ADDRESS and ADDRESS_TO_POINTER.
854 Consequence of recent symtab clarification. No marks for figuring out
855 who maintains the d10v.
859 Get the MIPS to correctly sign extend all address <-> pointer
862 Consequence of recent symtab clarification. No marks for figuring out
863 who maintains the MIPS.
867 GDB truncates 64 bit enums.
869 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00290.html
876 There are never to many testcases.
880 Better thread testsuite.
884 Better C++ testsuite.
888 Look at adding a GDB specific testsuite directory so that white box
889 tests of key internals can be added (eg ui_file).
893 Separate out tests that involve the floating point (FP).
895 (Something for people brining up new targets). FP and non-fp tests
896 are combined. I think there should be set of basic tests that
897 exercise pure integer support and then a more expanded set that
898 exercise FP and FP/integer interactions.
900 As an example, the MIPS, for n32 as problems with passing FP's and
901 structs. Since most inferior call tests include FP it is difficult to
902 determine of the integer tests are ok.
906 Architectural Changes: General
907 ==============================
909 These are harder than simple cleanups / fixes and, consequently
910 involve more work. Typically an Architectural Change will be broken
911 down into a more digestible set of cleanups and fixes.
915 Cleanup software single step.
917 At present many targets implement software single step by directly
918 blatting memory (see rs6000-tdep.c). Those targets should register
919 the applicable breakpoints using the breakpoint framework. Perhaphs a
920 new internal breakpoint class ``step'' is needed.
924 Replace READ_FP() with FRAME_HANDLE().
926 READ_FP() is a hangover from the days of the vax when the ABI really
927 did have a frame pointer register. Modern architectures typically
928 construct a virtual frame-handle from the stack pointer and various
929 other bits of string.
931 Unfortunately GDB still treats this synthetic FP register as though it
932 is real. That in turn really confuses users (arm and ``print $fp'' VS
933 ``info registers fp''). The synthetic FP should be separated out of
934 the true register set presented to the user.
938 Register Cache Cleanup (below from Andrew Cagney)
940 I would depict the current register architecture as something like:
947 register + REGISTER_BYTE(reg_nr)
950 -------------------------
951 | extern register[] |
952 -------------------------
954 where neither the high (valops.c et.al.) or low gdb (*-tdep.c) are
955 really clear on what mechanisms they should be using to manipulate that
956 buffer. Further, much code assumes, dangerously, that registers are
957 contigious. Having got mips-tdep.c to support multiple ABIs, believe
958 me, that is a bad assumption. Finally, that register cache layout is
959 determined by the current remote/local target and _not_ the less
960 specific target ISA. In fact, in many cases it is determined by the
961 somewhat arbitrary layout of the [gG] packets!
964 How I would like the register file to work is more like:
992 The main objectives being:
994 o a clear separation between the low
995 level target and the high level GDB
997 o a mechanism that solves the general
998 problem of register aliases, overlaps
999 etc instead of treating them as optional
1000 extras that can be wedged in as an after
1001 thought (that is a reasonable description
1002 of the current code).
1004 Identify then solve the hard case and the
1005 rest just falls out. GDB solved the easy
1006 case and then tried to ignore the real
1009 o a removal of the assumption that the
1010 mapping between the register cache
1011 and virtual registers is largely static.
1012 If you flip the USR/SSR stack register
1013 select bit in the status-register then
1014 the corresponding stack registers should
1017 o a mechanism that clearly separates the
1018 gdb internal register cache from any
1019 target (not architecture) dependent
1020 specifics such as [gG] packets.
1022 Of course, like anything, it sounds good in theory. In reality, it
1023 would have to contend with many<->many relationships at both the
1024 virt<->cache and cache<->target level. For instance:
1027 Modifying an mmx register may involve
1028 scattering values across both FP and
1029 mmpx specific parts of a buffer
1032 When writing back a SP it may need to
1033 both be written to both SP and USP.
