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 Remove old code that does not use ui_out functions and all the related
293 Eliminate more compiler warnings.
297 Restructure gdb directory tree so that it avoids any 8.3 and 14
302 Convert GDB build process to AUTOMAKE.
304 See also sub-directory configure below.
306 The current convention is (kind of) to use $(<header>_h) in all
307 dependency lists. It isn't done in a consistent way.
311 Code Cleanups: General
312 ======================
314 The following are more general cleanups and fixes. They are not tied
315 to any specific release.
319 The BFD directory requires bug-fixed AUTOMAKE et.al.
321 AUTOMAKE 1.4 incorrectly set the TEXINPUTS environment variable. It
322 contained the full path to texinfo.tex when it should have only
323 contained the directory. The bug has been fixed in the current
324 AUTOMAKE sources. Automake snapshots can be found in:
325 ftp://sourceware.cygnus.com/pub/gdb/snapshots
326 and ftp://sourceware.cygnus.com/pub/binutils
330 Find something better than DEFAULT_BFD_ARCH, DEFAULT_BFD_VEC to
331 determine the default isa/byte-order.
335 Rely on BFD_BIG_ENDIAN and BFD_LITTLE_ENDIAN instead of host dependent
336 BIG_ENDIAN and LITTLE_ENDIAN.
340 Eliminate more compiler warnings.
342 Of course there also needs to be the usual debate over which warnings
343 are valid and how to best go about this.
345 One method: choose a single option; get agreement that it is
346 reasonable; try it out to see if there isn't anything silly about it
347 (-Wunused-parameters is an example of that) then incrementally hack
350 The other method is to enable all warnings and eliminate them from one
355 Elimination of ``(catch_errors_ftype *) func''.
357 Like make_cleanup_func it isn't portable.
358 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00791.html
359 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00814.html
363 Nuke #define CONST_PTR.
371 [PATCH/5] src/intl/Makefile.in:distclean additions
372 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00363.html
374 Do not forget to merge the patch back into the trunk.
378 Rationalize the host-endian code (grep for HOST_BYTE_ORDER).
380 At present defs.h includes <endian.h> (which is linux specific) yet
381 almost nothing depends on it. Suggest "gdb_endian.h" which can also
382 handle <machine/endian.h> and only include that where it is really
387 Replace savestring() with something from libiberty.
389 An xstrldup()? but that would have different semantics.
393 Rationalize use of floatformat_unknown in GDB sources.
395 Instead of defaulting to floatformat_unknown, should hosts/targets
396 specify the value explicitly?
398 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
402 Add a ``name'' member to include/floatformat.h:struct floatformat.
403 Print that name in gdbarch.c.
407 Sort out the harris mess in include/floatformat.h (it hardwires two
408 different floating point formats).
412 See of the GDB local floatformat_do_doublest() and libiberty's
413 floatformat_to_double (which was once GDB's ...) can be merged some
418 Eliminate mmalloc(), mstrsave() et.al. from GDB.
420 Also eliminate it from defs.h.
424 Eliminate PTR. ISO-C allows ``void *''.
430 GDB should never abort. GDB should either throw ``error ()'' or
431 ``internal_error ()''. Better still GDB should naturally unwind with
436 Add __LINE__ and __FILE__ to internal_error().
440 GDB probably doesn't build on FreeBSD pre 2.2.x
441 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00378.html
443 Fixes to get FreeBSD working on 2.2.x, 3.x and 4.x caused the code to
448 Deprecate "fg". Apparently ``fg'' is actually continue.
450 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00417.html
454 Deprecate current use of ``floatformat_unknown''.
456 Require all targets to explicitly provide their float format instead
457 of defaulting to floatformat unknown. Doing the latter leads to nasty
460 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
464 Rationalize floatformat_to_double() vs floatformat_to_doublest().
466 Looks like GDB migrated floatformat_to_double() to libiberty but then
467 turned around and created a ..._to_doublest() the latter containing
470 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00472.html
474 Move floatformat_ia64_ext to libiberty/include floatformat.[ch].
476 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00466.html
480 The ``maintenance deprecate set endian big'' command doesn't notice
481 that it is deprecating ``set endian'' and not ``set endian big'' (big
482 is implemented using an enum). Is anyone going to notice this?
486 When tab expanding something like ``set arch<tab>'' ignore the
487 deprecated ``set archdebug'' and expand to ``set architecture''.
491 Eliminate ``arm_register_names[j] = (char *) regnames[j]'' and the
492 like from arm-tdep.c.
