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 Update GDB's coding standard documentation. Known topics:
17 o alloca/malloc et.al.
29 Wow, three bug reports for the same problem in one day! We should
30 probably make fixing this a real priority :-).
32 Anyway, thanks for reporting.
34 The following patch will fix the problems with setting breakpoints in
35 dynamically loaded objects:
37 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00230.html
39 This patch isn't checked in yet (ping Michael/JimB), but I hope this
40 will be in the next GDB release.
42 There should really be a test in the testsuite for this problem, since
43 it keeps coming up :-(. Any volunteers?
49 x86 linux GDB and SIGALRM (???)
50 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00803.html
52 This problem has been fixed, but a regression test still needs to be
53 added to the testsuite:
54 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00309.html
58 [The test has been submitted for approval - cagney]
62 RFD: infrun.c: No bpstat_stop_status call after proceed over break?
63 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00665.html
65 GDB misses watchpoint triggers after proceeding over a breakpoint on
70 GDB 5.0 doesn't work on Linux/SPARC
72 There are two parts to this.
74 o GDB 5.0 doesn't work on GNU/Linux/SPARC32
76 o GDB 5.0 doesn't work on the new target
79 GDB does build on both these targets.
81 The first problem is the one that should be fixed.
85 GDB 5.1 - New features
86 ======================
88 The following new features should be included in 5.1.
92 Enable MI by default. Old code can be deleted after 5.1 is out.
96 o syntax change where a list would
98 [ foo=a, foo=b, foo=c ]
100 { foo=a, foo=b, foo=c }
102 o kill off the idea of a reverse
111 Pascal (Pierre Muller, David Taylor)
113 Pierre Muller has contributed patches for adding Pascal Language
116 2 pascal language patches inserted in database
117 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00521.html
120 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00496.html
122 [I think this has been merged, need to confirm - cagney]
126 Java (Anthony Green, David Taylor)
128 Anthony Green has a number of Java patches that did not make it into
129 the 5.0 release. The first two are in cvs now, but the third needs
130 some fixing up before it can go in.
133 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00512.html
136 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00515.html
138 Patch: handle N_MAIN stab
139 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00527.html
145 A predicate to this is the multi-arching of SOFTWARE_SINGLE_STEP(). A
146 patch has been submitted.
153 The following code cleanups will hopefully be applied to GDB 5.1.
157 Resolve the build status of all broken targets as identified by the
160 o arm-* vs NetBSD's lack of ``unix''
161 o arm-* vs IRIX (see below)
169 Obsolete some targets.
171 Possible selection criteria are:
173 o uses a deprecated feature
177 o doesn't have a maintainer
181 o post proposals to gdb@ (DONE)
183 o post announcement to gdb-announce@
184 crossed with gdb@ reply-to to gdb@
206 configure.{in,host,tgt}
208 *-tdep.c *-nat.c *-xdep.c
215 Change documentation to GFDL license.
217 ``It is time to make an effort to start using the GFDL more
218 thoroughly. Would all GNU maintainers please change the license to
219 the GFDL, for all manuals and other major documentation files?
221 The GFDL and some instructions for using it can be found in
222 http://www.gnu.org/copyleft/''
228 Fix copyright notices.
230 Turns out that ``1998-2000'' isn't considered valid :-(
232 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00467.html
236 GDB 5.1 - Known Problems
237 ========================
243 The z8k has suffered bit rot and is known to not build. The problem
244 was occuring in the opcodes directory.
248 Solaris 8 x86 CURSES_H problem
249 http://sources.redhat.com/ml/gdb/2000-07/msg00038.html
251 The original problem was worked around with:
253 2000-06-06 Michael Snyder <msnyder@cygnus.com>
255 * configure.in: Enable autoconf to find curses.h on Solaris 2.8.
256 * configure: Regenerate.
258 When building both GDB and SID using the same source tree the problem
259 will still occure. sid/component/configure.in mis-configures
260 <curses.h> and leaves wrong information in the config cache.
269 Thread support. Right now, as soon as a thread finishes and exits,
270 you're hosed. This problem is reported once a week or so.
274 GDB 5.2 - New features
275 ======================
279 GCC 3.0 ABI support (but hopefully sooner...).
283 Objective C/C++ support (but hopefully sooner...).
290 The following cleanups have been identified as part of GDB 5.2.
294 Remove old code that does not use ui_out functions and all the related
295 "ifdef"s. This also allows the elimination of -DUI_OUT from
296 Makefile.in and configure.in.
302 Eliminate all warnings for at least one host/target for the flags:
303 -Wimplicit -Wreturn-type -Wcomment -Wtrigraphs -Wformat -Wparentheses
304 -Wpointer-arith -Wuninitialized
308 Deprecate, if not delete, the following:
314 Replaced by, on the target side
316 and on core-gdb side:
317 {read,write}_register_gen()
318 Remote.c will need to use something
319 other than REGISTER_BYTE() and
320 REGISTER_RAW_SIZE() when unpacking
323 STORE_PSEUDO_REGISTER
324 FETCH_PSEUDO_REGISTER
325 Now handed by the methods
326 gdbarch_{read,write}_register()
327 which sits between core GDB and
331 REGISTER_CONVERT_TO_RAW
332 REGISTER_CONVERT_TO_VIRTUAL
333 I think these three are redundant.
334 gdbarch_register_{read,write} can
335 do any conversion it likes.
337 REGISTER_VIRTUAL_SIZE
338 MAX_REGISTER_VIRTUAL_SIZE
339 REGISTER_VIRTUAL_TYPE
340 I think these can be replaced by
342 FRAME_REGISTER_TYPE(frame, regnum)
343 REGISTER_TYPE(regnum)
347 FRAME_REGISTER_INFO (frame, ...)
350 If nothing else rename this so that
351 how it relates to rawreg and the
355 The size of the cache can be computed
360 Restructure gdb directory tree so that it avoids any 8.3 and 14
365 Convert GDB build process to AUTOMAKE.
367 See also sub-directory configure below.
369 The current convention is (kind of) to use $(<header>_h) in all
370 dependency lists. It isn't done in a consistent way.
