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 RFD: infrun.c: No bpstat_stop_status call after proceed over break?
16 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00665.html
18 GDB misses watchpoint triggers after proceeding over a breakpoint on
23 x86 linux GDB and SIGALRM (???)
24 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00803.html
26 This problem has been fixed, but a regression test still needs to be
27 added to the testsuite:
28 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00309.html
34 Can't build IRIX -> arm GDB.
35 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00356.html
38 > Now I'm building for an embedded arm target. If there is a way of turning
39 > remote-rdi off, I couldn't find it. It looks like it gets built by default
40 > in gdb/configure.tgt(line 58) Anyway, the build dies in
41 > gdb/rdi-share/unixcomm.c. SERPORT1 et. al. never get defined because we
42 > aren't one of the architectures supported.
46 Problem with weak functions
47 http://sourceware.cygnus.com/ml/gdb/2000-05/msg00060.html
49 Dan Nicolaescu writes:
50 > It seems that gdb-4.95.1 does not display correctly the function when
51 > stoping in weak functions.
53 > It stops in a function that is defined as weak, not in the function
54 > that is actually run...
58 GDB 5.0 doesn't work on Linux/SPARC
62 Thread support. Right now, as soon as a thread finishes and exits,
63 you're hosed. This problem is reported once a week or so.
67 Wow, three bug reports for the same problem in one day! We should
68 probably make fixing this a real priority :-).
70 Anyway, thanks for reporting.
72 The following patch will fix the problems with setting breakpoints in
73 dynamically loaded objects:
75 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00230.html
77 This patch isn't checked in yet (ping Michael/JimB), but I hope this
78 will be in the next GDB release.
80 There should really be a test in the testsuite for this problem, since
81 it keeps coming up :-(. Any volunteers?
88 http://sources.redhat.com/ml/gdb/2000-07/msg00038.html
90 Is the Solaris 8 x86 problem fixed? When you configure it, configure
91 incorrectly determines that I have no curses.h. This causes mucho
92 compilation errors later on.
94 Simply editing the config.h to define CURSES_H fixes the problem, and
95 then the build works fine.
97 The status for this problem:
99 Solaris 8 x86 (PIII-560)
102 I had the same problem with several of the snapshots shortly before
103 5.0 became official, and 5.0 has the same problem.
105 I sent some mail in about it long ago, and never saw a reply.
107 I haven't had time to figure it out myself, especially since I get all
108 confused trying to figure out what configure does, I was happy to find
115 GDB 5.1 - New features
116 ======================
118 The following new features should be included in 5.1.
122 Enable MI by default. Old code can be deleted after 5.1 is out.
126 Pascal (Pierre Muller, David Taylor)
128 Pierre Muller has contributed patches for adding Pascal Language
131 2 pascal language patches inserted in database
132 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00521.html
135 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00496.html
139 Java (Anthony Green, David Taylor)
141 Anthony Green has a number of Java patches that did not make it into
142 the 5.0 release. The first two are in cvs now, but the third needs
143 some fixing up before it can go in.
146 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00512.html
149 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00515.html
151 Patch: handle N_MAIN stab
152 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00527.html
158 Modify gdb to work correctly with Pascal.
162 Revised UDP support (was: Re: [Fwd: [patch] UDP transport support])
163 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00000.html
165 (Broken) support for GDB's remote protocol across UDP is to be
166 included in the follow-on release.
168 It should be noted that UDP can only work when the [Gg] packet fits in
171 There is also much debate over the merit of this.
178 The following code cleanups will hopefully be applied to GDB 5.1.
182 Change documentation to GFDL license.
184 ``It is time to make an effort to start using the GFDL more
185 thoroughly. Would all GNU maintainers please change the license to
186 the GFDL, for all manuals and other major documentation files?
188 The GFDL and some instructions for using it can be found in
189 http://www.gnu.org/copyleft/''
195 Delete macro TARGET_BYTE_ORDER_SELECTABLE.
197 Patches in the database.
201 Fix copyright notices.
203 Turns out that ``1998-2000'' isn't considered valid :-(
205 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00467.html
211 Eliminate all uses of PARAMS in GDB's source code.
215 printcmd.c (print_address_numeric):
217 NOTE: This assumes that the significant address information is kept in
218 the least significant bits of ADDR - the upper bits were either zero
219 or sign extended. Should ADDRESS_TO_POINTER() or some
220 ADDRESS_TO_PRINTABLE() be used to do the conversion?
226 Eliminate all warnings for at least one host/target for the flags:
227 -Wimplicit -Wreturn-type -Wcomment -Wtrigraphs -Wformat -Wparentheses
228 -Wpointer-arith -Wuninitialized
232 Follow through `make check' with --enable-shared.
234 When the srcware tree is configured with --enable-shared, the `expect'
235 program won't run properly. Jim Wilson found out gdb has a local hack
236 to set LD_LIBRARY_PATH, but, AFAIK, no other project has been hacked
239 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00845.html
248 Fix at least one thread bug.
252 GDB 5.2 - New features
253 ======================
257 GCC 3.0 ABI support (but hopefully sooner...).
261 Objective C/C++ support (but hopefully sooner...).
268 The following cleanups have been identified as part of GDB 5.2.
272 Remove old code that does not use ui_out functions and all the related
273 "ifdef"s. This also allows the elimination of -DUI_OUT from
274 Makefile.in and configure.in.
278 Eliminate one more compiler warnings.
282 Restructure gdb directory tree so that it avoids any 8.3 and 14
287 Convert GDB build process to AUTOMAKE.
289 See also sub-directory configure below.
291 The current convention is (kind of) to use $(<header>_h) in all
292 dependency lists. It isn't done in a consistent way.