496 Fix uses of ->function.cfunc = set_function().
498 The command.c code calls sfunc() when a set command. Rather than
499 change it suggest fixing the callback function so that it is more
502 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
504 See also ``Fix implementation of ``target xxx''.'' below.
508 IRIX 3.x support is probably broken.
512 Delete sim/SIM_HAVE_BREAKPOINTS and gdb/SIM_HAS_BREAKPOINTS.
513 http://sourceware.cygnus.com/ml/gdb-patches/2000-07/msg00042.html
515 Apart from the d30v, are there any sim/common simulators that make use
518 A brief summary of what happened is that sim/common/sim-break.c was
519 created as a good idea. It turned out a better idea was to use
520 SIM_SIGBREAK and have GDB pass back sim_resume (..., SIGBREAK).
524 Move remote_remove_hw_breakpoint, remote_insert_hw_breakpoint,
525 remote_remove_watchpoint, remote_insert_watchpoint into target vector.
529 Eliminate ``extern'' from C files.
533 Replace ``STREQ()'' et.al. with ``strcmp() == 0'' et.al.
535 Extreme care is recommeded - perhaps only modify tests that are
536 exercised by the testsuite (as determined using some type of code
541 New Features and Fixes
542 ======================
544 These are harder than cleanups but easier than work involving
545 fundamental architectural change.
549 Add built-by, build-date, tm, xm, nm and anything else into gdb binary
550 so that you can see how the GDB was created.
554 Add an "info bfd" command that displays supported object formats,
555 similarly to objdump -i.
557 Is there a command already?
561 Fix ``I'm sorry, Dave, I can't do that.'' from symfile.c.
563 This requires internationalization.
569 (gdb) p fwprintf(stdout,L"%S\n", f)
570 No symbol "L" in current context.
574 Cleanup configury support for optional sub-directories.
576 Check how GCC handles multiple front ends for an example of how things
577 could work. A tentative first step is to rationalize things so that
578 all sub directories are handled in a fashion similar to gdb/mi.
580 See also automake above.
584 Add a transcript mechanism to GDB.
586 Such a mechanism might log all gdb input and output to a file in a
587 form that would allow it to be replayed. It could involve ``gdb
588 --transcript=FILE'' or it could involve ``(gdb) transcript file''.
592 Can the xdep files be replaced by autoconf?
596 Document trace machinery
600 Document ui-out and ui-file.
602 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00121.html
606 Update texinfo.tex to latest?
610 Incorporate agentexpr.texi into gdb.texinfo
612 agentexpr.texi mostly describes the details of the byte code used for
613 tracepoints, not the internals of the support for this in GDB. So it
614 looks like gdb.texinfo is a better place for this information.
616 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00566.html
620 Document overlay machinery.
624 ``(gdb) catch signal SIGNAL''
626 Overlaps with ``handle SIGNAL'' but the implied behavior is different.
627 You can attach commands to a catch but not a handle. A handle has a
628 limited number of hardwired actions.
632 Get the TUI working on all platforms.
636 Add support for ``gdb --- PROGRAM ARGS ...''.
637 Add support for ``gdb -cmd=...''
639 Along with many variations. Check:
641 ????? for a full discussion.
647 Implement ``(gdb) !ls''.
649 Which is very different from ``(gdb) ! ls''. Implementing the latter
652 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00034.html
656 Change the (char *list[]) to (const char (*)[]) so that dynamic lists can
661 When tab expanding something like ``set arch<tab>'' ignore the
662 deprecated ``set archdebug'' and expand to ``set architecture''.
666 Replace the code that uses the host FPU with an emulator of the target
671 The "ocd reset" command needs to flush the dcache, which requires breaking
672 the abstraction layer between the target independent and target code. One
673 way to address this is provide a generic "reset" command and target vector.
675 http://sources.redhat.com/ml/gdb-patches/2000-10/msg00011.html
684 Generic: lin-thread cannot handle thread exit (Mark Kettenis, Michael
685 Snyder) http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00525.html
687 The thread_db assisted debugging code doesn't handle exiting threads
688 properly, at least in combination with glibc 2.1.3 (the framework is
689 there, just not the actual code). There are at least two problems
690 that prevent this from working.
692 As an additional reference point, the pre thread_db code did not work
697 GNU/Linux/x86 and random thread signals (and Solaris/SPARC but not
699 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00336.html
701 Christopher Blizzard writes:
703 So, I've done some more digging into this and it looks like Jim
704 Kingdon has reported this problem in the past:
706 http://sourceware.cygnus.com/ml/bug-gdb/1999-10/msg00058.html
708 I can reproduce this problem both with and without Tom's patch. Has
709 anyone seen this before? Maybe have a solution for it hanging around?