374 GDB 5.2 - Known Problems
375 ========================
379 Code Cleanups: General
380 ======================
382 The following are more general cleanups and fixes. They are not tied
383 to any specific release.
387 Rename read_register{,_pid}() to read_unsigned_register{,_pid}().
391 Can't build IRIX -> arm GDB.
392 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00356.html
395 > Now I'm building for an embedded arm target. If there is a way of turning
396 > remote-rdi off, I couldn't find it. It looks like it gets built by default
397 > in gdb/configure.tgt(line 58) Anyway, the build dies in
398 > gdb/rdi-share/unixcomm.c. SERPORT1 et. al. never get defined because we
399 > aren't one of the architectures supported.
403 Problem with weak functions
404 http://sourceware.cygnus.com/ml/gdb/2000-05/msg00060.html
406 Dan Nicolaescu writes:
407 > It seems that gdb-4.95.1 does not display correctly the function when
408 > stoping in weak functions.
410 > It stops in a function that is defined as weak, not in the function
411 > that is actually run...
415 Follow through `make check' with --enable-shared.
417 When the srcware tree is configured with --enable-shared, the `expect'
418 program won't run properly. Jim Wilson found out gdb has a local hack
419 to set LD_LIBRARY_PATH, but, AFAIK, no other project has been hacked
422 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00845.html
426 Delete macro TARGET_BYTE_ORDER_SELECTABLE.
428 Patches in the database.
432 printcmd.c (print_address_numeric):
434 NOTE: This assumes that the significant address information is kept in
435 the least significant bits of ADDR - the upper bits were either zero
436 or sign extended. Should ADDRESS_TO_POINTER() or some
437 ADDRESS_TO_PRINTABLE() be used to do the conversion?
441 The BFD directory requires bug-fixed AUTOMAKE et.al.
443 AUTOMAKE 1.4 incorrectly set the TEXINPUTS environment variable. It
444 contained the full path to texinfo.tex when it should have only
445 contained the directory. The bug has been fixed in the current
446 AUTOMAKE sources. Automake snapshots can be found in:
447 ftp://sourceware.cygnus.com/pub/gdb/snapshots
448 and ftp://sourceware.cygnus.com/pub/binutils
452 Find something better than DEFAULT_BFD_ARCH, DEFAULT_BFD_VEC to
453 determine the default isa/byte-order.
457 Rely on BFD_BIG_ENDIAN and BFD_LITTLE_ENDIAN instead of host dependent
458 BIG_ENDIAN and LITTLE_ENDIAN.
462 Eliminate more compiler warnings.
464 Of course there also needs to be the usual debate over which warnings
465 are valid and how to best go about this.
467 One method: choose a single option; get agreement that it is
468 reasonable; try it out to see if there isn't anything silly about it
469 (-Wunused-parameters is an example of that) then incrementally hack
472 The other method is to enable all warnings and eliminate them from one
477 Elimination of ``(catch_errors_ftype *) func''.
479 Like make_cleanup_func it isn't portable.
480 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00791.html
481 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00814.html
485 Nuke #define CONST_PTR.
493 [PATCH/5] src/intl/Makefile.in:distclean additions
494 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00363.html
496 Do not forget to merge the patch back into the trunk.
500 Rationalize the host-endian code (grep for HOST_BYTE_ORDER).
502 At present defs.h includes <endian.h> (which is linux specific) yet
503 almost nothing depends on it. Suggest "gdb_endian.h" which can also
504 handle <machine/endian.h> and only include that where it is really
509 Replace savestring() with something from libiberty.
511 An xstrldup()? but that would have different semantics.
515 Rationalize use of floatformat_unknown in GDB sources.
517 Instead of defaulting to floatformat_unknown, should hosts/targets
518 specify the value explicitly?
520 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
524 Add a ``name'' member to include/floatformat.h:struct floatformat.
525 Print that name in gdbarch.c.
529 Sort out the harris mess in include/floatformat.h (it hardwires two
530 different floating point formats).
534 See of the GDB local floatformat_do_doublest() and libiberty's
535 floatformat_to_double (which was once GDB's ...) can be merged some
540 Eliminate mmalloc(), mstrsave() et.al. from GDB.
542 Also eliminate it from defs.h.
546 Eliminate PTR. ISO-C allows ``void *''.
552 GDB should never abort. GDB should either throw ``error ()'' or
553 ``internal_error ()''. Better still GDB should naturally unwind with
558 GDB probably doesn't build on FreeBSD pre 2.2.x
559 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00378.html
561 Fixes to get FreeBSD working on 2.2.x, 3.x and 4.x caused the code to
566 Deprecate "fg". Apparently ``fg'' is actually continue.
568 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00417.html
572 Deprecate current use of ``floatformat_unknown''.
574 Require all targets to explicitly provide their float format instead
575 of defaulting to floatformat unknown. Doing the latter leads to nasty
578 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
582 Rationalize floatformat_to_double() vs floatformat_to_doublest().
584 Looks like GDB migrated floatformat_to_double() to libiberty but then
585 turned around and created a ..._to_doublest() the latter containing
588 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00472.html
592 Move floatformat_ia64_ext to libiberty/include floatformat.[ch].
594 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00466.html
598 The ``maintenance deprecate set endian big'' command doesn't notice
599 that it is deprecating ``set endian'' and not ``set endian big'' (big
600 is implemented using an enum). Is anyone going to notice this?
604 When tab expanding something like ``set arch<tab>'' ignore the
605 deprecated ``set archdebug'' and expand to ``set architecture''.
609 Eliminate ``arm_register_names[j] = (char *) regnames[j]'' and the
610 like from arm-tdep.c.