296 Code Cleanups: General
297 ======================
299 The following are more general cleanups and fixes. They are not tied
300 to any specific release.
304 The BFD directory requires bug-fixed AUTOMAKE et.al.
306 AUTOMAKE 1.4 incorrectly set the TEXINPUTS environment variable. It
307 contained the full path to texinfo.tex when it should have only
308 contained the directory. The bug has been fixed in the current
309 AUTOMAKE sources. Automake snapshots can be found in:
310 ftp://sourceware.cygnus.com/pub/gdb/snapshots
311 and ftp://sourceware.cygnus.com/pub/binutils
315 Find something better than DEFAULT_BFD_ARCH, DEFAULT_BFD_VEC to
316 determine the default isa/byte-order.
320 Rely on BFD_BIG_ENDIAN and BFD_LITTLE_ENDIAN instead of host dependent
321 BIG_ENDIAN and LITTLE_ENDIAN.
325 Eliminate more compiler warnings.
327 Of course there also needs to be the usual debate over which warnings
328 are valid and how to best go about this.
330 One method: choose a single option; get agreement that it is
331 reasonable; try it out to see if there isn't anything silly about it
332 (-Wunused-parameters is an example of that) then incrementally hack
335 The other method is to enable all warnings and eliminate them from one
340 Elimination of ``(catch_errors_ftype *) func''.
342 Like make_cleanup_func it isn't portable.
343 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00791.html
344 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00814.html
348 Nuke #define CONST_PTR.
356 [PATCH/5] src/intl/Makefile.in:distclean additions
357 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00363.html
359 Do not forget to merge the patch back into the trunk.
363 Rationalize the host-endian code (grep for HOST_BYTE_ORDER).
365 At present defs.h includes <endian.h> (which is linux specific) yet
366 almost nothing depends on it. Suggest "gdb_endian.h" which can also
367 handle <machine/endian.h> and only include that where it is really
372 Replace savestring() with something from libiberty.
374 An xstrldup()? but that would have different semantics.
378 Rationalize use of floatformat_unknown in GDB sources.
380 Instead of defaulting to floatformat_unknown, should hosts/targets
381 specify the value explicitly?
383 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
387 Add a ``name'' member to include/floatformat.h:struct floatformat.
388 Print that name in gdbarch.c.
392 Sort out the harris mess in include/floatformat.h (it hardwires two
393 different floating point formats).
397 See of the GDB local floatformat_do_doublest() and libiberty's
398 floatformat_to_double (which was once GDB's ...) can be merged some
403 Eliminate mmalloc(), mstrsave() et.al. from GDB.
405 Also eliminate it from defs.h.
409 Eliminate PTR. ISO-C allows ``void *''.
415 GDB should never abort. GDB should either throw ``error ()'' or
416 ``internal_error ()''. Better still GDB should naturally unwind with
421 GDB probably doesn't build on FreeBSD pre 2.2.x
422 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00378.html
424 Fixes to get FreeBSD working on 2.2.x, 3.x and 4.x caused the code to
429 Deprecate "fg". Apparently ``fg'' is actually continue.
431 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00417.html
435 Deprecate current use of ``floatformat_unknown''.
437 Require all targets to explicitly provide their float format instead
438 of defaulting to floatformat unknown. Doing the latter leads to nasty
441 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
445 Rationalize floatformat_to_double() vs floatformat_to_doublest().
447 Looks like GDB migrated floatformat_to_double() to libiberty but then
448 turned around and created a ..._to_doublest() the latter containing
451 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00472.html
455 Move floatformat_ia64_ext to libiberty/include floatformat.[ch].
457 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00466.html
461 The ``maintenance deprecate set endian big'' command doesn't notice
462 that it is deprecating ``set endian'' and not ``set endian big'' (big
463 is implemented using an enum). Is anyone going to notice this?
467 When tab expanding something like ``set arch<tab>'' ignore the
468 deprecated ``set archdebug'' and expand to ``set architecture''.
472 Eliminate ``arm_register_names[j] = (char *) regnames[j]'' and the
473 like from arm-tdep.c.
477 Fix uses of ->function.cfunc = set_function().
479 The command.c code calls sfunc() when a set command. Rather than
480 change it suggest fixing the callback function so that it is more
483 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
485 See also ``Fix implementation of ``target xxx''.'' below.
489 IRIX 3.x support is probably broken.
493 Delete sim/SIM_HAVE_BREAKPOINTS and gdb/SIM_HAS_BREAKPOINTS.
494 http://sourceware.cygnus.com/ml/gdb-patches/2000-07/msg00042.html
496 Apart from the d30v, are there any sim/common simulators that make use
499 A brief summary of what happened is that sim/common/sim-break.c was
500 created as a good idea. It turned out a better idea was to use
501 SIM_SIGBREAK and have GDB pass back sim_resume (..., SIGBREAK).
505 Move remote_remove_hw_breakpoint, remote_insert_hw_breakpoint,
506 remote_remove_watchpoint, remote_insert_watchpoint into target vector.
