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)
167 Obsolete some targets.
169 Possible selection criteria are:
171 o uses a deprecated feature
175 o doesn't have a maintainer
179 o post proposals to gdb@ (DONE)
181 o post announcement to gdb-announce@
182 crossed with gdb@ reply-to to gdb@
199 configure.{in,host,tgt}
201 *-tdep.c *-nat.c *-xdep.c
208 Change documentation to GFDL license.
210 ``It is time to make an effort to start using the GFDL more
211 thoroughly. Would all GNU maintainers please change the license to
212 the GFDL, for all manuals and other major documentation files?
214 The GFDL and some instructions for using it can be found in
215 http://www.gnu.org/copyleft/''
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
229 GDB 5.1 - Known Problems
230 ========================
236 The z8k has suffered bit rot and is known to not build. The problem
237 was occuring in the opcodes directory.
241 Solaris 8 x86 CURSES_H problem
242 http://sources.redhat.com/ml/gdb/2000-07/msg00038.html
244 The original problem was worked around with:
246 2000-06-06 Michael Snyder <msnyder@cygnus.com>
248 * configure.in: Enable autoconf to find curses.h on Solaris 2.8.
249 * configure: Regenerate.
251 When building both GDB and SID using the same source tree the problem
252 will still occure. sid/component/configure.in mis-configures
253 <curses.h> and leaves wrong information in the config cache.
262 Thread support. Right now, as soon as a thread finishes and exits,
263 you're hosed. This problem is reported once a week or so.
267 GDB 5.2 - New features
268 ======================
272 GCC 3.0 ABI support (but hopefully sooner...).
276 Objective C/C++ support (but hopefully sooner...).
283 The following cleanups have been identified as part of GDB 5.2.
287 Remove old code that does not use ui_out functions and all the related
288 "ifdef"s. This also allows the elimination of -DUI_OUT from
289 Makefile.in and configure.in.
295 Eliminate all warnings for at least one host/target for the flags:
296 -Wimplicit -Wreturn-type -Wcomment -Wtrigraphs -Wformat -Wparentheses
297 -Wpointer-arith -Wuninitialized
301 Restructure gdb directory tree so that it avoids any 8.3 and 14
306 Convert GDB build process to AUTOMAKE.
308 See also sub-directory configure below.
310 The current convention is (kind of) to use $(<header>_h) in all
311 dependency lists. It isn't done in a consistent way.
315 GDB 5.2 - Known Problems
316 ========================
320 Code Cleanups: General
321 ======================
323 The following are more general cleanups and fixes. They are not tied
324 to any specific release.
328 Revised UDP support (was: Re: [Fwd: [patch] UDP transport support])
329 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00000.html
331 (Broken) support for GDB's remote protocol across UDP is to be
332 included in the follow-on release.
334 It should be noted that UDP can only work when the [Gg] packet fits in
337 There is also much debate over the merit of this.
341 Can't build IRIX -> arm GDB.
342 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00356.html
345 > Now I'm building for an embedded arm target. If there is a way of turning
346 > remote-rdi off, I couldn't find it. It looks like it gets built by default
347 > in gdb/configure.tgt(line 58) Anyway, the build dies in
348 > gdb/rdi-share/unixcomm.c. SERPORT1 et. al. never get defined because we
349 > aren't one of the architectures supported.
353 Problem with weak functions
354 http://sourceware.cygnus.com/ml/gdb/2000-05/msg00060.html
356 Dan Nicolaescu writes:
357 > It seems that gdb-4.95.1 does not display correctly the function when
358 > stoping in weak functions.
360 > It stops in a function that is defined as weak, not in the function
361 > that is actually run...
365 Follow through `make check' with --enable-shared.
367 When the srcware tree is configured with --enable-shared, the `expect'
368 program won't run properly. Jim Wilson found out gdb has a local hack
369 to set LD_LIBRARY_PATH, but, AFAIK, no other project has been hacked
372 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00845.html
376 Delete macro TARGET_BYTE_ORDER_SELECTABLE.
378 Patches in the database.
382 printcmd.c (print_address_numeric):
384 NOTE: This assumes that the significant address information is kept in
385 the least significant bits of ADDR - the upper bits were either zero
386 or sign extended. Should ADDRESS_TO_POINTER() or some
387 ADDRESS_TO_PRINTABLE() be used to do the conversion?
391 The BFD directory requires bug-fixed AUTOMAKE et.al.
393 AUTOMAKE 1.4 incorrectly set the TEXINPUTS environment variable. It
394 contained the full path to texinfo.tex when it should have only
395 contained the directory. The bug has been fixed in the current
396 AUTOMAKE sources. Automake snapshots can be found in:
397 ftp://sourceware.cygnus.com/pub/gdb/snapshots
398 and ftp://sourceware.cygnus.com/pub/binutils
402 Find something better than DEFAULT_BFD_ARCH, DEFAULT_BFD_VEC to
403 determine the default isa/byte-order.