614 Fix uses of ->function.cfunc = set_function().
616 The command.c code calls sfunc() when a set command. Rather than
617 change it suggest fixing the callback function so that it is more
620 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
622 See also ``Fix implementation of ``target xxx''.'' below.
626 IRIX 3.x support is probably broken.
630 Delete sim/SIM_HAVE_BREAKPOINTS and gdb/SIM_HAS_BREAKPOINTS.
631 http://sourceware.cygnus.com/ml/gdb-patches/2000-07/msg00042.html
633 Apart from the d30v, are there any sim/common simulators that make use
636 A brief summary of what happened is that sim/common/sim-break.c was
637 created as a good idea. It turned out a better idea was to use
638 SIM_SIGBREAK and have GDB pass back sim_resume (..., SIGBREAK).
642 Move remote_remove_hw_breakpoint, remote_insert_hw_breakpoint,
643 remote_remove_watchpoint, remote_insert_watchpoint into target vector.
647 Eliminate ``extern'' from C files.
651 Replace ``STREQ()'' et.al. with ``strcmp() == 0'' et.al.
653 Extreme care is recommeded - perhaps only modify tests that are
654 exercised by the testsuite (as determined using some type of code
659 Replace the file gdb/CONTRIBUTE with a file that is generated from the
660 gdb/doc/*.texinfo directory.
664 New Features and Fixes
665 ======================
667 These are harder than cleanups but easier than work involving
668 fundamental architectural change.
672 Hardware watchpoint problems on x86 OSes, including Linux:
674 1. Delete/disable hardware watchpoints should free hardware debug
676 2. Watch for different values on a viariable with one hardware debug
679 According to Eli Zaretskii <eliz@delorie.com>:
681 These are not GDB/ia32 issues per se: the above features are all
682 implemented in the DJGPP port of GDB and work in v5.0. Every
683 x86-based target should be able to lift the relevant parts of
684 go32-nat.c and use them almost verbatim. You get debug register
685 sharing through reference counts, and the ability to watch large
686 regions (up to 16 bytes) using multiple registers. (The required
687 infrastructure in high-level GDB application code, mostly in
688 breakpoint.c, is also working since v5.0.)
692 Add built-by, build-date, tm, xm, nm and anything else into gdb binary
693 so that you can see how the GDB was created.
697 Add an "info bfd" command that displays supported object formats,
698 similarly to objdump -i.
700 Is there a command already?
704 Fix ``I'm sorry, Dave, I can't do that.'' from symfile.c.
706 This requires internationalization.
712 (gdb) p fwprintf(stdout,L"%S\n", f)
713 No symbol "L" in current context.
717 Cleanup configury support for optional sub-directories.
719 Check how GCC handles multiple front ends for an example of how things
720 could work. A tentative first step is to rationalize things so that
721 all sub directories are handled in a fashion similar to gdb/mi.
723 See also automake above.
727 Add a transcript mechanism to GDB.
729 Such a mechanism might log all gdb input and output to a file in a
730 form that would allow it to be replayed. It could involve ``gdb
731 --transcript=FILE'' or it could involve ``(gdb) transcript file''.
735 Can the xdep files be replaced by autoconf?
739 Document trace machinery
743 Document ui-out and ui-file.
745 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00121.html
749 Update texinfo.tex to latest?
753 Incorporate agentexpr.texi into gdb.texinfo
755 agentexpr.texi mostly describes the details of the byte code used for
756 tracepoints, not the internals of the support for this in GDB. So it
757 looks like gdb.texinfo is a better place for this information.
759 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00566.html
763 Document overlay machinery.
767 ``(gdb) catch signal SIGNAL''
769 Overlaps with ``handle SIGNAL'' but the implied behavior is different.
770 You can attach commands to a catch but not a handle. A handle has a
771 limited number of hardwired actions.
777 o readline/*.h bitrot
779 The TUI isn't up-to-date with
780 respect to the readline currently
781 bundled with GDB. Importing a
782 new readline is on the 5.1 wish
783 list so this can only get worse.
785 Grep for things like term_cursor_move.
787 (To be honest, I don't see anyone
788 importing a new readline before 5.1 is
791 o tui.c:va_catch_errors() bitrot
793 This nasty piece of work used knowledge
794 of the internals of GDBs error functions :-(
795 Ever since those internals were cleaned
796 up this code has been broken. :-(
798 o tuiWin.c:c_makeVisibleWithNewHeight() broken
799 tuiLayout.c:_extractDisplayStartAddr() broken
801 Both these function call find_line_pc()
802 incorrectly (wrong args, wrong return value).
804 I suspect this bug has always been there!
805 It had been hidden because those files
806 didn't include the necessary header files
809 o tuiRegs() host dependant
811 Not suprisingly, this isn't a very portable
812 section of code. However, I'm sure people
813 could live with no regs in the short to
816 o defs.h: #include "tui.h" et.al.
818 I'm not sure where this came from.
819 It was a really bad idea.
821 To get things to compile I did a nasty
822 hack (Just declare what was needed and
823 replace any expressions like xx->y.z()
824 in GDB proper with function calls). I
825 could commit it slightly cleaned up if
828 Medium Term. the #ifdef TUI and TuiDo()
829 should be changed to hooks (like GDBTK).
830 The gdb-events.[hc] is there for that
833 o tui.c:_tuiReset() host dependant
835 tui.c contains a lump of termio[s]
836 I suspect an equivalent block of
837 code can be lifted from readline.
838 An equivalent readline function may
841 o curses.h vs ncurses.h.
843 Simple portability problem.
847 This function is a mystery - where is it?
849 o tui-file.[hc] cleanup
851 This can be significantly simplified.
853 o The code should be pacified. (-Werror -W...)
855 There are plenty of #includes,
856 duplicate #includes, missing function decls
859 Some of the problems I found were through
860 fixing a few of the warnings.
862 o The code should be GNUtified.
864 It would be very nice to have this code
865 look like the rest of GDB. That way people
866 would be more accepting of it as a true
869 Until it is GNUtified it is going to stick
870 out like a sore thumb to the programmer.