712 There's a test case for this documented at:
714 when debugging threaded applications you get extra SIGTRAPs
715 http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=9565
717 [There should be a GDB testcase - cagney]
721 GDB5 TOT on unixware 7
722 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00119.html
725 > I just spun the top of tree of the GDB5 branch on UnixWare 7. As a
726 > practical matter, the current thread support is somewhat more annoying
727 > than when GDB was thread-unaware.
731 Migrate qfThreadInfo packet -> qThreadInfo. (Andrew Cagney)
733 Add support for packet enable/disable commands with these thread
734 packets. General cleanup.
736 [PATCH] Document the ThreadInfo remote protocol queries
737 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00832.html
739 [PATCH] "info threads" queries for remote.c
740 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00831.html
747 New languages come onto the scene all the time.
751 Re: Various C++ things
753 value_headof/value_from_vtable_info are worthless, and should be
754 removed. The one place in printcmd.c that uses it should use the RTTI
757 RTTI for g++ should be using the typeinfo functions rather than the
758 vtables. The typeinfo functions are always at offset 4 from the
759 beginning of the vtable, and are always right. The vtables will have
760 weird names like E::VB sometimes. The typeinfo function will always
761 be "E type_info function", or somesuch.
763 value_virtual_fn_field needs to be fixed so there are no failures for
764 virtual functions for C++ using g++.
766 Testsuite cases are the major priority right now for C++ support,
767 since i have to make a lot of changes that could potentially break
772 Add support for Modula3
774 Get DEC/Compaq to contribute their Modula-3 support.
778 Remote Protocol Support
779 =======================
783 Remote protocol doco feedback.
785 Too much feedback to mention needs to be merged in (901660). Search
786 for the word ``remote''.
789 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00023.html
790 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00056.html
791 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00382.html
795 GDB doesn't recover gracefully from remote protocol errors.
797 GDB wasn't checking for NAKs from the remote target. Instead a NAK is
798 ignored and a timeout is required before GDB retries. A pre-cursor to
799 fixing this this is making GDB's remote protocol packet more robust.
801 While downloading to a remote protocol target, gdb ignores packet
802 errors in so far as it will continue to download with chunk N+1 even
803 if chunk N was not correctly sent. This causes gdb.base/remote.exp to
804 take a painfully long time to run. As a PS that test needs to be
805 fixed so that it builds on 16 bit machines.
809 Add the cycle step command.
811 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00237.html
815 Resolve how to scale things to support very large packets.
819 Resolve how to handle a target that changes things like its endianess
820 on the fly - should it be returned in the ``T'' packet?
822 Underlying problem is that the register file is target endian. If the
823 target endianess changes gdb doesn't know.
827 Rename read_register{,_pid}() to read_unsigned_register{,_pid}().
834 If / when GDB starts to support the debugging of multi-processor
835 (rather than multi-thread) applications the symtab code will need to
836 be updated a little so that several independent symbol tables are
837 active at a given time.
839 The other interesting change is a clarification of the exact meaning
840 of CORE_ADDR and that has had consequences for a few targets (that
841 were abusing that data type).
845 Investiagate ways of reducing memory.
849 Investigate ways of improving load time.
853 Get the d10v to use POINTER_TO_ADDRESS and ADDRESS_TO_POINTER.
855 Consequence of recent symtab clarification. No marks for figuring out
856 who maintains the d10v.
860 Get the MIPS to correctly sign extend all address <-> pointer
863 Consequence of recent symtab clarification. No marks for figuring out
864 who maintains the MIPS.
868 GDB truncates 64 bit enums.
870 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00290.html
877 There are never to many testcases.
881 Better thread testsuite.
885 Better C++ testsuite.
889 Look at adding a GDB specific testsuite directory so that white box
890 tests of key internals can be added (eg ui_file).
894 Separate out tests that involve the floating point (FP).
896 (Something for people brining up new targets). FP and non-fp tests
897 are combined. I think there should be set of basic tests that
898 exercise pure integer support and then a more expanded set that
899 exercise FP and FP/integer interactions.
901 As an example, the MIPS, for n32 as problems with passing FP's and
902 structs. Since most inferior call tests include FP it is difficult to
903 determine of the integer tests are ok.