407 Rely on BFD_BIG_ENDIAN and BFD_LITTLE_ENDIAN instead of host dependent
408 BIG_ENDIAN and LITTLE_ENDIAN.
412 Eliminate more compiler warnings.
414 Of course there also needs to be the usual debate over which warnings
415 are valid and how to best go about this.
417 One method: choose a single option; get agreement that it is
418 reasonable; try it out to see if there isn't anything silly about it
419 (-Wunused-parameters is an example of that) then incrementally hack
422 The other method is to enable all warnings and eliminate them from one
427 Elimination of ``(catch_errors_ftype *) func''.
429 Like make_cleanup_func it isn't portable.
430 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00791.html
431 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00814.html
435 Nuke #define CONST_PTR.
443 [PATCH/5] src/intl/Makefile.in:distclean additions
444 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00363.html
446 Do not forget to merge the patch back into the trunk.
450 Rationalize the host-endian code (grep for HOST_BYTE_ORDER).
452 At present defs.h includes <endian.h> (which is linux specific) yet
453 almost nothing depends on it. Suggest "gdb_endian.h" which can also
454 handle <machine/endian.h> and only include that where it is really
459 Replace savestring() with something from libiberty.
461 An xstrldup()? but that would have different semantics.
465 Rationalize use of floatformat_unknown in GDB sources.
467 Instead of defaulting to floatformat_unknown, should hosts/targets
468 specify the value explicitly?
470 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
474 Add a ``name'' member to include/floatformat.h:struct floatformat.
475 Print that name in gdbarch.c.
479 Sort out the harris mess in include/floatformat.h (it hardwires two
480 different floating point formats).
484 See of the GDB local floatformat_do_doublest() and libiberty's
485 floatformat_to_double (which was once GDB's ...) can be merged some
490 Eliminate mmalloc(), mstrsave() et.al. from GDB.
492 Also eliminate it from defs.h.
496 Eliminate PTR. ISO-C allows ``void *''.
502 GDB should never abort. GDB should either throw ``error ()'' or
503 ``internal_error ()''. Better still GDB should naturally unwind with
508 GDB probably doesn't build on FreeBSD pre 2.2.x
509 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00378.html
511 Fixes to get FreeBSD working on 2.2.x, 3.x and 4.x caused the code to
516 Deprecate "fg". Apparently ``fg'' is actually continue.
518 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00417.html
522 Deprecate current use of ``floatformat_unknown''.
524 Require all targets to explicitly provide their float format instead
525 of defaulting to floatformat unknown. Doing the latter leads to nasty
528 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
532 Rationalize floatformat_to_double() vs floatformat_to_doublest().
534 Looks like GDB migrated floatformat_to_double() to libiberty but then
535 turned around and created a ..._to_doublest() the latter containing
538 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00472.html
542 Move floatformat_ia64_ext to libiberty/include floatformat.[ch].
544 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00466.html
548 The ``maintenance deprecate set endian big'' command doesn't notice
549 that it is deprecating ``set endian'' and not ``set endian big'' (big
550 is implemented using an enum). Is anyone going to notice this?
554 When tab expanding something like ``set arch<tab>'' ignore the
555 deprecated ``set archdebug'' and expand to ``set architecture''.
559 Eliminate ``arm_register_names[j] = (char *) regnames[j]'' and the
560 like from arm-tdep.c.