872 o The code should be clearly copyrighted
874 (FSF, with due credit to HP)
878 Add support for ``gdb --- PROGRAM ARGS ...''.
879 Add support for ``gdb -cmd=...''
881 Along with many variations. Check:
883 ????? for a full discussion.
889 Implement ``(gdb) !ls''.
891 Which is very different from ``(gdb) ! ls''. Implementing the latter
894 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00034.html
898 Change the (char *list[]) to (const char (*)[]) so that dynamic lists can
903 When tab expanding something like ``set arch<tab>'' ignore the
904 deprecated ``set archdebug'' and expand to ``set architecture''.
908 Replace the code that uses the host FPU with an emulator of the target
913 The "ocd reset" command needs to flush the dcache, which requires breaking
914 the abstraction layer between the target independent and target code. One
915 way to address this is provide a generic "reset" command and target vector.
917 http://sources.redhat.com/ml/gdb-patches/2000-10/msg00011.html
926 Generic: lin-thread cannot handle thread exit (Mark Kettenis, Michael
927 Snyder) http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00525.html
929 The thread_db assisted debugging code doesn't handle exiting threads
930 properly, at least in combination with glibc 2.1.3 (the framework is
931 there, just not the actual code). There are at least two problems
932 that prevent this from working.
934 As an additional reference point, the pre thread_db code did not work
939 GNU/Linux/x86 and random thread signals (and Solaris/SPARC but not
941 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00336.html
943 Christopher Blizzard writes:
945 So, I've done some more digging into this and it looks like Jim
946 Kingdon has reported this problem in the past:
948 http://sourceware.cygnus.com/ml/bug-gdb/1999-10/msg00058.html
950 I can reproduce this problem both with and without Tom's patch. Has
951 anyone seen this before? Maybe have a solution for it hanging around?
954 There's a test case for this documented at:
956 when debugging threaded applications you get extra SIGTRAPs
957 http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=9565
959 [There should be a GDB testcase - cagney]
963 GDB5 TOT on unixware 7
964 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00119.html
967 > I just spun the top of tree of the GDB5 branch on UnixWare 7. As a
968 > practical matter, the current thread support is somewhat more annoying
969 > than when GDB was thread-unaware.
976 New languages come onto the scene all the time.
980 Re: Various C++ things
982 value_headof/value_from_vtable_info are worthless, and should be
983 removed. The one place in printcmd.c that uses it should use the RTTI
986 RTTI for g++ should be using the typeinfo functions rather than the
987 vtables. The typeinfo functions are always at offset 4 from the
988 beginning of the vtable, and are always right. The vtables will have
989 weird names like E::VB sometimes. The typeinfo function will always
990 be "E type_info function", or somesuch.
992 value_virtual_fn_field needs to be fixed so there are no failures for
993 virtual functions for C++ using g++.
995 Testsuite cases are the major priority right now for C++ support,
996 since i have to make a lot of changes that could potentially break
1001 Add support for Modula3
1003 Get DEC/Compaq to contribute their Modula-3 support.
1007 Remote Protocol Support
1008 =======================
1012 Revised UDP support (was: Re: [Fwd: [patch] UDP transport support])
1013 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00000.html
1015 (Broken) support for GDB's remote protocol across UDP is to be
1016 included in the follow-on release.
1018 It should be noted that UDP can only work when the [Gg] packet fits in
1019 a single UDP packet.
1021 There is also much debate over the merit of this.
1025 Migrate qfThreadInfo packet -> qThreadInfo. (Andrew Cagney)
1027 Add support for packet enable/disable commands with these thread
1028 packets. General cleanup.
1030 [PATCH] Document the ThreadInfo remote protocol queries
1031 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00832.html
1033 [PATCH] "info threads" queries for remote.c
1034 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00831.html
1038 Remote protocol doco feedback.
1040 Too much feedback to mention needs to be merged in (901660). Search
1041 for the word ``remote''.
1044 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00023.html
1045 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00056.html
1046 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00382.html
1050 GDB doesn't recover gracefully from remote protocol errors.
1052 GDB wasn't checking for NAKs from the remote target. Instead a NAK is
1053 ignored and a timeout is required before GDB retries. A pre-cursor to
1054 fixing this this is making GDB's remote protocol packet more robust.
1056 While downloading to a remote protocol target, gdb ignores packet
1057 errors in so far as it will continue to download with chunk N+1 even
1058 if chunk N was not correctly sent. This causes gdb.base/remote.exp to
1059 take a painfully long time to run. As a PS that test needs to be
1060 fixed so that it builds on 16 bit machines.
1064 Fix the ``!'' packet.
1066 JT reported that the existing targets do, in fact return ``OK'' so it
1067 is possible to merge remote and extended-remote targets.
1071 Drop ``<address>'' from the [SsCc] packets.
1073 I don't think that GDB generates them so having it in the protocol is
1078 Fix doco on the ``q'' packet.
1080 It has evolved into a generic RPC. The notes should reflect this and,
1081 perhaps, the ``Q'' packet can be deprecated.
1083 The doco should mention that ``OK'' is a valid packet response.
1085 The doco should explain why ``OK'' needs to be a valid packet
1090 Add the cycle step command.
1092 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00237.html
1096 Resolve how to scale things to support very large packets.
1100 Resolve how to handle a target that changes things like its endianess
1101 on the fly - should it be returned in the ``T'' packet?
1103 Underlying problem is that the register file is target endian. If the
1104 target endianess changes gdb doesn't know.
1111 If / when GDB starts to support the debugging of multi-processor
1112 (rather than multi-thread) applications the symtab code will need to
1113 be updated a little so that several independent symbol tables are
1114 active at a given time.
1116 The other interesting change is a clarification of the exact meaning
1117 of CORE_ADDR and that has had consequences for a few targets (that
1118 were abusing that data type).