1038 Rather than let this like the last time it was discussed, just slip, I'm
1039 first going to add this e-mail (+ references) to TODO. I'd then like to
1040 sketch out a broad strategy I think could get us there.
1043 First thing I'd suggest is separating out the ``extern registers[]''
1044 code so that we can at least identify what is using it. At present
1045 things are scattered across many files. That way we can at least
1046 pretend that there is a cache instead of a global array :-)
1048 I'd then suggest someone putting up a proposal for the pseudo-reg /
1049 high-level side interface so that code can be adopted to it. For old
1050 code, initially a blanket rename of write_register_bytes() to
1051 deprecated_write_register_bytes() would help.
1053 Following that would, finaly be the corresponding changes to the target.
1057 Check that GDB can handle all BFD architectures (Andrew Cagney)
1059 There should be a test that checks that BFD/GDB are in sync with
1060 regard to architecture changes. Something like a test that first
1061 queries GDB for all supported architectures and then feeds each back
1062 to GDB.. Anyone interested in learning how to write tests? :-)
1066 Architectural Change: Multi-arch et al.
1067 =======================================
1069 The long term objective is to remove all assumptions that there is a
1070 single target with a single address space with a single instruction
1071 set architecture and single application binary interface.
1073 This is an ongoing effort. The first milestone is to enable
1074 ``multi-arch'' where by all architectural decisions are made at
1077 It should be noted that ``gdbarch'' is really ``gdbabi'' and
1078 ``gdbisa''. Once things are multi-arched breaking that down correctly
1079 will become much easier.
1083 GDBARCH cleanup (Andrew Cagney)
1085 The non-generated parts of gdbarch.{sh,h,c} should be separated out
1086 into arch-utils.[hc].
1088 Document that gdbarch_init_ftype could easily fail because it didn't
1089 identify an architecture.
1093 Fix BELIEVE_PPC_PROMOTION. Change it to BELIEVE_PPC_PROMOTION_P?
1095 At present there is still #ifdef BELIEVE_PPC_PROMOTION code in the
1100 Fix target_signal_from_host() etc.
1102 The name is wrong for starters. ``target_signal'' should probably be
1103 ``gdb_signal''. ``from_host'' should be ``from_target_signal''.
1104 After that it needs to be multi-arched and made independent of any
1105 host signal numbering.
1109 Update ALPHA so that it uses ``struct frame_extra_info'' instead of
1112 This is a barrier to replacing mips_extra_func_info with something
1113 that works with multi-arch.
1117 Multi-arch mips_extra_func_info.
1119 This first needs the alpha to be updated so that it uses ``struct
1124 Rationalize TARGET_SINGLE_FORMAT and TARGET_SINGLE_BIT et al.
1126 Surely one of them is redundant.
1130 Convert ALL architectures to MULTI-ARCH.
1134 Select the initial multi-arch ISA / ABI based on --target or similar.
1136 At present the default is based on what ever is first in the BFD
1137 archures table. It should be determined based on the ``--target=...''
1142 Make MIPS pure multi-arch.
1144 It is only at the multi-arch enabled stage.
1150 Enable the code to recognize --enable-targets=.... like BINUTILS does.
1152 Can the tm.h and nm.h files be eliminated by multi-arch.
1156 Architectural Change: MI, LIBGDB and scripting languages
1157 ========================================================
1159 See also architectural changes related to the event loop. LIBGDB
1160 can't be finished until there is a generic event loop being used by
1163 The long term objective is it to be possible to integrate GDB into
1164 scripting languages.
1168 Implement generic ``(gdb) commmand > file''
1170 Once everything is going through ui_file it should be come fairly
1173 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00104.html
1177 Replace gdb_stdtarg with gdb_targout (and possibly gdb_targerr).
1179 gdb_stdtarg is easily confused with gdb_stdarg.