564 Fix uses of ->function.cfunc = set_function().
566 The command.c code calls sfunc() when a set command. Rather than
567 change it suggest fixing the callback function so that it is more
570 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
572 See also ``Fix implementation of ``target xxx''.'' below.
576 IRIX 3.x support is probably broken.
580 Delete sim/SIM_HAVE_BREAKPOINTS and gdb/SIM_HAS_BREAKPOINTS.
581 http://sourceware.cygnus.com/ml/gdb-patches/2000-07/msg00042.html
583 Apart from the d30v, are there any sim/common simulators that make use
586 A brief summary of what happened is that sim/common/sim-break.c was
587 created as a good idea. It turned out a better idea was to use
588 SIM_SIGBREAK and have GDB pass back sim_resume (..., SIGBREAK).
592 Move remote_remove_hw_breakpoint, remote_insert_hw_breakpoint,
593 remote_remove_watchpoint, remote_insert_watchpoint into target vector.
597 Eliminate ``extern'' from C files.
601 Replace ``STREQ()'' et.al. with ``strcmp() == 0'' et.al.
603 Extreme care is recommeded - perhaps only modify tests that are
604 exercised by the testsuite (as determined using some type of code
609 Replace the file gdb/CONTRIBUTE with a file that is generated from the
610 gdb/doc/*.texinfo directory.
614 New Features and Fixes
615 ======================
617 These are harder than cleanups but easier than work involving
618 fundamental architectural change.
622 Hardware watchpoint problems on x86 OSes, including Linux:
624 1. Delete/disable hardware watchpoints should free hardware debug
626 2. Watch for different values on a viariable with one hardware debug
629 According to Eli Zaretskii <eliz@delorie.com>:
631 These are not GDB/ia32 issues per se: the above features are all
632 implemented in the DJGPP port of GDB and work in v5.0. Every
633 x86-based target should be able to lift the relevant parts of
634 go32-nat.c and use them almost verbatim. You get debug register
635 sharing through reference counts, and the ability to watch large
636 regions (up to 16 bytes) using multiple registers. (The required
637 infrastructure in high-level GDB application code, mostly in
638 breakpoint.c, is also working since v5.0.)
642 Add built-by, build-date, tm, xm, nm and anything else into gdb binary
643 so that you can see how the GDB was created.
647 Add an "info bfd" command that displays supported object formats,
648 similarly to objdump -i.
650 Is there a command already?
654 Fix ``I'm sorry, Dave, I can't do that.'' from symfile.c.
656 This requires internationalization.
662 (gdb) p fwprintf(stdout,L"%S\n", f)
663 No symbol "L" in current context.
667 Cleanup configury support for optional sub-directories.
669 Check how GCC handles multiple front ends for an example of how things
670 could work. A tentative first step is to rationalize things so that
671 all sub directories are handled in a fashion similar to gdb/mi.
673 See also automake above.
677 Add a transcript mechanism to GDB.
679 Such a mechanism might log all gdb input and output to a file in a
680 form that would allow it to be replayed. It could involve ``gdb
681 --transcript=FILE'' or it could involve ``(gdb) transcript file''.
685 Can the xdep files be replaced by autoconf?
689 Document trace machinery
693 Document ui-out and ui-file.
695 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00121.html
699 Update texinfo.tex to latest?
703 Incorporate agentexpr.texi into gdb.texinfo
705 agentexpr.texi mostly describes the details of the byte code used for
706 tracepoints, not the internals of the support for this in GDB. So it
707 looks like gdb.texinfo is a better place for this information.
709 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00566.html
713 Document overlay machinery.
717 ``(gdb) catch signal SIGNAL''
719 Overlaps with ``handle SIGNAL'' but the implied behavior is different.
720 You can attach commands to a catch but not a handle. A handle has a
721 limited number of hardwired actions.
727 o readline/*.h bitrot
729 The TUI isn't up-to-date with
730 respect to the readline currently
731 bundled with GDB. Importing a
732 new readline is on the 5.1 wish
733 list so this can only get worse.
735 Grep for things like term_cursor_move.
737 (To be honest, I don't see anyone
738 importing a new readline before 5.1 is
741 o tui.c:va_catch_errors() bitrot
743 This nasty piece of work used knowledge
744 of the internals of GDBs error functions :-(
745 Ever since those internals were cleaned
746 up this code has been broken. :-(
748 o tuiWin.c:c_makeVisibleWithNewHeight() broken
749 tuiLayout.c:_extractDisplayStartAddr() broken
751 Both these function call find_line_pc()
752 incorrectly (wrong args, wrong return value).