907 Architectural Changes: General
908 ==============================
910 These are harder than simple cleanups / fixes and, consequently
911 involve more work. Typically an Architectural Change will be broken
912 down into a more digestible set of cleanups and fixes.
916 Cleanup software single step.
918 At present many targets implement software single step by directly
919 blatting memory (see rs6000-tdep.c). Those targets should register
920 the applicable breakpoints using the breakpoint framework. Perhaphs a
921 new internal breakpoint class ``step'' is needed.
925 Replace READ_FP() with FRAME_HANDLE().
927 READ_FP() is a hangover from the days of the vax when the ABI really
928 did have a frame pointer register. Modern architectures typically
929 construct a virtual frame-handle from the stack pointer and various
930 other bits of string.
932 Unfortunately GDB still treats this synthetic FP register as though it
933 is real. That in turn really confuses users (arm and ``print $fp'' VS
934 ``info registers fp''). The synthetic FP should be separated out of
935 the true register set presented to the user.
939 Register Cache Cleanup (below from Andrew Cagney)
941 I would depict the current register architecture as something like:
948 register + REGISTER_BYTE(reg_nr)
951 -------------------------
952 | extern register[] |
953 -------------------------
955 where neither the high (valops.c et.al.) or low gdb (*-tdep.c) are
956 really clear on what mechanisms they should be using to manipulate that
957 buffer. Further, much code assumes, dangerously, that registers are
958 contigious. Having got mips-tdep.c to support multiple ABIs, believe
959 me, that is a bad assumption. Finally, that register cache layout is
960 determined by the current remote/local target and _not_ the less
961 specific target ISA. In fact, in many cases it is determined by the
962 somewhat arbitrary layout of the [gG] packets!
965 How I would like the register file to work is more like:
993 The main objectives being:
995 o a clear separation between the low
996 level target and the high level GDB
998 o a mechanism that solves the general
999 problem of register aliases, overlaps
1000 etc instead of treating them as optional
1001 extras that can be wedged in as an after
1002 thought (that is a reasonable description
1003 of the current code).
1005 Identify then solve the hard case and the
1006 rest just falls out. GDB solved the easy
1007 case and then tried to ignore the real
1010 o a removal of the assumption that the
1011 mapping between the register cache
1012 and virtual registers is largely static.
1013 If you flip the USR/SSR stack register
1014 select bit in the status-register then
1015 the corresponding stack registers should
1018 o a mechanism that clearly separates the
1019 gdb internal register cache from any
1020 target (not architecture) dependent
1021 specifics such as [gG] packets.
1023 Of course, like anything, it sounds good in theory. In reality, it
1024 would have to contend with many<->many relationships at both the
1025 virt<->cache and cache<->target level. For instance:
1028 Modifying an mmx register may involve
1029 scattering values across both FP and
1030 mmpx specific parts of a buffer
1033 When writing back a SP it may need to
1034 both be written to both SP and USP.
1039 Rather than let this like the last time it was discussed, just slip, I'm
1040 first going to add this e-mail (+ references) to TODO. I'd then like to
1041 sketch out a broad strategy I think could get us there.
1044 First thing I'd suggest is separating out the ``extern registers[]''
1045 code so that we can at least identify what is using it. At present
1046 things are scattered across many files. That way we can at least
1047 pretend that there is a cache instead of a global array :-)
1049 I'd then suggest someone putting up a proposal for the pseudo-reg /
1050 high-level side interface so that code can be adopted to it. For old
1051 code, initially a blanket rename of write_register_bytes() to
1052 deprecated_write_register_bytes() would help.
1054 Following that would, finaly be the corresponding changes to the target.
1058 Check that GDB can handle all BFD architectures (Andrew Cagney)
1060 There should be a test that checks that BFD/GDB are in sync with
1061 regard to architecture changes. Something like a test that first
1062 queries GDB for all supported architectures and then feeds each back
1063 to GDB.. Anyone interested in learning how to write tests? :-)
1067 Architectural Change: Multi-arch et al.
1068 =======================================
1070 The long term objective is to remove all assumptions that there is a
1071 single target with a single address space with a single instruction
1072 set architecture and single application binary interface.
1074 This is an ongoing effort. The first milestone is to enable
1075 ``multi-arch'' where by all architectural decisions are made at
1078 It should be noted that ``gdbarch'' is really ``gdbabi'' and
1079 ``gdbisa''. Once things are multi-arched breaking that down correctly
1080 will become much easier.