1183 Extra ui_file methods - dump.
1185 Very useful for whitebox testing.
1189 Eliminate error_begin().
1191 With ui_file, there is no need for the statefull error_begin ()
1196 Send normal output to gdb_stdout.
1197 Send error messages to gdb_stderror.
1198 Send debug and log output log gdb_stdlog.
1200 GDB still contains many cases where (f)printf or printf_filtered () is
1201 used when it should be sending the messages to gdb_stderror or
1202 gdb_stdlog. The thought of #defining printf to something has crossed
1207 Re-do GDB's output pager.
1209 GDB's output pager still relies on people correctly using *_filtered
1210 for gdb_stdout and *_unfiltered for gdb_stdlog / gdb_stderr.
1211 Hopefully, with all normal output going to gdb_stdout, the pager can
1212 just look at the ui_file that the output is on and then use that to
1213 decide what to do about paging. Sounds good in theory.
1217 Check/cleanup MI documentation.
1219 The list of commands specified in the documentation needs to be
1220 checked against the mi-cmds.c table in a mechanical way (so that they
1221 two can be kept up-to-date).
1225 Convert MI into libgdb
1227 MI provides a text interface into what should be many of the libgdb
1228 functions. The implementation of those functions should be separated
1229 into the MI interface and the functions proper. Those functions being
1230 moved to gdb/lib say.
1236 The first part can already be found in defs.h.
1240 MI's input does not use buffering.
1242 At present the MI interface reads raw characters of from an unbuffered
1243 FD. This is to avoid several nasty buffer/race conditions. That code
1244 should be changed so that it registers its self with the event loop
1245 (on the input FD) and then push commands up to MI as they arrive.
1247 The serial code already does this.
1251 Make MI interface accessible from existing CLI.
1255 Add a breakpoint-edit command to MI.
1257 It would be similar to MI's breakpoint create but would apply to an
1258 existing breakpoint. It saves the need to delete/create breakpoints
1259 when ever they are changed.
1263 Add directory path to MI breakpoint.
1265 That way the GUI's task of finding the file within which the
1266 breakpoint was set is simplified.
1270 Add a mechanism to reject certain expression classes to MI
1272 There are situtations where you don't want GDB's expression
1273 parser/evaluator to perform inferior function calls or variable
1274 assignments. A way of restricting the expression parser so that such
1275 operations are not accepted would be very helpful.
1279 Remove sideffects from libgdb breakpoint create function.
1281 The user can use the CLI to create a breakpoint with partial
1282 information - no file (gdb would use the file from the last
1285 The libgdb interface currently affects that environment which can lead
1286 to confusion when a user is setting breakpoints via both the MI and
1289 This is also a good example of how getting the CLI ``right'' will be
1294 Move gdb_lasterr to ui_out?
1296 The way GDB throws errors and records them needs a re-think. ui_out
1297 handles the correct output well. It doesn't resolve what to do with
1298 output / error-messages when things go wrong.
1302 do_setshow_command contains a 1024 byte buffer.
1304 The function assumes that there will never be any more than 1024 bytes
1305 of enum. It should use mem_file.
1309 Should struct cmd_list_element . completer take the command as an
1314 Should the bulk of top.c:line_completion_function() be moved to
1315 command.[hc]? complete_on_cmdlist() and complete_on_enums() could
1316 then be made private.
1320 top.c (execute_command): Should a command being valid when the target
1321 is running be made an attribute (predicate) to the command rather than
1322 an explicit set of tests.
1326 top.c (execute_command): Should the bulk of this function be moved
1327 into command.[hc] so that top.c doesn't grub around in the command
1332 Architectural Change: Async
1333 ===========================
1335 While GDB uses an event loop when prompting the user for input. That
1336 event loop is not exploited by targets when they allow the target
1337 program to continue. Typically targets still block in (target_wait())
1338 until the program again halts.
1340 The closest a target comes to supporting full asynchronous mode are
1341 the remote targets ``async'' and ``extended-async''.