510 Eliminate ``extern'' from C files.
514 Replace ``STREQ()'' et.al. with ``strcmp() == 0'' et.al.
516 Extreme care is recommeded - perhaps only modify tests that are
517 exercised by the testsuite (as determined using some type of code
522 Replace the file gdb/CONTRIBUTE with a file that is generated from the
523 gdb/doc/*.texinfo directory.
527 New Features and Fixes
528 ======================
530 These are harder than cleanups but easier than work involving
531 fundamental architectural change.
535 Hardware watchpoint problems on x86 OSes, including Linux:
537 1. Delete/disable hardware watchpoints should free hardware debug
539 2. Watch for different values on a viariable with one hardware debug
542 According to Eli Zaretskii <eliz@delorie.com>:
544 These are not GDB/ia32 issues per se: the above features are all
545 implemented in the DJGPP port of GDB and work in v5.0. Every
546 x86-based target should be able to lift the relevant parts of
547 go32-nat.c and use them almost verbatim. You get debug register
548 sharing through reference counts, and the ability to watch large
549 regions (up to 16 bytes) using multiple registers. (The required
550 infrastructure in high-level GDB application code, mostly in
551 breakpoint.c, is also working since v5.0.)
555 Add built-by, build-date, tm, xm, nm and anything else into gdb binary
556 so that you can see how the GDB was created.
560 Add an "info bfd" command that displays supported object formats,
561 similarly to objdump -i.
563 Is there a command already?
567 Fix ``I'm sorry, Dave, I can't do that.'' from symfile.c.
569 This requires internationalization.
575 (gdb) p fwprintf(stdout,L"%S\n", f)
576 No symbol "L" in current context.
580 Cleanup configury support for optional sub-directories.
582 Check how GCC handles multiple front ends for an example of how things
583 could work. A tentative first step is to rationalize things so that
584 all sub directories are handled in a fashion similar to gdb/mi.
586 See also automake above.
590 Add a transcript mechanism to GDB.
592 Such a mechanism might log all gdb input and output to a file in a
593 form that would allow it to be replayed. It could involve ``gdb
594 --transcript=FILE'' or it could involve ``(gdb) transcript file''.
598 Can the xdep files be replaced by autoconf?
602 Document trace machinery
606 Document ui-out and ui-file.
608 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00121.html
612 Update texinfo.tex to latest?
616 Incorporate agentexpr.texi into gdb.texinfo
618 agentexpr.texi mostly describes the details of the byte code used for
619 tracepoints, not the internals of the support for this in GDB. So it
620 looks like gdb.texinfo is a better place for this information.
622 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00566.html
626 Document overlay machinery.
630 ``(gdb) catch signal SIGNAL''
632 Overlaps with ``handle SIGNAL'' but the implied behavior is different.
633 You can attach commands to a catch but not a handle. A handle has a
634 limited number of hardwired actions.
638 Get the TUI working on all platforms.
642 Add support for ``gdb --- PROGRAM ARGS ...''.
643 Add support for ``gdb -cmd=...''
645 Along with many variations. Check:
647 ????? for a full discussion.
653 Implement ``(gdb) !ls''.
655 Which is very different from ``(gdb) ! ls''. Implementing the latter
658 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00034.html
662 Change the (char *list[]) to (const char (*)[]) so that dynamic lists can
667 When tab expanding something like ``set arch<tab>'' ignore the
668 deprecated ``set archdebug'' and expand to ``set architecture''.
672 Replace the code that uses the host FPU with an emulator of the target
677 The "ocd reset" command needs to flush the dcache, which requires breaking
678 the abstraction layer between the target independent and target code. One
679 way to address this is provide a generic "reset" command and target vector.
681 http://sources.redhat.com/ml/gdb-patches/2000-10/msg00011.html
690 Generic: lin-thread cannot handle thread exit (Mark Kettenis, Michael
691 Snyder) http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00525.html
693 The thread_db assisted debugging code doesn't handle exiting threads
694 properly, at least in combination with glibc 2.1.3 (the framework is
695 there, just not the actual code). There are at least two problems
696 that prevent this from working.
698 As an additional reference point, the pre thread_db code did not work
703 GNU/Linux/x86 and random thread signals (and Solaris/SPARC but not
705 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00336.html
707 Christopher Blizzard writes:
709 So, I've done some more digging into this and it looks like Jim
710 Kingdon has reported this problem in the past:
712 http://sourceware.cygnus.com/ml/bug-gdb/1999-10/msg00058.html
714 I can reproduce this problem both with and without Tom's patch. Has
715 anyone seen this before? Maybe have a solution for it hanging around?
718 There's a test case for this documented at:
720 when debugging threaded applications you get extra SIGTRAPs
721 http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=9565
723 [There should be a GDB testcase - cagney]
727 GDB5 TOT on unixware 7
728 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00119.html
731 > I just spun the top of tree of the GDB5 branch on UnixWare 7. As a
732 > practical matter, the current thread support is somewhat more annoying
733 > than when GDB was thread-unaware.