1084 GDBARCH cleanup (Andrew Cagney)
1086 The non-generated parts of gdbarch.{sh,h,c} should be separated out
1087 into arch-utils.[hc].
1089 Document that gdbarch_init_ftype could easily fail because it didn't
1090 identify an architecture.
1094 Fix BELIEVE_PPC_PROMOTION. Change it to BELIEVE_PPC_PROMOTION_P?
1096 At present there is still #ifdef BELIEVE_PPC_PROMOTION code in the
1101 Fix target_signal_from_host() etc.
1103 The name is wrong for starters. ``target_signal'' should probably be
1104 ``gdb_signal''. ``from_host'' should be ``from_target_signal''.
1105 After that it needs to be multi-arched and made independent of any
1106 host signal numbering.
1110 Update ALPHA so that it uses ``struct frame_extra_info'' instead of
1113 This is a barrier to replacing mips_extra_func_info with something
1114 that works with multi-arch.
1118 Multi-arch mips_extra_func_info.
1120 This first needs the alpha to be updated so that it uses ``struct
1125 Rationalize TARGET_SINGLE_FORMAT and TARGET_SINGLE_BIT et al.
1127 Surely one of them is redundant.
1131 Convert ALL architectures to MULTI-ARCH.
1135 Select the initial multi-arch ISA / ABI based on --target or similar.
1137 At present the default is based on what ever is first in the BFD
1138 archures table. It should be determined based on the ``--target=...''
1143 Make MIPS pure multi-arch.
1145 It is only at the multi-arch enabled stage.
1151 Enable the code to recognize --enable-targets=.... like BINUTILS does.
1153 Can the tm.h and nm.h files be eliminated by multi-arch.
1157 Architectural Change: MI, LIBGDB and scripting languages
1158 ========================================================
1160 See also architectural changes related to the event loop. LIBGDB
1161 can't be finished until there is a generic event loop being used by
1164 The long term objective is it to be possible to integrate GDB into
1165 scripting languages.
1169 Implement generic ``(gdb) commmand > file''
1171 Once everything is going through ui_file it should be come fairly
1174 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00104.html
1178 Replace gdb_stdtarg with gdb_targout (and possibly gdb_targerr).
1180 gdb_stdtarg is easily confused with gdb_stdarg.
1184 Extra ui_file methods - dump.
1186 Very useful for whitebox testing.
1190 Eliminate error_begin().
1192 With ui_file, there is no need for the statefull error_begin ()
1197 Send normal output to gdb_stdout.
1198 Send error messages to gdb_stderror.
1199 Send debug and log output log gdb_stdlog.
1201 GDB still contains many cases where (f)printf or printf_filtered () is
1202 used when it should be sending the messages to gdb_stderror or
1203 gdb_stdlog. The thought of #defining printf to something has crossed
1208 Re-do GDB's output pager.
1210 GDB's output pager still relies on people correctly using *_filtered
1211 for gdb_stdout and *_unfiltered for gdb_stdlog / gdb_stderr.
1212 Hopefully, with all normal output going to gdb_stdout, the pager can
1213 just look at the ui_file that the output is on and then use that to
1214 decide what to do about paging. Sounds good in theory.
1218 Check/cleanup MI documentation.
1220 The list of commands specified in the documentation needs to be
1221 checked against the mi-cmds.c table in a mechanical way (so that they
1222 two can be kept up-to-date).
1226 Convert MI into libgdb
1228 MI provides a text interface into what should be many of the libgdb
1229 functions. The implementation of those functions should be separated
1230 into the MI interface and the functions proper. Those functions being
1231 moved to gdb/lib say.
1237 The first part can already be found in defs.h.
1241 MI's input does not use buffering.
1243 At present the MI interface reads raw characters of from an unbuffered
1244 FD. This is to avoid several nasty buffer/race conditions. That code
1245 should be changed so that it registers its self with the event loop
1246 (on the input FD) and then push commands up to MI as they arrive.
1248 The serial code already does this.
1252 Make MI interface accessible from existing CLI.
1256 Add a breakpoint-edit command to MI.
1258 It would be similar to MI's breakpoint create but would apply to an
1259 existing breakpoint. It saves the need to delete/create breakpoints
1260 when ever they are changed.
1264 Add directory path to MI breakpoint.
1266 That way the GUI's task of finding the file within which the
1267 breakpoint was set is simplified.
1271 Add a mechanism to reject certain expression classes to MI
1273 There are situtations where you don't want GDB's expression
1274 parser/evaluator to perform inferior function calls or variable
1275 assignments. A way of restricting the expression parser so that such
1276 operations are not accepted would be very helpful.