754 I suspect this bug has always been there!
755 It had been hidden because those files
756 didn't include the necessary header files
759 o tuiRegs() host dependant
761 Not suprisingly, this isn't a very portable
762 section of code. However, I'm sure people
763 could live with no regs in the short to
766 o defs.h: #include "tui.h" et.al.
768 I'm not sure where this came from.
769 It was a really bad idea.
771 To get things to compile I did a nasty
772 hack (Just declare what was needed and
773 replace any expressions like xx->y.z()
774 in GDB proper with function calls). I
775 could commit it slightly cleaned up if
778 Medium Term. the #ifdef TUI and TuiDo()
779 should be changed to hooks (like GDBTK).
780 The gdb-events.[hc] is there for that
783 o tui.c:_tuiReset() host dependant
785 tui.c contains a lump of termio[s]
786 I suspect an equivalent block of
787 code can be lifted from readline.
788 An equivalent readline function may
791 o curses.h vs ncurses.h.
793 Simple portability problem.
797 This function is a mystery - where is it?
799 o tui-file.[hc] cleanup
801 This can be significantly simplified.
803 o The code should be pacified. (-Werror -W...)
805 There are plenty of #includes,
806 duplicate #includes, missing function decls
809 Some of the problems I found were through
810 fixing a few of the warnings.
812 o The code should be GNUtified.
814 It would be very nice to have this code
815 look like the rest of GDB. That way people
816 would be more accepting of it as a true
819 Until it is GNUtified it is going to stick
820 out like a sore thumb to the programmer.
822 o The code should be clearly copyrighted
824 (FSF, with due credit to HP)
828 Add support for ``gdb --- PROGRAM ARGS ...''.
829 Add support for ``gdb -cmd=...''
831 Along with many variations. Check:
833 ????? for a full discussion.
839 Implement ``(gdb) !ls''.
841 Which is very different from ``(gdb) ! ls''. Implementing the latter
844 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00034.html
848 Change the (char *list[]) to (const char (*)[]) so that dynamic lists can
853 When tab expanding something like ``set arch<tab>'' ignore the
854 deprecated ``set archdebug'' and expand to ``set architecture''.
858 Replace the code that uses the host FPU with an emulator of the target
863 The "ocd reset" command needs to flush the dcache, which requires breaking
864 the abstraction layer between the target independent and target code. One
865 way to address this is provide a generic "reset" command and target vector.
867 http://sources.redhat.com/ml/gdb-patches/2000-10/msg00011.html
876 Generic: lin-thread cannot handle thread exit (Mark Kettenis, Michael
877 Snyder) http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00525.html
879 The thread_db assisted debugging code doesn't handle exiting threads
880 properly, at least in combination with glibc 2.1.3 (the framework is
881 there, just not the actual code). There are at least two problems
882 that prevent this from working.
884 As an additional reference point, the pre thread_db code did not work
889 GNU/Linux/x86 and random thread signals (and Solaris/SPARC but not
891 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00336.html
893 Christopher Blizzard writes:
895 So, I've done some more digging into this and it looks like Jim
896 Kingdon has reported this problem in the past:
898 http://sourceware.cygnus.com/ml/bug-gdb/1999-10/msg00058.html
900 I can reproduce this problem both with and without Tom's patch. Has
901 anyone seen this before? Maybe have a solution for it hanging around?
904 There's a test case for this documented at:
906 when debugging threaded applications you get extra SIGTRAPs
907 http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=9565
909 [There should be a GDB testcase - cagney]
913 GDB5 TOT on unixware 7
914 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00119.html
917 > I just spun the top of tree of the GDB5 branch on UnixWare 7. As a
918 > practical matter, the current thread support is somewhat more annoying
919 > than when GDB was thread-unaware.
923 Migrate qfThreadInfo packet -> qThreadInfo. (Andrew Cagney)
925 Add support for packet enable/disable commands with these thread
926 packets. General cleanup.
928 [PATCH] Document the ThreadInfo remote protocol queries
929 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00832.html
931 [PATCH] "info threads" queries for remote.c
932 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00831.html
939 New languages come onto the scene all the time.
943 Re: Various C++ things
945 value_headof/value_from_vtable_info are worthless, and should be
946 removed. The one place in printcmd.c that uses it should use the RTTI
949 RTTI for g++ should be using the typeinfo functions rather than the
950 vtables. The typeinfo functions are always at offset 4 from the
951 beginning of the vtable, and are always right. The vtables will have
952 weird names like E::VB sometimes. The typeinfo function will always
953 be "E type_info function", or somesuch.