1122 Investiagate ways of reducing memory.
1126 Investigate ways of improving load time.
1130 Get the d10v to use POINTER_TO_ADDRESS and ADDRESS_TO_POINTER.
1132 Consequence of recent symtab clarification. No marks for figuring out
1133 who maintains the d10v.
1137 Get the MIPS to correctly sign extend all address <-> pointer
1140 Consequence of recent symtab clarification. No marks for figuring out
1141 who maintains the MIPS.
1145 GDB truncates 64 bit enums.
1147 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00290.html
1154 There are never to many testcases.
1158 Better thread testsuite.
1162 Better C++ testsuite.
1166 Look at adding a GDB specific testsuite directory so that white box
1167 tests of key internals can be added (eg ui_file).
1171 Separate out tests that involve the floating point (FP).
1173 (Something for people brining up new targets). FP and non-fp tests
1174 are combined. I think there should be set of basic tests that
1175 exercise pure integer support and then a more expanded set that
1176 exercise FP and FP/integer interactions.
1178 As an example, the MIPS, for n32 as problems with passing FP's and
1179 structs. Since most inferior call tests include FP it is difficult to
1180 determine of the integer tests are ok.
1184 Architectural Changes: General
1185 ==============================
1187 These are harder than simple cleanups / fixes and, consequently
1188 involve more work. Typically an Architectural Change will be broken
1189 down into a more digestible set of cleanups and fixes.
1193 Cleanup software single step.
1195 At present many targets implement software single step by directly
1196 blatting memory (see rs6000-tdep.c). Those targets should register
1197 the applicable breakpoints using the breakpoint framework. Perhaphs a
1198 new internal breakpoint class ``step'' is needed.
1202 Replace READ_FP() with FRAME_HANDLE().
1204 READ_FP() is a hangover from the days of the vax when the ABI really
1205 did have a frame pointer register. Modern architectures typically
1206 construct a virtual frame-handle from the stack pointer and various
1207 other bits of string.
1209 Unfortunately GDB still treats this synthetic FP register as though it
1210 is real. That in turn really confuses users (arm and ``print $fp'' VS
1211 ``info registers fp''). The synthetic FP should be separated out of
1212 the true register set presented to the user.
1216 Register Cache Cleanup (below from Andrew Cagney)
1218 I would depict the current register architecture as something like:
1220 High GDB --> Low GDB
1225 register + REGISTER_BYTE(reg_nr)
1228 -------------------------
1229 | extern register[] |
1230 -------------------------
1232 where neither the high (valops.c et.al.) or low gdb (*-tdep.c) are
1233 really clear on what mechanisms they should be using to manipulate that
1234 buffer. Further, much code assumes, dangerously, that registers are
1235 contigious. Having got mips-tdep.c to support multiple ABIs, believe
1236 me, that is a bad assumption. Finally, that register cache layout is
1237 determined by the current remote/local target and _not_ the less
1238 specific target ISA. In fact, in many cases it is determined by the
1239 somewhat arbitrary layout of the [gG] packets!
1242 How I would like the register file to work is more like:
1270 The main objectives being:
1272 o a clear separation between the low
1273 level target and the high level GDB
1275 o a mechanism that solves the general
1276 problem of register aliases, overlaps
1277 etc instead of treating them as optional
1278 extras that can be wedged in as an after
1279 thought (that is a reasonable description
1280 of the current code).
1282 Identify then solve the hard case and the
1283 rest just falls out. GDB solved the easy
1284 case and then tried to ignore the real
1287 o a removal of the assumption that the
1288 mapping between the register cache
1289 and virtual registers is largely static.
1290 If you flip the USR/SSR stack register
1291 select bit in the status-register then
1292 the corresponding stack registers should
1295 o a mechanism that clearly separates the
1296 gdb internal register cache from any
1297 target (not architecture) dependent
1298 specifics such as [gG] packets.
1300 Of course, like anything, it sounds good in theory. In reality, it
1301 would have to contend with many<->many relationships at both the
1302 virt<->cache and cache<->target level. For instance:
1305 Modifying an mmx register may involve
1306 scattering values across both FP and
1307 mmpx specific parts of a buffer
1310 When writing back a SP it may need to
1311 both be written to both SP and USP.
1316 Rather than let this like the last time it was discussed, just slip, I'm
1317 first going to add this e-mail (+ references) to TODO. I'd then like to
1318 sketch out a broad strategy I think could get us there.
1321 First thing I'd suggest is separating out the ``extern registers[]''
1322 code so that we can at least identify what is using it. At present
1323 things are scattered across many files. That way we can at least
1324 pretend that there is a cache instead of a global array :-)
1326 I'd then suggest someone putting up a proposal for the pseudo-reg /
1327 high-level side interface so that code can be adopted to it. For old
1328 code, initially a blanket rename of write_register_bytes() to
1329 deprecated_write_register_bytes() would help.
1331 Following that would, finaly be the corresponding changes to the target.
1335 Check that GDB can handle all BFD architectures (Andrew Cagney)
1337 There should be a test that checks that BFD/GDB are in sync with
1338 regard to architecture changes. Something like a test that first
1339 queries GDB for all supported architectures and then feeds each back
1340 to GDB.. Anyone interested in learning how to write tests? :-)
1344 Architectural Change: Multi-arch et al.
1345 =======================================
1347 The long term objective is to remove all assumptions that there is a
1348 single target with a single address space with a single instruction
1349 set architecture and single application binary interface.
1351 This is an ongoing effort. The first milestone is to enable
1352 ``multi-arch'' where by all architectural decisions are made at
1355 It should be noted that ``gdbarch'' is really ``gdbabi'' and
1356 ``gdbisa''. Once things are multi-arched breaking that down correctly
1357 will become much easier.
1361 GDBARCH cleanup (Andrew Cagney)
1363 The non-generated parts of gdbarch.{sh,h,c} should be separated out
1364 into arch-utils.[hc].