1280 Remove sideffects from libgdb breakpoint create function.
1282 The user can use the CLI to create a breakpoint with partial
1283 information - no file (gdb would use the file from the last
1286 The libgdb interface currently affects that environment which can lead
1287 to confusion when a user is setting breakpoints via both the MI and
1290 This is also a good example of how getting the CLI ``right'' will be
1295 Move gdb_lasterr to ui_out?
1297 The way GDB throws errors and records them needs a re-think. ui_out
1298 handles the correct output well. It doesn't resolve what to do with
1299 output / error-messages when things go wrong.
1303 do_setshow_command contains a 1024 byte buffer.
1305 The function assumes that there will never be any more than 1024 bytes
1306 of enum. It should use mem_file.
1310 Should struct cmd_list_element . completer take the command as an
1315 Should the bulk of top.c:line_completion_function() be moved to
1316 command.[hc]? complete_on_cmdlist() and complete_on_enums() could
1317 then be made private.
1321 top.c (execute_command): Should a command being valid when the target
1322 is running be made an attribute (predicate) to the command rather than
1323 an explicit set of tests.
1327 top.c (execute_command): Should the bulk of this function be moved
1328 into command.[hc] so that top.c doesn't grub around in the command
1333 Architectural Change: Async
1334 ===========================
1336 While GDB uses an event loop when prompting the user for input. That
1337 event loop is not exploited by targets when they allow the target
1338 program to continue. Typically targets still block in (target_wait())
1339 until the program again halts.
1341 The closest a target comes to supporting full asynchronous mode are
1342 the remote targets ``async'' and ``extended-async''.
1346 Asynchronous expression evaluator
1348 Inferior function calls hang GDB.
1352 Fix implementation of ``target xxx''.
1354 At present when the user specifies ``target xxxx'', the CLI maps that
1355 directly onto a target open method. It is then assumed that the
1356 target open method should do all sorts of complicated things as this
1357 is the only chance it has. Check how the various remote targets
1358 duplicate the target operations. Check also how the various targets
1359 behave differently for purely arbitrary reasons.
1361 What should happen is that ``target xxxx'' should call a generic
1362 ``target'' function and that should then co-ordinate the opening of
1363 ``xxxx''. This becomes especially important when you're trying to
1364 open an asynchronous target that may need to perform background tasks
1365 as part of the ``attach'' phase.
1367 Unfortunately, due to limitations in the old/creaking command.h
1368 interface, that isn't possible. The function being called isn't told
1369 of the ``xxx'' or any other context information.
1371 Consequently a precursor to fixing ``target xxxx'' is to clean up the
1372 CLI code so that it passes to the callback function (attatched to a
1373 command) useful information such as the actual command and a context
1374 for that command. Other changes such as making ``struct command''
1375 opaque may also help.
1378 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
1382 Make "target xxx" command interruptible.
1384 As things become async this becomes possible. A target would start
1385 the connect and then return control to the event loop. A cntrl-c
1386 would notify the target that the operation is to be abandoned and the
1387 target code could respond.
1391 Add a "suspend" subcommand of the "continue" command to suspend gdb
1392 while continuing execution of the subprocess. Useful when you are
1393 debugging servers and you want to dodge out and initiate a connection
1394 to a server running under gdb.
1403 Frequently requested but not approved requests.
1407 Eliminate unused argument warnings using ATTRIBUTE_UNUSED.
1409 The benefits on this one are thought to be marginal - GDBs design
1410 means that unused parameters are very common. GCC 3.0 will also
1411 include the option -Wno-unused-parameter which means that ``-Wall
1412 -Wno-unused-parameters -Werror'' can be specified.
1421 This list is not up to date, and opinions vary about the importance or
1422 even desirability of some of the items. If you do fix something, it
1423 always pays to check the below.
1427 @c This does not work (yet if ever). FIXME.
1428 @c @item --parse=@var{lang} @dots{}
1429 @c Configure the @value{GDBN} expression parser to parse the listed languages.
1430 @c @samp{all} configures @value{GDBN} for all supported languages. To get a
1431 @c list of all supported languages, omit the argument. Without this
1432 @c option, @value{GDBN} is configured to parse all supported languages.
1436 START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that
1437 is its default value. Clean this up.
1441 It should be possible to use symbols from shared libraries before we know
1442 exactly where the libraries will be loaded. E.g. "b perror" before running
1443 the program. This could maybe be done as an extension of the "breakpoint
1444 re-evaluation" after new symbols are loaded.