737 Migrate qfThreadInfo packet -> qThreadInfo. (Andrew Cagney)
739 Add support for packet enable/disable commands with these thread
740 packets. General cleanup.
742 [PATCH] Document the ThreadInfo remote protocol queries
743 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00832.html
745 [PATCH] "info threads" queries for remote.c
746 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00831.html
753 New languages come onto the scene all the time.
757 Re: Various C++ things
759 value_headof/value_from_vtable_info are worthless, and should be
760 removed. The one place in printcmd.c that uses it should use the RTTI
763 RTTI for g++ should be using the typeinfo functions rather than the
764 vtables. The typeinfo functions are always at offset 4 from the
765 beginning of the vtable, and are always right. The vtables will have
766 weird names like E::VB sometimes. The typeinfo function will always
767 be "E type_info function", or somesuch.
769 value_virtual_fn_field needs to be fixed so there are no failures for
770 virtual functions for C++ using g++.
772 Testsuite cases are the major priority right now for C++ support,
773 since i have to make a lot of changes that could potentially break
778 Add support for Modula3
780 Get DEC/Compaq to contribute their Modula-3 support.
784 Remote Protocol Support
785 =======================
789 Remote protocol doco feedback.
791 Too much feedback to mention needs to be merged in (901660). Search
792 for the word ``remote''.
795 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00023.html
796 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00056.html
797 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00382.html
801 GDB doesn't recover gracefully from remote protocol errors.
803 GDB wasn't checking for NAKs from the remote target. Instead a NAK is
804 ignored and a timeout is required before GDB retries. A pre-cursor to
805 fixing this this is making GDB's remote protocol packet more robust.
807 While downloading to a remote protocol target, gdb ignores packet
808 errors in so far as it will continue to download with chunk N+1 even
809 if chunk N was not correctly sent. This causes gdb.base/remote.exp to
810 take a painfully long time to run. As a PS that test needs to be
811 fixed so that it builds on 16 bit machines.
815 Add the cycle step command.
817 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00237.html
821 Resolve how to scale things to support very large packets.
825 Resolve how to handle a target that changes things like its endianess
826 on the fly - should it be returned in the ``T'' packet?
828 Underlying problem is that the register file is target endian. If the
829 target endianess changes gdb doesn't know.
833 Rename read_register{,_pid}() to read_unsigned_register{,_pid}().
840 If / when GDB starts to support the debugging of multi-processor
841 (rather than multi-thread) applications the symtab code will need to
842 be updated a little so that several independent symbol tables are
843 active at a given time.
845 The other interesting change is a clarification of the exact meaning
846 of CORE_ADDR and that has had consequences for a few targets (that
847 were abusing that data type).
851 Investiagate ways of reducing memory.
855 Investigate ways of improving load time.
859 Get the d10v to use POINTER_TO_ADDRESS and ADDRESS_TO_POINTER.
861 Consequence of recent symtab clarification. No marks for figuring out
862 who maintains the d10v.
866 Get the MIPS to correctly sign extend all address <-> pointer
869 Consequence of recent symtab clarification. No marks for figuring out
870 who maintains the MIPS.
874 GDB truncates 64 bit enums.
876 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00290.html
883 There are never to many testcases.
887 Better thread testsuite.
891 Better C++ testsuite.
895 Look at adding a GDB specific testsuite directory so that white box
896 tests of key internals can be added (eg ui_file).
900 Separate out tests that involve the floating point (FP).
902 (Something for people brining up new targets). FP and non-fp tests
903 are combined. I think there should be set of basic tests that
904 exercise pure integer support and then a more expanded set that
905 exercise FP and FP/integer interactions.
907 As an example, the MIPS, for n32 as problems with passing FP's and
908 structs. Since most inferior call tests include FP it is difficult to
909 determine of the integer tests are ok.
913 Architectural Changes: General
914 ==============================
916 These are harder than simple cleanups / fixes and, consequently
917 involve more work. Typically an Architectural Change will be broken
918 down into a more digestible set of cleanups and fixes.
922 Cleanup software single step.
924 At present many targets implement software single step by directly
925 blatting memory (see rs6000-tdep.c). Those targets should register
926 the applicable breakpoints using the breakpoint framework. Perhaphs a
927 new internal breakpoint class ``step'' is needed.
931 Replace READ_FP() with FRAME_HANDLE().
933 READ_FP() is a hangover from the days of the vax when the ABI really
934 did have a frame pointer register. Modern architectures typically
935 construct a virtual frame-handle from the stack pointer and various
936 other bits of string.
938 Unfortunately GDB still treats this synthetic FP register as though it
939 is real. That in turn really confuses users (arm and ``print $fp'' VS
940 ``info registers fp''). The synthetic FP should be separated out of
941 the true register set presented to the user.
945 Register Cache Cleanup (below from Andrew Cagney)
947 I would depict the current register architecture as something like:
954 register + REGISTER_BYTE(reg_nr)
957 -------------------------
958 | extern register[] |
959 -------------------------
961 where neither the high (valops.c et.al.) or low gdb (*-tdep.c) are
962 really clear on what mechanisms they should be using to manipulate that
963 buffer. Further, much code assumes, dangerously, that registers are
964 contigious. Having got mips-tdep.c to support multiple ABIs, believe
965 me, that is a bad assumption. Finally, that register cache layout is
966 determined by the current remote/local target and _not_ the less
967 specific target ISA. In fact, in many cases it is determined by the
968 somewhat arbitrary layout of the [gG] packets!