1366 Document that gdbarch_init_ftype could easily fail because it didn't
1367 identify an architecture.
1371 Fix BELIEVE_PPC_PROMOTION. Change it to BELIEVE_PPC_PROMOTION_P?
1373 At present there is still #ifdef BELIEVE_PPC_PROMOTION code in the
1378 Fix target_signal_from_host() etc.
1380 The name is wrong for starters. ``target_signal'' should probably be
1381 ``gdb_signal''. ``from_host'' should be ``from_target_signal''.
1382 After that it needs to be multi-arched and made independent of any
1383 host signal numbering.
1387 Update ALPHA so that it uses ``struct frame_extra_info'' instead of
1390 This is a barrier to replacing mips_extra_func_info with something
1391 that works with multi-arch.
1395 Multi-arch mips_extra_func_info.
1397 This first needs the alpha to be updated so that it uses ``struct
1402 Rationalize TARGET_SINGLE_FORMAT and TARGET_SINGLE_BIT et al.
1404 Surely one of them is redundant.
1408 Convert ALL architectures to MULTI-ARCH.
1412 Select the initial multi-arch ISA / ABI based on --target or similar.
1414 At present the default is based on what ever is first in the BFD
1415 archures table. It should be determined based on the ``--target=...''
1420 Make MIPS pure multi-arch.
1422 It is only at the multi-arch enabled stage.
1428 Enable the code to recognize --enable-targets=.... like BINUTILS does.
1430 Can the tm.h and nm.h files be eliminated by multi-arch.
1434 Architectural Change: MI, LIBGDB and scripting languages
1435 ========================================================
1437 See also architectural changes related to the event loop. LIBGDB
1438 can't be finished until there is a generic event loop being used by
1441 The long term objective is it to be possible to integrate GDB into
1442 scripting languages.
1446 Implement generic ``(gdb) commmand > file''
1448 Once everything is going through ui_file it should be come fairly
1451 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00104.html
1455 Replace gdb_stdtarg with gdb_targout (and possibly gdb_targerr).
1457 gdb_stdtarg is easily confused with gdb_stdarg.
1461 Extra ui_file methods - dump.
1463 Very useful for whitebox testing.
1467 Eliminate error_begin().
1469 With ui_file, there is no need for the statefull error_begin ()
1474 Send normal output to gdb_stdout.
1475 Send error messages to gdb_stderror.
1476 Send debug and log output log gdb_stdlog.
1478 GDB still contains many cases where (f)printf or printf_filtered () is
1479 used when it should be sending the messages to gdb_stderror or
1480 gdb_stdlog. The thought of #defining printf to something has crossed
1485 Re-do GDB's output pager.
1487 GDB's output pager still relies on people correctly using *_filtered
1488 for gdb_stdout and *_unfiltered for gdb_stdlog / gdb_stderr.
1489 Hopefully, with all normal output going to gdb_stdout, the pager can
1490 just look at the ui_file that the output is on and then use that to
1491 decide what to do about paging. Sounds good in theory.
1495 Check/cleanup MI documentation.
1497 The list of commands specified in the documentation needs to be
1498 checked against the mi-cmds.c table in a mechanical way (so that they
1499 two can be kept up-to-date).
1503 Convert MI into libgdb
1505 MI provides a text interface into what should be many of the libgdb
1506 functions. The implementation of those functions should be separated
1507 into the MI interface and the functions proper. Those functions being
1508 moved to gdb/lib say.
1514 The first part can already be found in defs.h.
1518 MI's input does not use buffering.
1520 At present the MI interface reads raw characters of from an unbuffered
1521 FD. This is to avoid several nasty buffer/race conditions. That code
1522 should be changed so that it registers its self with the event loop
1523 (on the input FD) and then push commands up to MI as they arrive.
1525 The serial code already does this.
1529 Make MI interface accessible from existing CLI.
1533 Add a breakpoint-edit command to MI.
1535 It would be similar to MI's breakpoint create but would apply to an
1536 existing breakpoint. It saves the need to delete/create breakpoints
1537 when ever they are changed.
1541 Add directory path to MI breakpoint.
1543 That way the GUI's task of finding the file within which the
1544 breakpoint was set is simplified.
1548 Add a mechanism to reject certain expression classes to MI
1550 There are situtations where you don't want GDB's expression
1551 parser/evaluator to perform inferior function calls or variable
1552 assignments. A way of restricting the expression parser so that such
1553 operations are not accepted would be very helpful.
1557 Remove sideffects from libgdb breakpoint create function.
1559 The user can use the CLI to create a breakpoint with partial
1560 information - no file (gdb would use the file from the last
1563 The libgdb interface currently affects that environment which can lead
1564 to confusion when a user is setting breakpoints via both the MI and
1567 This is also a good example of how getting the CLI ``right'' will be
1572 Move gdb_lasterr to ui_out?
1574 The way GDB throws errors and records them needs a re-think. ui_out
1575 handles the correct output well. It doesn't resolve what to do with
1576 output / error-messages when things go wrong.
1580 do_setshow_command contains a 1024 byte buffer.
1582 The function assumes that there will never be any more than 1024 bytes
1583 of enum. It should use mem_file.
1587 Should struct cmd_list_element . completer take the command as an
1592 Should the bulk of top.c:line_completion_function() be moved to
1593 command.[hc]? complete_on_cmdlist() and complete_on_enums() could
1594 then be made private.
1598 top.c (execute_command): Should a command being valid when the target
1599 is running be made an attribute (predicate) to the command rather than
1600 an explicit set of tests.
1604 top.c (execute_command): Should the bulk of this function be moved
1605 into command.[hc] so that top.c doesn't grub around in the command
1610 Architectural Change: Async
1611 ===========================
1613 While GDB uses an event loop when prompting the user for input. That
1614 event loop is not exploited by targets when they allow the target
1615 program to continue. Typically targets still block in (target_wait())
1616 until the program again halts.