1345 Asynchronous expression evaluator
1347 Inferior function calls hang GDB.
1351 Fix implementation of ``target xxx''.
1353 At present when the user specifies ``target xxxx'', the CLI maps that
1354 directly onto a target open method. It is then assumed that the
1355 target open method should do all sorts of complicated things as this
1356 is the only chance it has. Check how the various remote targets
1357 duplicate the target operations. Check also how the various targets
1358 behave differently for purely arbitrary reasons.
1360 What should happen is that ``target xxxx'' should call a generic
1361 ``target'' function and that should then co-ordinate the opening of
1362 ``xxxx''. This becomes especially important when you're trying to
1363 open an asynchronous target that may need to perform background tasks
1364 as part of the ``attach'' phase.
1366 Unfortunately, due to limitations in the old/creaking command.h
1367 interface, that isn't possible. The function being called isn't told
1368 of the ``xxx'' or any other context information.
1370 Consequently a precursor to fixing ``target xxxx'' is to clean up the
1371 CLI code so that it passes to the callback function (attatched to a
1372 command) useful information such as the actual command and a context
1373 for that command. Other changes such as making ``struct command''
1374 opaque may also help.
1377 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
1381 Make "target xxx" command interruptible.
1383 As things become async this becomes possible. A target would start
1384 the connect and then return control to the event loop. A cntrl-c
1385 would notify the target that the operation is to be abandoned and the
1386 target code could respond.
1390 Add a "suspend" subcommand of the "continue" command to suspend gdb
1391 while continuing execution of the subprocess. Useful when you are
1392 debugging servers and you want to dodge out and initiate a connection
1393 to a server running under gdb.
1402 Frequently requested but not approved requests.
1406 Eliminate unused argument warnings using ATTRIBUTE_UNUSED.
1408 The benefits on this one are thought to be marginal - GDBs design
1409 means that unused parameters are very common. GCC 3.0 will also
1410 include the option -Wno-unused-parameter which means that ``-Wall
1411 -Wno-unused-parameters -Werror'' can be specified.
1420 This list is not up to date, and opinions vary about the importance or
1421 even desirability of some of the items. If you do fix something, it
1422 always pays to check the below.
1426 @c This does not work (yet if ever). FIXME.
1427 @c @item --parse=@var{lang} @dots{}
1428 @c Configure the @value{GDBN} expression parser to parse the listed languages.
1429 @c @samp{all} configures @value{GDBN} for all supported languages. To get a
1430 @c list of all supported languages, omit the argument. Without this
1431 @c option, @value{GDBN} is configured to parse all supported languages.
1435 START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that
1436 is its default value. Clean this up.
1440 It should be possible to use symbols from shared libraries before we know
1441 exactly where the libraries will be loaded. E.g. "b perror" before running
1442 the program. This could maybe be done as an extension of the "breakpoint
1443 re-evaluation" after new symbols are loaded.
1447 Make single_step() insert and remove breakpoints in one operation.
1449 [If this is talking about having single_step() insert the breakpoints,
1450 run the target then pull the breakpoints then it is wrong. The
1451 function has to return as control has to eventually be passed back to
1452 the main event loop.]
1456 Speed up single stepping by avoiding extraneous ptrace calls.
1460 Speed up single stepping by not inserting and removing breakpoints
1461 each time the inferior starts and stops.
1463 Breakpoints should not be inserted and deleted all the time. Only the
1464 one(s) there should be removed when we have to step over one. Support
1465 breakpoints that don't have to be removed to step over them.
1467 [this has resulted in numerous debates. The issue isn't clear cut]
1471 Provide "voodoo" debugging of core files. This creates a zombie
1472 process as a child of the debugger, and loads it up with the data,
1473 stack, and regs of the core file. This allows you to call functions
1474 in the executable, to manipulate the data in the core file.
1480 GDB reopens the source file on every line, as you "next" through it.
1482 [still true? I've a memory of this being fixed]
1486 Perhaps "i source" should take an argument like that of "list".
1490 Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if
1491 it matches the source line indicated.