971 How I would like the register file to work is more like:
999 The main objectives being:
1001 o a clear separation between the low
1002 level target and the high level GDB
1004 o a mechanism that solves the general
1005 problem of register aliases, overlaps
1006 etc instead of treating them as optional
1007 extras that can be wedged in as an after
1008 thought (that is a reasonable description
1009 of the current code).
1011 Identify then solve the hard case and the
1012 rest just falls out. GDB solved the easy
1013 case and then tried to ignore the real
1016 o a removal of the assumption that the
1017 mapping between the register cache
1018 and virtual registers is largely static.
1019 If you flip the USR/SSR stack register
1020 select bit in the status-register then
1021 the corresponding stack registers should
1024 o a mechanism that clearly separates the
1025 gdb internal register cache from any
1026 target (not architecture) dependent
1027 specifics such as [gG] packets.
1029 Of course, like anything, it sounds good in theory. In reality, it
1030 would have to contend with many<->many relationships at both the
1031 virt<->cache and cache<->target level. For instance:
1034 Modifying an mmx register may involve
1035 scattering values across both FP and
1036 mmpx specific parts of a buffer
1039 When writing back a SP it may need to
1040 both be written to both SP and USP.
1045 Rather than let this like the last time it was discussed, just slip, I'm
1046 first going to add this e-mail (+ references) to TODO. I'd then like to
1047 sketch out a broad strategy I think could get us there.
1050 First thing I'd suggest is separating out the ``extern registers[]''
1051 code so that we can at least identify what is using it. At present
1052 things are scattered across many files. That way we can at least
1053 pretend that there is a cache instead of a global array :-)
1055 I'd then suggest someone putting up a proposal for the pseudo-reg /
1056 high-level side interface so that code can be adopted to it. For old
1057 code, initially a blanket rename of write_register_bytes() to
1058 deprecated_write_register_bytes() would help.
1060 Following that would, finaly be the corresponding changes to the target.
1064 Check that GDB can handle all BFD architectures (Andrew Cagney)
1066 There should be a test that checks that BFD/GDB are in sync with
1067 regard to architecture changes. Something like a test that first
1068 queries GDB for all supported architectures and then feeds each back
1069 to GDB.. Anyone interested in learning how to write tests? :-)
1073 Architectural Change: Multi-arch et al.
1074 =======================================
1076 The long term objective is to remove all assumptions that there is a
1077 single target with a single address space with a single instruction
1078 set architecture and single application binary interface.
1080 This is an ongoing effort. The first milestone is to enable
1081 ``multi-arch'' where by all architectural decisions are made at
1084 It should be noted that ``gdbarch'' is really ``gdbabi'' and
1085 ``gdbisa''. Once things are multi-arched breaking that down correctly
1086 will become much easier.
1090 GDBARCH cleanup (Andrew Cagney)
1092 The non-generated parts of gdbarch.{sh,h,c} should be separated out
1093 into arch-utils.[hc].
1095 Document that gdbarch_init_ftype could easily fail because it didn't
1096 identify an architecture.
1100 Fix BELIEVE_PPC_PROMOTION. Change it to BELIEVE_PPC_PROMOTION_P?
1102 At present there is still #ifdef BELIEVE_PPC_PROMOTION code in the
1107 Fix target_signal_from_host() etc.
1109 The name is wrong for starters. ``target_signal'' should probably be
1110 ``gdb_signal''. ``from_host'' should be ``from_target_signal''.
1111 After that it needs to be multi-arched and made independent of any
1112 host signal numbering.
1116 Update ALPHA so that it uses ``struct frame_extra_info'' instead of
1119 This is a barrier to replacing mips_extra_func_info with something
1120 that works with multi-arch.
1124 Multi-arch mips_extra_func_info.
1126 This first needs the alpha to be updated so that it uses ``struct
1131 Rationalize TARGET_SINGLE_FORMAT and TARGET_SINGLE_BIT et al.
1133 Surely one of them is redundant.
1137 Convert ALL architectures to MULTI-ARCH.
1141 Select the initial multi-arch ISA / ABI based on --target or similar.
1143 At present the default is based on what ever is first in the BFD
1144 archures table. It should be determined based on the ``--target=...''
1149 Make MIPS pure multi-arch.
1151 It is only at the multi-arch enabled stage.
1157 Enable the code to recognize --enable-targets=.... like BINUTILS does.
1159 Can the tm.h and nm.h files be eliminated by multi-arch.
1163 Architectural Change: MI, LIBGDB and scripting languages
1164 ========================================================
1166 See also architectural changes related to the event loop. LIBGDB
1167 can't be finished until there is a generic event loop being used by
1170 The long term objective is it to be possible to integrate GDB into
1171 scripting languages.
1175 Implement generic ``(gdb) commmand > file''
1177 Once everything is going through ui_file it should be come fairly
1180 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00104.html
1184 Replace gdb_stdtarg with gdb_targout (and possibly gdb_targerr).
1186 gdb_stdtarg is easily confused with gdb_stdarg.