955 value_virtual_fn_field needs to be fixed so there are no failures for
956 virtual functions for C++ using g++.
958 Testsuite cases are the major priority right now for C++ support,
959 since i have to make a lot of changes that could potentially break
964 Add support for Modula3
966 Get DEC/Compaq to contribute their Modula-3 support.
970 Remote Protocol Support
971 =======================
975 Remote protocol doco feedback.
977 Too much feedback to mention needs to be merged in (901660). Search
978 for the word ``remote''.
981 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00023.html
982 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00056.html
983 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00382.html
987 GDB doesn't recover gracefully from remote protocol errors.
989 GDB wasn't checking for NAKs from the remote target. Instead a NAK is
990 ignored and a timeout is required before GDB retries. A pre-cursor to
991 fixing this this is making GDB's remote protocol packet more robust.
993 While downloading to a remote protocol target, gdb ignores packet
994 errors in so far as it will continue to download with chunk N+1 even
995 if chunk N was not correctly sent. This causes gdb.base/remote.exp to
996 take a painfully long time to run. As a PS that test needs to be
997 fixed so that it builds on 16 bit machines.
1001 Add the cycle step command.
1003 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00237.html
1007 Resolve how to scale things to support very large packets.
1011 Resolve how to handle a target that changes things like its endianess
1012 on the fly - should it be returned in the ``T'' packet?
1014 Underlying problem is that the register file is target endian. If the
1015 target endianess changes gdb doesn't know.
1019 Rename read_register{,_pid}() to read_unsigned_register{,_pid}().
1026 If / when GDB starts to support the debugging of multi-processor
1027 (rather than multi-thread) applications the symtab code will need to
1028 be updated a little so that several independent symbol tables are
1029 active at a given time.
1031 The other interesting change is a clarification of the exact meaning
1032 of CORE_ADDR and that has had consequences for a few targets (that
1033 were abusing that data type).
1037 Investiagate ways of reducing memory.
1041 Investigate ways of improving load time.
1045 Get the d10v to use POINTER_TO_ADDRESS and ADDRESS_TO_POINTER.
1047 Consequence of recent symtab clarification. No marks for figuring out
1048 who maintains the d10v.
1052 Get the MIPS to correctly sign extend all address <-> pointer
1055 Consequence of recent symtab clarification. No marks for figuring out
1056 who maintains the MIPS.
1060 GDB truncates 64 bit enums.
1062 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00290.html
1069 There are never to many testcases.
1073 Better thread testsuite.
1077 Better C++ testsuite.
1081 Look at adding a GDB specific testsuite directory so that white box
1082 tests of key internals can be added (eg ui_file).
1086 Separate out tests that involve the floating point (FP).
1088 (Something for people brining up new targets). FP and non-fp tests
1089 are combined. I think there should be set of basic tests that
1090 exercise pure integer support and then a more expanded set that
1091 exercise FP and FP/integer interactions.
1093 As an example, the MIPS, for n32 as problems with passing FP's and
1094 structs. Since most inferior call tests include FP it is difficult to
1095 determine of the integer tests are ok.
1099 Architectural Changes: General
1100 ==============================
1102 These are harder than simple cleanups / fixes and, consequently
1103 involve more work. Typically an Architectural Change will be broken
1104 down into a more digestible set of cleanups and fixes.
1108 Cleanup software single step.
1110 At present many targets implement software single step by directly
1111 blatting memory (see rs6000-tdep.c). Those targets should register
1112 the applicable breakpoints using the breakpoint framework. Perhaphs a
1113 new internal breakpoint class ``step'' is needed.
1117 Replace READ_FP() with FRAME_HANDLE().
1119 READ_FP() is a hangover from the days of the vax when the ABI really
1120 did have a frame pointer register. Modern architectures typically
1121 construct a virtual frame-handle from the stack pointer and various
1122 other bits of string.
1124 Unfortunately GDB still treats this synthetic FP register as though it
1125 is real. That in turn really confuses users (arm and ``print $fp'' VS
1126 ``info registers fp''). The synthetic FP should be separated out of
1127 the true register set presented to the user.