1448 Make single_step() insert and remove breakpoints in one operation.
1450 [If this is talking about having single_step() insert the breakpoints,
1451 run the target then pull the breakpoints then it is wrong. The
1452 function has to return as control has to eventually be passed back to
1453 the main event loop.]
1457 Speed up single stepping by avoiding extraneous ptrace calls.
1461 Speed up single stepping by not inserting and removing breakpoints
1462 each time the inferior starts and stops.
1464 Breakpoints should not be inserted and deleted all the time. Only the
1465 one(s) there should be removed when we have to step over one. Support
1466 breakpoints that don't have to be removed to step over them.
1468 [this has resulted in numerous debates. The issue isn't clear cut]
1472 Provide "voodoo" debugging of core files. This creates a zombie
1473 process as a child of the debugger, and loads it up with the data,
1474 stack, and regs of the core file. This allows you to call functions
1475 in the executable, to manipulate the data in the core file.
1481 GDB reopens the source file on every line, as you "next" through it.
1483 [still true? I've a memory of this being fixed]
1487 Perhaps "i source" should take an argument like that of "list".
1491 Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if
1492 it matches the source line indicated.
1496 The prompt at end of screen should accept space as well as CR.
1500 Backtrace should point out what the currently selected frame is, in
1501 its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar,
1502 ...)" rather than "#3 foo (bar, ...)".
1506 "i program" should work for core files, and display more info, like what
1507 actually caused it to die.
1511 "x/10i" should shorten the long name, if any, on subsequent lines.
1515 "next" over a function that longjumps, never stops until next time you happen
1516 to get to that spot by accident. E.g. "n" over execute_command which has
1521 "set zeroprint off", don't bother printing members of structs which
1522 are entirely zero. Useful for those big structs with few useful
1527 GDB does four ioctl's for every command, probably switching terminal modes
1528 to/from inferior or for readline or something.
1532 terminal_ours versus terminal_inferior: cache state. Switch should be a noop
1533 if the state is the same, too.
1537 "i frame" shows wrong "arglist at" location, doesn't show where the args
1538 should be found, only their actual values.
1542 There should be a way for "set" commands to validate the new setting
1543 before it takes effect.
1547 "ena d" is ambiguous, why? "ena delete" seems to think it is a command!
1551 i line VAR produces "Line number not known for symbol ``var''.". I
1552 thought we were stashing that info now!
1556 We should be able to write to random files at hex offsets like adb.
1560 [elena - delete this]
1562 Handle add_file with separate text, data, and bss addresses. Maybe
1563 handle separate addresses for each segment in the object file?
1567 [Jimb/Elena delete this one]
1569 Handle free_named_symtab to cope with multiply-loaded object files
1570 in a dynamic linking environment. Should remember the last copy loaded,
1571 but not get too snowed if it finds references to the older copy.
1575 [elena delete this also]
1577 Remove all references to:
1584 now that we have BFD. All remaining are in machine dependent files.
1588 Re-organize help categories into things that tend to fit on a screen
1593 Add in commands like ADB's for searching for patterns, etc. We should
1594 be able to examine and patch raw unsymboled binaries as well in gdb as
1595 we can in adb. (E.g. increase the timeout in /bin/login without source).
1597 [actually, add ADB interface :-]
1601 When doing "step" or "next", if a few lines of source are skipped between
1602 the previous line and the current one, print those lines, not just the
1603 last line of a multiline statement.
1607 Handling of "&" address-of operator needs some serious overhaul
1608 for ANSI C and consistency on arrays and functions.
1609 For "float point[15];":
1610 ptype &point[4] ==> Attempt to take address of non-lvalue.
1611 For "char *malloc();":
1612 ptype malloc ==> "char *()"; should be same as
1613 ptype &malloc ==> "char *(*)()"
1614 call printf ("%x\n", malloc) ==> weird value, should be same as
1615 call printf ("%x\n", &malloc) ==> correct value
1619 Fix dbxread.c symbol reading in the presence of interrupts. It
1620 currently leaves a cleanup to blow away the entire symbol table when a
1621 QUIT occurs. (What's wrong with that? -kingdon, 28 Oct 1993).
1623 [I suspect that the grype was that, on a slow system, you might want
1624 to cntrl-c and get just half the symbols and then load the rest later
1625 - scary to be honest]
1629 Mipsread.c reads include files depth-first, because the dependencies
1630 in the psymtabs are way too inclusive (it seems to me). Figure out what
1631 really depends on what, to avoid recursing 20 or 30 times while reading
1636 value_add() should be subtracting the lower bound of arrays, if known,
1637 and possibly checking against the upper bound for error reporting.