1495 The prompt at end of screen should accept space as well as CR.
1499 Backtrace should point out what the currently selected frame is, in
1500 its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar,
1501 ...)" rather than "#3 foo (bar, ...)".
1505 "i program" should work for core files, and display more info, like what
1506 actually caused it to die.
1510 "x/10i" should shorten the long name, if any, on subsequent lines.
1514 "next" over a function that longjumps, never stops until next time you happen
1515 to get to that spot by accident. E.g. "n" over execute_command which has
1520 "set zeroprint off", don't bother printing members of structs which
1521 are entirely zero. Useful for those big structs with few useful
1526 GDB does four ioctl's for every command, probably switching terminal modes
1527 to/from inferior or for readline or something.
1531 terminal_ours versus terminal_inferior: cache state. Switch should be a noop
1532 if the state is the same, too.
1536 "i frame" shows wrong "arglist at" location, doesn't show where the args
1537 should be found, only their actual values.
1541 There should be a way for "set" commands to validate the new setting
1542 before it takes effect.
1546 "ena d" is ambiguous, why? "ena delete" seems to think it is a command!
1550 i line VAR produces "Line number not known for symbol ``var''.". I
1551 thought we were stashing that info now!
1555 We should be able to write to random files at hex offsets like adb.
1559 [elena - delete this]
1561 Handle add_file with separate text, data, and bss addresses. Maybe
1562 handle separate addresses for each segment in the object file?
1566 [Jimb/Elena delete this one]
1568 Handle free_named_symtab to cope with multiply-loaded object files
1569 in a dynamic linking environment. Should remember the last copy loaded,
1570 but not get too snowed if it finds references to the older copy.
1574 [elena delete this also]
1576 Remove all references to:
1583 now that we have BFD. All remaining are in machine dependent files.
1587 Re-organize help categories into things that tend to fit on a screen
1592 Add in commands like ADB's for searching for patterns, etc. We should
1593 be able to examine and patch raw unsymboled binaries as well in gdb as
1594 we can in adb. (E.g. increase the timeout in /bin/login without source).
1596 [actually, add ADB interface :-]
1600 When doing "step" or "next", if a few lines of source are skipped between
1601 the previous line and the current one, print those lines, not just the
1602 last line of a multiline statement.
1606 Handling of "&" address-of operator needs some serious overhaul
1607 for ANSI C and consistency on arrays and functions.
1608 For "float point[15];":
1609 ptype &point[4] ==> Attempt to take address of non-lvalue.
1610 For "char *malloc();":
1611 ptype malloc ==> "char *()"; should be same as
1612 ptype &malloc ==> "char *(*)()"
1613 call printf ("%x\n", malloc) ==> weird value, should be same as
1614 call printf ("%x\n", &malloc) ==> correct value
1618 Fix dbxread.c symbol reading in the presence of interrupts. It
1619 currently leaves a cleanup to blow away the entire symbol table when a
1620 QUIT occurs. (What's wrong with that? -kingdon, 28 Oct 1993).
1622 [I suspect that the grype was that, on a slow system, you might want
1623 to cntrl-c and get just half the symbols and then load the rest later
1624 - scary to be honest]
1628 Mipsread.c reads include files depth-first, because the dependencies
1629 in the psymtabs are way too inclusive (it seems to me). Figure out what
1630 really depends on what, to avoid recursing 20 or 30 times while reading
1635 value_add() should be subtracting the lower bound of arrays, if known,
1636 and possibly checking against the upper bound for error reporting.
1640 When listing source lines, check for a preceding \n, to verify that
1641 the file hasn't changed out from under us.
1643 [fixed by some other means I think. That hack wouldn't actually work
1644 reliably - the file might move such that another \n appears. ]
1648 Get all the remote systems (where the protocol allows it) to be able to
1649 stop the remote system when the GDB user types ^C (like remote.c
1650 does). For ebmon, use ^Ak.