1618 The closest a target comes to supporting full asynchronous mode are
1619 the remote targets ``async'' and ``extended-async''.
1623 Asynchronous expression evaluator
1625 Inferior function calls hang GDB.
1629 Fix implementation of ``target xxx''.
1631 At present when the user specifies ``target xxxx'', the CLI maps that
1632 directly onto a target open method. It is then assumed that the
1633 target open method should do all sorts of complicated things as this
1634 is the only chance it has. Check how the various remote targets
1635 duplicate the target operations. Check also how the various targets
1636 behave differently for purely arbitrary reasons.
1638 What should happen is that ``target xxxx'' should call a generic
1639 ``target'' function and that should then co-ordinate the opening of
1640 ``xxxx''. This becomes especially important when you're trying to
1641 open an asynchronous target that may need to perform background tasks
1642 as part of the ``attach'' phase.
1644 Unfortunately, due to limitations in the old/creaking command.h
1645 interface, that isn't possible. The function being called isn't told
1646 of the ``xxx'' or any other context information.
1648 Consequently a precursor to fixing ``target xxxx'' is to clean up the
1649 CLI code so that it passes to the callback function (attatched to a
1650 command) useful information such as the actual command and a context
1651 for that command. Other changes such as making ``struct command''
1652 opaque may also help.
1655 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
1659 Make "target xxx" command interruptible.
1661 As things become async this becomes possible. A target would start
1662 the connect and then return control to the event loop. A cntrl-c
1663 would notify the target that the operation is to be abandoned and the
1664 target code could respond.
1668 Add a "suspend" subcommand of the "continue" command to suspend gdb
1669 while continuing execution of the subprocess. Useful when you are
1670 debugging servers and you want to dodge out and initiate a connection
1671 to a server running under gdb.
1680 Frequently requested but not approved requests.
1684 Eliminate unused argument warnings using ATTRIBUTE_UNUSED.
1686 The benefits on this one are thought to be marginal - GDBs design
1687 means that unused parameters are very common. GCC 3.0 will also
1688 include the option -Wno-unused-parameter which means that ``-Wall
1689 -Wno-unused-parameters -Werror'' can be specified.
1698 This list is not up to date, and opinions vary about the importance or
1699 even desirability of some of the items. If you do fix something, it
1700 always pays to check the below.
1704 @c This does not work (yet if ever). FIXME.
1705 @c @item --parse=@var{lang} @dots{}
1706 @c Configure the @value{GDBN} expression parser to parse the listed languages.
1707 @c @samp{all} configures @value{GDBN} for all supported languages. To get a
1708 @c list of all supported languages, omit the argument. Without this
1709 @c option, @value{GDBN} is configured to parse all supported languages.
1713 START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that
1714 is its default value. Clean this up.
1718 It should be possible to use symbols from shared libraries before we know
1719 exactly where the libraries will be loaded. E.g. "b perror" before running
1720 the program. This could maybe be done as an extension of the "breakpoint
1721 re-evaluation" after new symbols are loaded.
1725 Make single_step() insert and remove breakpoints in one operation.
1727 [If this is talking about having single_step() insert the breakpoints,
1728 run the target then pull the breakpoints then it is wrong. The
1729 function has to return as control has to eventually be passed back to
1730 the main event loop.]
1734 Speed up single stepping by avoiding extraneous ptrace calls.
1738 Speed up single stepping by not inserting and removing breakpoints
1739 each time the inferior starts and stops.
1741 Breakpoints should not be inserted and deleted all the time. Only the
1742 one(s) there should be removed when we have to step over one. Support
1743 breakpoints that don't have to be removed to step over them.
1745 [this has resulted in numerous debates. The issue isn't clear cut]
1749 Provide "voodoo" debugging of core files. This creates a zombie
1750 process as a child of the debugger, and loads it up with the data,
1751 stack, and regs of the core file. This allows you to call functions
1752 in the executable, to manipulate the data in the core file.
1758 GDB reopens the source file on every line, as you "next" through it.
1760 [still true? I've a memory of this being fixed]
1764 Perhaps "i source" should take an argument like that of "list".
1768 Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if
1769 it matches the source line indicated.
1773 The prompt at end of screen should accept space as well as CR.
1777 Backtrace should point out what the currently selected frame is, in
1778 its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar,
1779 ...)" rather than "#3 foo (bar, ...)".
1783 "i program" should work for core files, and display more info, like what
1784 actually caused it to die.
1788 "x/10i" should shorten the long name, if any, on subsequent lines.
1792 "next" over a function that longjumps, never stops until next time you happen
1793 to get to that spot by accident. E.g. "n" over execute_command which has
1798 "set zeroprint off", don't bother printing members of structs which
1799 are entirely zero. Useful for those big structs with few useful
1804 GDB does four ioctl's for every command, probably switching terminal modes
1805 to/from inferior or for readline or something.
1809 terminal_ours versus terminal_inferior: cache state. Switch should be a noop
1810 if the state is the same, too.
1814 "i frame" shows wrong "arglist at" location, doesn't show where the args
1815 should be found, only their actual values.
1819 There should be a way for "set" commands to validate the new setting
1820 before it takes effect.
1824 "ena d" is ambiguous, why? "ena delete" seems to think it is a command!
1828 i line VAR produces "Line number not known for symbol ``var''.". I
1829 thought we were stashing that info now!
1833 We should be able to write to random files at hex offsets like adb.
1837 [elena - delete this]
1839 Handle add_file with separate text, data, and bss addresses. Maybe
1840 handle separate addresses for each segment in the object file?
1844 [Jimb/Elena delete this one]
1846 Handle free_named_symtab to cope with multiply-loaded object files
1847 in a dynamic linking environment. Should remember the last copy loaded,
1848 but not get too snowed if it finds references to the older copy.
1852 [elena delete this also]
1854 Remove all references to:
1861 now that we have BFD. All remaining are in machine dependent files.