1641 When listing source lines, check for a preceding \n, to verify that
1642 the file hasn't changed out from under us.
1644 [fixed by some other means I think. That hack wouldn't actually work
1645 reliably - the file might move such that another \n appears. ]
1649 Get all the remote systems (where the protocol allows it) to be able to
1650 stop the remote system when the GDB user types ^C (like remote.c
1651 does). For ebmon, use ^Ak.
1655 Possible feature: A version of the "disassemble" command which shows
1656 both source and assembly code ("set symbol-filename on" is a partial
1659 [has this been done? It was certainly done for MI and GDBtk]
1663 investigate "x/s 0" (right now stops early) (I think maybe GDB is
1664 using a 0 address for bad purposes internally).
1668 Make "info path" and path_command work again (but independent of the
1669 environment either of gdb or that we'll pass to the inferior).
1673 Make GDB understand the GCC feature for putting octal constants in
1674 enums. Make it so overflow on an enum constant does not error_type
1675 the whole type. Allow arbitrarily large enums with type attributes.
1676 Put all this stuff in the testsuite.
1680 Make TYPE_CODE_ERROR with a non-zero TYPE_LENGTH more useful (print
1681 the value in hex; process type attributes). Add this to the
1682 testsuite. This way future compilers can add new types and old
1683 versions of GDB can do something halfway reasonable.
1687 Fix mdebugread.c:parse_type to do fundamental types right (see
1688 rs6000_builtin_type in stabsread.c for what "right" is--the point is
1689 that the debug format fixes the sizes of these things and it shouldn't
1690 depend on stuff like TARGET_PTR_BIT and so on. For mdebug, there seem
1691 to be separate bt* codes for 64 bit and 32 bit things, and GDB should
1692 be aware of that). Also use a switch statement for clarity and speed.
1696 Investigate adding symbols in target_load--some targets do, some
1701 Put dirname in psymtabs and change lookup*symtab to use dirname (so
1702 /foo/bar.c works whether compiled by cc /foo/bar.c, or cd /foo; cc
1707 Merge xcoffread.c and coffread.c. Use breakpoint_re_set instead of
1712 Make a watchpoint which contains a function call an error (it is
1713 broken now, making it work is probably not worth the effort).
1717 New test case based on weird.exp but in which type numbers are not
1718 renumbered (thus multiply defining a type). This currently causes an
1719 infinite loop on "p v_comb".
1723 [Hey! Hint Hint Delete Delete!!!]
1725 Fix 386 floating point so that floating point registers are real
1726 registers (but code can deal at run-time if they are missing, like
1727 mips and 68k). This would clean up "info float" and related stuff.
1731 gcc -g -c enummask.c then gdb enummask.o, then "p v". GDB complains
1732 about not being able to access memory location 0.
1734 -------------------- enummask.c
1755 If try to modify value in file with "set write off" should give
1756 appropriate error not "cannot access memory at address 0x65e0".
1760 Allow core file without exec file on RS/6000.
1764 Make sure "shell" with no arguments works right on DOS.
1768 Make gdb.ini (as well as .gdbinit) be checked on all platforms, so
1769 the same directory can be NFS-mounted on unix or DOS, and work the
1774 [Is this another delete???]
1776 Get SECT_OFF_TEXT stuff out of objfile_relocate (might be needed to
1777 get RS/6000 to work right, might not be immediately relevant).
1781 Work out some kind of way to allow running the inferior to be done as
1782 a sub-execution of, eg. breakpoint command lists. Currently running
1783 the inferior interupts any command list execution. This would require
1784 some rewriting of wait_for_inferior & friends, and hence should
1785 probably be done in concert with the above.
1789 Add function arguments to gdb user defined functions.
1793 Add convenience variables that refer to exec file, symbol file,
1794 selected frame source file, selected frame function, selected frame
1799 Modify the handling of symbols grouped through BINCL/EINCL stabs to
1800 allocate a partial symtab for each BINCL/EINCL grouping. This will
1801 seriously decrease the size of inter-psymtab dependencies and hence
1802 lessen the amount that needs to be read in when a new source file is
1807 Add a command for searching memory, a la adb. It specifies size,
1808 mask, value, start address. ADB searches until it finds it or hits
1809 an error (or is interrupted).
1813 Remove the range and type checking code and documentation, if not