1654 Possible feature: A version of the "disassemble" command which shows
1655 both source and assembly code ("set symbol-filename on" is a partial
1658 [has this been done? It was certainly done for MI and GDBtk]
1662 investigate "x/s 0" (right now stops early) (I think maybe GDB is
1663 using a 0 address for bad purposes internally).
1667 Make "info path" and path_command work again (but independent of the
1668 environment either of gdb or that we'll pass to the inferior).
1672 Make GDB understand the GCC feature for putting octal constants in
1673 enums. Make it so overflow on an enum constant does not error_type
1674 the whole type. Allow arbitrarily large enums with type attributes.
1675 Put all this stuff in the testsuite.
1679 Make TYPE_CODE_ERROR with a non-zero TYPE_LENGTH more useful (print
1680 the value in hex; process type attributes). Add this to the
1681 testsuite. This way future compilers can add new types and old
1682 versions of GDB can do something halfway reasonable.
1686 Fix mdebugread.c:parse_type to do fundamental types right (see
1687 rs6000_builtin_type in stabsread.c for what "right" is--the point is
1688 that the debug format fixes the sizes of these things and it shouldn't
1689 depend on stuff like TARGET_PTR_BIT and so on. For mdebug, there seem
1690 to be separate bt* codes for 64 bit and 32 bit things, and GDB should
1691 be aware of that). Also use a switch statement for clarity and speed.
1695 Investigate adding symbols in target_load--some targets do, some
1700 Put dirname in psymtabs and change lookup*symtab to use dirname (so
1701 /foo/bar.c works whether compiled by cc /foo/bar.c, or cd /foo; cc
1706 Merge xcoffread.c and coffread.c. Use breakpoint_re_set instead of
1711 Make a watchpoint which contains a function call an error (it is
1712 broken now, making it work is probably not worth the effort).
1716 New test case based on weird.exp but in which type numbers are not
1717 renumbered (thus multiply defining a type). This currently causes an
1718 infinite loop on "p v_comb".
1722 [Hey! Hint Hint Delete Delete!!!]
1724 Fix 386 floating point so that floating point registers are real
1725 registers (but code can deal at run-time if they are missing, like
1726 mips and 68k). This would clean up "info float" and related stuff.
1730 gcc -g -c enummask.c then gdb enummask.o, then "p v". GDB complains
1731 about not being able to access memory location 0.
1733 -------------------- enummask.c
1754 If try to modify value in file with "set write off" should give
1755 appropriate error not "cannot access memory at address 0x65e0".
1759 Allow core file without exec file on RS/6000.
1763 Make sure "shell" with no arguments works right on DOS.
1767 Make gdb.ini (as well as .gdbinit) be checked on all platforms, so
1768 the same directory can be NFS-mounted on unix or DOS, and work the
1773 [Is this another delete???]
1775 Get SECT_OFF_TEXT stuff out of objfile_relocate (might be needed to
1776 get RS/6000 to work right, might not be immediately relevant).
1780 Work out some kind of way to allow running the inferior to be done as
1781 a sub-execution of, eg. breakpoint command lists. Currently running
1782 the inferior interupts any command list execution. This would require
1783 some rewriting of wait_for_inferior & friends, and hence should
1784 probably be done in concert with the above.
1788 Add function arguments to gdb user defined functions.
1792 Add convenience variables that refer to exec file, symbol file,
1793 selected frame source file, selected frame function, selected frame
1798 Modify the handling of symbols grouped through BINCL/EINCL stabs to
1799 allocate a partial symtab for each BINCL/EINCL grouping. This will
1800 seriously decrease the size of inter-psymtab dependencies and hence
1801 lessen the amount that needs to be read in when a new source file is
1806 Add a command for searching memory, a la adb. It specifies size,
1807 mask, value, start address. ADB searches until it finds it or hits
1808 an error (or is interrupted).
1812 Remove the range and type checking code and documentation, if not