1865 Re-organize help categories into things that tend to fit on a screen
1870 Add in commands like ADB's for searching for patterns, etc. We should
1871 be able to examine and patch raw unsymboled binaries as well in gdb as
1872 we can in adb. (E.g. increase the timeout in /bin/login without source).
1874 [actually, add ADB interface :-]
1878 When doing "step" or "next", if a few lines of source are skipped between
1879 the previous line and the current one, print those lines, not just the
1880 last line of a multiline statement.
1884 Handling of "&" address-of operator needs some serious overhaul
1885 for ANSI C and consistency on arrays and functions.
1886 For "float point[15];":
1887 ptype &point[4] ==> Attempt to take address of non-lvalue.
1888 For "char *malloc();":
1889 ptype malloc ==> "char *()"; should be same as
1890 ptype &malloc ==> "char *(*)()"
1891 call printf ("%x\n", malloc) ==> weird value, should be same as
1892 call printf ("%x\n", &malloc) ==> correct value
1896 Fix dbxread.c symbol reading in the presence of interrupts. It
1897 currently leaves a cleanup to blow away the entire symbol table when a
1898 QUIT occurs. (What's wrong with that? -kingdon, 28 Oct 1993).
1900 [I suspect that the grype was that, on a slow system, you might want
1901 to cntrl-c and get just half the symbols and then load the rest later
1902 - scary to be honest]
1906 Mipsread.c reads include files depth-first, because the dependencies
1907 in the psymtabs are way too inclusive (it seems to me). Figure out what
1908 really depends on what, to avoid recursing 20 or 30 times while reading
1913 value_add() should be subtracting the lower bound of arrays, if known,
1914 and possibly checking against the upper bound for error reporting.
1918 When listing source lines, check for a preceding \n, to verify that
1919 the file hasn't changed out from under us.
1921 [fixed by some other means I think. That hack wouldn't actually work
1922 reliably - the file might move such that another \n appears. ]
1926 Get all the remote systems (where the protocol allows it) to be able to
1927 stop the remote system when the GDB user types ^C (like remote.c
1928 does). For ebmon, use ^Ak.
1932 Possible feature: A version of the "disassemble" command which shows
1933 both source and assembly code ("set symbol-filename on" is a partial
1936 [has this been done? It was certainly done for MI and GDBtk]
1940 investigate "x/s 0" (right now stops early) (I think maybe GDB is
1941 using a 0 address for bad purposes internally).
1945 Make "info path" and path_command work again (but independent of the
1946 environment either of gdb or that we'll pass to the inferior).
1950 Make GDB understand the GCC feature for putting octal constants in
1951 enums. Make it so overflow on an enum constant does not error_type
1952 the whole type. Allow arbitrarily large enums with type attributes.
1953 Put all this stuff in the testsuite.
1957 Make TYPE_CODE_ERROR with a non-zero TYPE_LENGTH more useful (print
1958 the value in hex; process type attributes). Add this to the
1959 testsuite. This way future compilers can add new types and old
1960 versions of GDB can do something halfway reasonable.
1964 Fix mdebugread.c:parse_type to do fundamental types right (see
1965 rs6000_builtin_type in stabsread.c for what "right" is--the point is
1966 that the debug format fixes the sizes of these things and it shouldn't
1967 depend on stuff like TARGET_PTR_BIT and so on. For mdebug, there seem
1968 to be separate bt* codes for 64 bit and 32 bit things, and GDB should
1969 be aware of that). Also use a switch statement for clarity and speed.
1973 Investigate adding symbols in target_load--some targets do, some
1978 Put dirname in psymtabs and change lookup*symtab to use dirname (so
1979 /foo/bar.c works whether compiled by cc /foo/bar.c, or cd /foo; cc
1984 Merge xcoffread.c and coffread.c. Use breakpoint_re_set instead of
1989 Make a watchpoint which contains a function call an error (it is
1990 broken now, making it work is probably not worth the effort).
1994 New test case based on weird.exp but in which type numbers are not
1995 renumbered (thus multiply defining a type). This currently causes an
1996 infinite loop on "p v_comb".
2000 [Hey! Hint Hint Delete Delete!!!]
2002 Fix 386 floating point so that floating point registers are real
2003 registers (but code can deal at run-time if they are missing, like
2004 mips and 68k). This would clean up "info float" and related stuff.
2008 gcc -g -c enummask.c then gdb enummask.o, then "p v". GDB complains
2009 about not being able to access memory location 0.
2011 -------------------- enummask.c
2032 If try to modify value in file with "set write off" should give
2033 appropriate error not "cannot access memory at address 0x65e0".
2037 Allow core file without exec file on RS/6000.
2041 Make sure "shell" with no arguments works right on DOS.
2045 Make gdb.ini (as well as .gdbinit) be checked on all platforms, so
2046 the same directory can be NFS-mounted on unix or DOS, and work the
2051 [Is this another delete???]
2053 Get SECT_OFF_TEXT stuff out of objfile_relocate (might be needed to
2054 get RS/6000 to work right, might not be immediately relevant).
2058 Work out some kind of way to allow running the inferior to be done as
2059 a sub-execution of, eg. breakpoint command lists. Currently running
2060 the inferior interupts any command list execution. This would require
2061 some rewriting of wait_for_inferior & friends, and hence should
2062 probably be done in concert with the above.
2066 Add function arguments to gdb user defined functions.
2070 Add convenience variables that refer to exec file, symbol file,
2071 selected frame source file, selected frame function, selected frame
2076 Modify the handling of symbols grouped through BINCL/EINCL stabs to
2077 allocate a partial symtab for each BINCL/EINCL grouping. This will
2078 seriously decrease the size of inter-psymtab dependencies and hence
2079 lessen the amount that needs to be read in when a new source file is
2084 Add a command for searching memory, a la adb. It specifies size,
2085 mask, value, start address. ADB searches until it finds it or hits
2086 an error (or is interrupted).
2090 Remove the range and type checking code and documentation, if not