1190 Extra ui_file methods - dump.
1192 Very useful for whitebox testing.
1196 Eliminate error_begin().
1198 With ui_file, there is no need for the statefull error_begin ()
1203 Send normal output to gdb_stdout.
1204 Send error messages to gdb_stderror.
1205 Send debug and log output log gdb_stdlog.
1207 GDB still contains many cases where (f)printf or printf_filtered () is
1208 used when it should be sending the messages to gdb_stderror or
1209 gdb_stdlog. The thought of #defining printf to something has crossed
1214 Re-do GDB's output pager.
1216 GDB's output pager still relies on people correctly using *_filtered
1217 for gdb_stdout and *_unfiltered for gdb_stdlog / gdb_stderr.
1218 Hopefully, with all normal output going to gdb_stdout, the pager can
1219 just look at the ui_file that the output is on and then use that to
1220 decide what to do about paging. Sounds good in theory.
1224 Check/cleanup MI documentation.
1226 The list of commands specified in the documentation needs to be
1227 checked against the mi-cmds.c table in a mechanical way (so that they
1228 two can be kept up-to-date).
1232 Convert MI into libgdb
1234 MI provides a text interface into what should be many of the libgdb
1235 functions. The implementation of those functions should be separated
1236 into the MI interface and the functions proper. Those functions being
1237 moved to gdb/lib say.
1243 The first part can already be found in defs.h.
1247 MI's input does not use buffering.
1249 At present the MI interface reads raw characters of from an unbuffered
1250 FD. This is to avoid several nasty buffer/race conditions. That code
1251 should be changed so that it registers its self with the event loop
1252 (on the input FD) and then push commands up to MI as they arrive.
1254 The serial code already does this.
1258 Make MI interface accessible from existing CLI.
1262 Add a breakpoint-edit command to MI.
1264 It would be similar to MI's breakpoint create but would apply to an
1265 existing breakpoint. It saves the need to delete/create breakpoints
1266 when ever they are changed.
1270 Add directory path to MI breakpoint.
1272 That way the GUI's task of finding the file within which the
1273 breakpoint was set is simplified.
1277 Add a mechanism to reject certain expression classes to MI
1279 There are situtations where you don't want GDB's expression
1280 parser/evaluator to perform inferior function calls or variable
1281 assignments. A way of restricting the expression parser so that such
1282 operations are not accepted would be very helpful.
1286 Remove sideffects from libgdb breakpoint create function.
1288 The user can use the CLI to create a breakpoint with partial
1289 information - no file (gdb would use the file from the last
1292 The libgdb interface currently affects that environment which can lead
1293 to confusion when a user is setting breakpoints via both the MI and
1296 This is also a good example of how getting the CLI ``right'' will be
1301 Move gdb_lasterr to ui_out?
1303 The way GDB throws errors and records them needs a re-think. ui_out
1304 handles the correct output well. It doesn't resolve what to do with
1305 output / error-messages when things go wrong.
1309 do_setshow_command contains a 1024 byte buffer.
1311 The function assumes that there will never be any more than 1024 bytes
1312 of enum. It should use mem_file.
1316 Should struct cmd_list_element . completer take the command as an
1321 Should the bulk of top.c:line_completion_function() be moved to
1322 command.[hc]? complete_on_cmdlist() and complete_on_enums() could
1323 then be made private.
1327 top.c (execute_command): Should a command being valid when the target
1328 is running be made an attribute (predicate) to the command rather than
1329 an explicit set of tests.
1333 top.c (execute_command): Should the bulk of this function be moved
1334 into command.[hc] so that top.c doesn't grub around in the command
1339 Architectural Change: Async
1340 ===========================
1342 While GDB uses an event loop when prompting the user for input. That
1343 event loop is not exploited by targets when they allow the target
1344 program to continue. Typically targets still block in (target_wait())
1345 until the program again halts.
1347 The closest a target comes to supporting full asynchronous mode are
1348 the remote targets ``async'' and ``extended-async''.
1352 Asynchronous expression evaluator
1354 Inferior function calls hang GDB.
1358 Fix implementation of ``target xxx''.
1360 At present when the user specifies ``target xxxx'', the CLI maps that
1361 directly onto a target open method. It is then assumed that the
1362 target open method should do all sorts of complicated things as this
1363 is the only chance it has. Check how the various remote targets
1364 duplicate the target operations. Check also how the various targets
1365 behave differently for purely arbitrary reasons.
1367 What should happen is that ``target xxxx'' should call a generic
1368 ``target'' function and that should then co-ordinate the opening of
1369 ``xxxx''. This becomes especially important when you're trying to
1370 open an asynchronous target that may need to perform background tasks
1371 as part of the ``attach'' phase.
1373 Unfortunately, due to limitations in the old/creaking command.h
1374 interface, that isn't possible. The function being called isn't told
1375 of the ``xxx'' or any other context information.
1377 Consequently a precursor to fixing ``target xxxx'' is to clean up the
1378 CLI code so that it passes to the callback function (attatched to a
1379 command) useful information such as the actual command and a context
1380 for that command. Other changes such as making ``struct command''
1381 opaque may also help.
1384 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
1388 Make "target xxx" command interruptible.
1390 As things become async this becomes possible. A target would start
1391 the connect and then return control to the event loop. A cntrl-c
1392 would notify the target that the operation is to be abandoned and the
1393 target code could respond.