1131 Register Cache Cleanup (below from Andrew Cagney)
1133 I would depict the current register architecture as something like:
1135 High GDB --> Low GDB
1140 register + REGISTER_BYTE(reg_nr)
1143 -------------------------
1144 | extern register[] |
1145 -------------------------
1147 where neither the high (valops.c et.al.) or low gdb (*-tdep.c) are
1148 really clear on what mechanisms they should be using to manipulate that
1149 buffer. Further, much code assumes, dangerously, that registers are
1150 contigious. Having got mips-tdep.c to support multiple ABIs, believe
1151 me, that is a bad assumption. Finally, that register cache layout is
1152 determined by the current remote/local target and _not_ the less
1153 specific target ISA. In fact, in many cases it is determined by the
1154 somewhat arbitrary layout of the [gG] packets!
1157 How I would like the register file to work is more like:
1185 The main objectives being:
1187 o a clear separation between the low
1188 level target and the high level GDB
1190 o a mechanism that solves the general
1191 problem of register aliases, overlaps
1192 etc instead of treating them as optional
1193 extras that can be wedged in as an after
1194 thought (that is a reasonable description
1195 of the current code).
1197 Identify then solve the hard case and the
1198 rest just falls out. GDB solved the easy
1199 case and then tried to ignore the real
1202 o a removal of the assumption that the
1203 mapping between the register cache
1204 and virtual registers is largely static.
1205 If you flip the USR/SSR stack register
1206 select bit in the status-register then
1207 the corresponding stack registers should
1210 o a mechanism that clearly separates the
1211 gdb internal register cache from any
1212 target (not architecture) dependent
1213 specifics such as [gG] packets.
1215 Of course, like anything, it sounds good in theory. In reality, it
1216 would have to contend with many<->many relationships at both the
1217 virt<->cache and cache<->target level. For instance:
1220 Modifying an mmx register may involve
1221 scattering values across both FP and
1222 mmpx specific parts of a buffer
1225 When writing back a SP it may need to
1226 both be written to both SP and USP.
1231 Rather than let this like the last time it was discussed, just slip, I'm
1232 first going to add this e-mail (+ references) to TODO. I'd then like to
1233 sketch out a broad strategy I think could get us there.
1236 First thing I'd suggest is separating out the ``extern registers[]''
1237 code so that we can at least identify what is using it. At present
1238 things are scattered across many files. That way we can at least
1239 pretend that there is a cache instead of a global array :-)
1241 I'd then suggest someone putting up a proposal for the pseudo-reg /
1242 high-level side interface so that code can be adopted to it. For old
1243 code, initially a blanket rename of write_register_bytes() to
1244 deprecated_write_register_bytes() would help.
1246 Following that would, finaly be the corresponding changes to the target.
1250 Check that GDB can handle all BFD architectures (Andrew Cagney)
1252 There should be a test that checks that BFD/GDB are in sync with
1253 regard to architecture changes. Something like a test that first
1254 queries GDB for all supported architectures and then feeds each back
1255 to GDB.. Anyone interested in learning how to write tests? :-)
1259 Architectural Change: Multi-arch et al.
1260 =======================================
1262 The long term objective is to remove all assumptions that there is a
1263 single target with a single address space with a single instruction
1264 set architecture and single application binary interface.
1266 This is an ongoing effort. The first milestone is to enable
1267 ``multi-arch'' where by all architectural decisions are made at
1270 It should be noted that ``gdbarch'' is really ``gdbabi'' and
1271 ``gdbisa''. Once things are multi-arched breaking that down correctly
1272 will become much easier.
1276 GDBARCH cleanup (Andrew Cagney)
1278 The non-generated parts of gdbarch.{sh,h,c} should be separated out
1279 into arch-utils.[hc].
1281 Document that gdbarch_init_ftype could easily fail because it didn't
1282 identify an architecture.
1286 Fix BELIEVE_PPC_PROMOTION. Change it to BELIEVE_PPC_PROMOTION_P?
1288 At present there is still #ifdef BELIEVE_PPC_PROMOTION code in the
1293 Fix target_signal_from_host() etc.
1295 The name is wrong for starters. ``target_signal'' should probably be
1296 ``gdb_signal''. ``from_host'' should be ``from_target_signal''.
1297 After that it needs to be multi-arched and made independent of any
1298 host signal numbering.
1302 Update ALPHA so that it uses ``struct frame_extra_info'' instead of
1305 This is a barrier to replacing mips_extra_func_info with something
1306 that works with multi-arch.