1397 Add a "suspend" subcommand of the "continue" command to suspend gdb
1398 while continuing execution of the subprocess. Useful when you are
1399 debugging servers and you want to dodge out and initiate a connection
1400 to a server running under gdb.
1409 Frequently requested but not approved requests.
1413 Eliminate unused argument warnings using ATTRIBUTE_UNUSED.
1415 The benefits on this one are thought to be marginal - GDBs design
1416 means that unused parameters are very common. GCC 3.0 will also
1417 include the option -Wno-unused-parameter which means that ``-Wall
1418 -Wno-unused-parameters -Werror'' can be specified.
1427 This list is not up to date, and opinions vary about the importance or
1428 even desirability of some of the items. If you do fix something, it
1429 always pays to check the below.
1433 @c This does not work (yet if ever). FIXME.
1434 @c @item --parse=@var{lang} @dots{}
1435 @c Configure the @value{GDBN} expression parser to parse the listed languages.
1436 @c @samp{all} configures @value{GDBN} for all supported languages. To get a
1437 @c list of all supported languages, omit the argument. Without this
1438 @c option, @value{GDBN} is configured to parse all supported languages.
1442 START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that
1443 is its default value. Clean this up.
1447 It should be possible to use symbols from shared libraries before we know
1448 exactly where the libraries will be loaded. E.g. "b perror" before running
1449 the program. This could maybe be done as an extension of the "breakpoint
1450 re-evaluation" after new symbols are loaded.
1454 Make single_step() insert and remove breakpoints in one operation.
1456 [If this is talking about having single_step() insert the breakpoints,
1457 run the target then pull the breakpoints then it is wrong. The
1458 function has to return as control has to eventually be passed back to
1459 the main event loop.]
1463 Speed up single stepping by avoiding extraneous ptrace calls.
1467 Speed up single stepping by not inserting and removing breakpoints
1468 each time the inferior starts and stops.
1470 Breakpoints should not be inserted and deleted all the time. Only the
1471 one(s) there should be removed when we have to step over one. Support
1472 breakpoints that don't have to be removed to step over them.
1474 [this has resulted in numerous debates. The issue isn't clear cut]
1478 Provide "voodoo" debugging of core files. This creates a zombie
1479 process as a child of the debugger, and loads it up with the data,
1480 stack, and regs of the core file. This allows you to call functions
1481 in the executable, to manipulate the data in the core file.
1487 GDB reopens the source file on every line, as you "next" through it.
1489 [still true? I've a memory of this being fixed]
1493 Perhaps "i source" should take an argument like that of "list".
1497 Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if
1498 it matches the source line indicated.
1502 The prompt at end of screen should accept space as well as CR.
1506 Backtrace should point out what the currently selected frame is, in
1507 its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar,
1508 ...)" rather than "#3 foo (bar, ...)".
1512 "i program" should work for core files, and display more info, like what
1513 actually caused it to die.
1517 "x/10i" should shorten the long name, if any, on subsequent lines.
1521 "next" over a function that longjumps, never stops until next time you happen
1522 to get to that spot by accident. E.g. "n" over execute_command which has
1527 "set zeroprint off", don't bother printing members of structs which
1528 are entirely zero. Useful for those big structs with few useful
1533 GDB does four ioctl's for every command, probably switching terminal modes
1534 to/from inferior or for readline or something.
1538 terminal_ours versus terminal_inferior: cache state. Switch should be a noop
1539 if the state is the same, too.
1543 "i frame" shows wrong "arglist at" location, doesn't show where the args
1544 should be found, only their actual values.
1548 There should be a way for "set" commands to validate the new setting
1549 before it takes effect.
1553 "ena d" is ambiguous, why? "ena delete" seems to think it is a command!
1557 i line VAR produces "Line number not known for symbol ``var''.". I
1558 thought we were stashing that info now!
1562 We should be able to write to random files at hex offsets like adb.
1566 [elena - delete this]
1568 Handle add_file with separate text, data, and bss addresses. Maybe
1569 handle separate addresses for each segment in the object file?
1573 [Jimb/Elena delete this one]
1575 Handle free_named_symtab to cope with multiply-loaded object files
1576 in a dynamic linking environment. Should remember the last copy loaded,
1577 but not get too snowed if it finds references to the older copy.
1581 [elena delete this also]
1583 Remove all references to:
1590 now that we have BFD. All remaining are in machine dependent files.
1594 Re-organize help categories into things that tend to fit on a screen
1599 Add in commands like ADB's for searching for patterns, etc. We should
1600 be able to examine and patch raw unsymboled binaries as well in gdb as
1601 we can in adb. (E.g. increase the timeout in /bin/login without source).
1603 [actually, add ADB interface :-]
1607 When doing "step" or "next", if a few lines of source are skipped between
1608 the previous line and the current one, print those lines, not just the
1609 last line of a multiline statement.
1613 Handling of "&" address-of operator needs some serious overhaul
1614 for ANSI C and consistency on arrays and functions.
1615 For "float point[15];":
1616 ptype &point[4] ==> Attempt to take address of non-lvalue.
1617 For "char *malloc();":
1618 ptype malloc ==> "char *()"; should be same as
1619 ptype &malloc ==> "char *(*)()"
1620 call printf ("%x\n", malloc) ==> weird value, should be same as
1621 call printf ("%x\n", &malloc) ==> correct value
1625 Fix dbxread.c symbol reading in the presence of interrupts. It
1626 currently leaves a cleanup to blow away the entire symbol table when a
1627 QUIT occurs. (What's wrong with that? -kingdon, 28 Oct 1993).
1629 [I suspect that the grype was that, on a slow system, you might want
1630 to cntrl-c and get just half the symbols and then load the rest later
1631 - scary to be honest]
1635 Mipsread.c reads include files depth-first, because the dependencies
1636 in the psymtabs are way too inclusive (it seems to me). Figure out what
1637 really depends on what, to avoid recursing 20 or 30 times while reading
1642 value_add() should be subtracting the lower bound of arrays, if known,
1643 and possibly checking against the upper bound for error reporting.
1647 When listing source lines, check for a preceding \n, to verify that
1648 the file hasn't changed out from under us.
1650 [fixed by some other means I think. That hack wouldn't actually work
1651 reliably - the file might move such that another \n appears. ]
1655 Get all the remote systems (where the protocol allows it) to be able to
1656 stop the remote system when the GDB user types ^C (like remote.c
1657 does). For ebmon, use ^Ak.
1661 Possible feature: A version of the "disassemble" command which shows
1662 both source and assembly code ("set symbol-filename on" is a partial
1665 [has this been done? It was certainly done for MI and GDBtk]
1669 investigate "x/s 0" (right now stops early) (I think maybe GDB is
1670 using a 0 address for bad purposes internally).
1674 Make "info path" and path_command work again (but independent of the
1675 environment either of gdb or that we'll pass to the inferior).
1679 Make GDB understand the GCC feature for putting octal constants in
1680 enums. Make it so overflow on an enum constant does not error_type
1681 the whole type. Allow arbitrarily large enums with type attributes.
1682 Put all this stuff in the testsuite.
1686 Make TYPE_CODE_ERROR with a non-zero TYPE_LENGTH more useful (print
1687 the value in hex; process type attributes). Add this to the
1688 testsuite. This way future compilers can add new types and old
1689 versions of GDB can do something halfway reasonable.
1693 Fix mdebugread.c:parse_type to do fundamental types right (see
1694 rs6000_builtin_type in stabsread.c for what "right" is--the point is
1695 that the debug format fixes the sizes of these things and it shouldn't
1696 depend on stuff like TARGET_PTR_BIT and so on. For mdebug, there seem
1697 to be separate bt* codes for 64 bit and 32 bit things, and GDB should
1698 be aware of that). Also use a switch statement for clarity and speed.
1702 Investigate adding symbols in target_load--some targets do, some
1707 Put dirname in psymtabs and change lookup*symtab to use dirname (so
1708 /foo/bar.c works whether compiled by cc /foo/bar.c, or cd /foo; cc
1713 Merge xcoffread.c and coffread.c. Use breakpoint_re_set instead of
1718 Make a watchpoint which contains a function call an error (it is
1719 broken now, making it work is probably not worth the effort).
1723 New test case based on weird.exp but in which type numbers are not
1724 renumbered (thus multiply defining a type). This currently causes an
1725 infinite loop on "p v_comb".
1729 [Hey! Hint Hint Delete Delete!!!]
1731 Fix 386 floating point so that floating point registers are real
1732 registers (but code can deal at run-time if they are missing, like
1733 mips and 68k). This would clean up "info float" and related stuff.
1737 gcc -g -c enummask.c then gdb enummask.o, then "p v". GDB complains
1738 about not being able to access memory location 0.
1740 -------------------- enummask.c
1761 If try to modify value in file with "set write off" should give
1762 appropriate error not "cannot access memory at address 0x65e0".
1766 Allow core file without exec file on RS/6000.
1770 Make sure "shell" with no arguments works right on DOS.
1774 Make gdb.ini (as well as .gdbinit) be checked on all platforms, so
1775 the same directory can be NFS-mounted on unix or DOS, and work the
1780 [Is this another delete???]
1782 Get SECT_OFF_TEXT stuff out of objfile_relocate (might be needed to
1783 get RS/6000 to work right, might not be immediately relevant).
1787 Work out some kind of way to allow running the inferior to be done as
1788 a sub-execution of, eg. breakpoint command lists. Currently running
1789 the inferior interupts any command list execution. This would require
1790 some rewriting of wait_for_inferior & friends, and hence should
1791 probably be done in concert with the above.
1795 Add function arguments to gdb user defined functions.
1799 Add convenience variables that refer to exec file, symbol file,
1800 selected frame source file, selected frame function, selected frame
1805 Modify the handling of symbols grouped through BINCL/EINCL stabs to
1806 allocate a partial symtab for each BINCL/EINCL grouping. This will
1807 seriously decrease the size of inter-psymtab dependencies and hence
1808 lessen the amount that needs to be read in when a new source file is
1813 Add a command for searching memory, a la adb. It specifies size,
1814 mask, value, start address. ADB searches until it finds it or hits
1815 an error (or is interrupted).
1819 Remove the range and type checking code and documentation, if not