1310 Multi-arch mips_extra_func_info.
1312 This first needs the alpha to be updated so that it uses ``struct
1317 Rationalize TARGET_SINGLE_FORMAT and TARGET_SINGLE_BIT et al.
1319 Surely one of them is redundant.
1323 Convert ALL architectures to MULTI-ARCH.
1327 Select the initial multi-arch ISA / ABI based on --target or similar.
1329 At present the default is based on what ever is first in the BFD
1330 archures table. It should be determined based on the ``--target=...''
1335 Make MIPS pure multi-arch.
1337 It is only at the multi-arch enabled stage.
1343 Enable the code to recognize --enable-targets=.... like BINUTILS does.
1345 Can the tm.h and nm.h files be eliminated by multi-arch.
1349 Architectural Change: MI, LIBGDB and scripting languages
1350 ========================================================
1352 See also architectural changes related to the event loop. LIBGDB
1353 can't be finished until there is a generic event loop being used by
1356 The long term objective is it to be possible to integrate GDB into
1357 scripting languages.
1361 Implement generic ``(gdb) commmand > file''
1363 Once everything is going through ui_file it should be come fairly
1366 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00104.html
1370 Replace gdb_stdtarg with gdb_targout (and possibly gdb_targerr).
1372 gdb_stdtarg is easily confused with gdb_stdarg.
1376 Extra ui_file methods - dump.
1378 Very useful for whitebox testing.
1382 Eliminate error_begin().
1384 With ui_file, there is no need for the statefull error_begin ()
1389 Send normal output to gdb_stdout.
1390 Send error messages to gdb_stderror.
1391 Send debug and log output log gdb_stdlog.
1393 GDB still contains many cases where (f)printf or printf_filtered () is
1394 used when it should be sending the messages to gdb_stderror or
1395 gdb_stdlog. The thought of #defining printf to something has crossed
1400 Re-do GDB's output pager.
1402 GDB's output pager still relies on people correctly using *_filtered
1403 for gdb_stdout and *_unfiltered for gdb_stdlog / gdb_stderr.
1404 Hopefully, with all normal output going to gdb_stdout, the pager can
1405 just look at the ui_file that the output is on and then use that to
1406 decide what to do about paging. Sounds good in theory.
1410 Check/cleanup MI documentation.
1412 The list of commands specified in the documentation needs to be
1413 checked against the mi-cmds.c table in a mechanical way (so that they
1414 two can be kept up-to-date).
1418 Convert MI into libgdb
1420 MI provides a text interface into what should be many of the libgdb
1421 functions. The implementation of those functions should be separated
1422 into the MI interface and the functions proper. Those functions being
1423 moved to gdb/lib say.
1429 The first part can already be found in defs.h.
1433 MI's input does not use buffering.
1435 At present the MI interface reads raw characters of from an unbuffered
1436 FD. This is to avoid several nasty buffer/race conditions. That code
1437 should be changed so that it registers its self with the event loop
1438 (on the input FD) and then push commands up to MI as they arrive.
1440 The serial code already does this.
1444 Make MI interface accessible from existing CLI.
1448 Add a breakpoint-edit command to MI.
1450 It would be similar to MI's breakpoint create but would apply to an
1451 existing breakpoint. It saves the need to delete/create breakpoints
1452 when ever they are changed.
1456 Add directory path to MI breakpoint.
1458 That way the GUI's task of finding the file within which the
1459 breakpoint was set is simplified.
1463 Add a mechanism to reject certain expression classes to MI
1465 There are situtations where you don't want GDB's expression
1466 parser/evaluator to perform inferior function calls or variable
1467 assignments. A way of restricting the expression parser so that such
1468 operations are not accepted would be very helpful.
1472 Remove sideffects from libgdb breakpoint create function.
1474 The user can use the CLI to create a breakpoint with partial
1475 information - no file (gdb would use the file from the last
1478 The libgdb interface currently affects that environment which can lead
1479 to confusion when a user is setting breakpoints via both the MI and
1482 This is also a good example of how getting the CLI ``right'' will be
1487 Move gdb_lasterr to ui_out?
1489 The way GDB throws errors and records them needs a re-think. ui_out
1490 handles the correct output well. It doesn't resolve what to do with
1491 output / error-messages when things go wrong.
1495 do_setshow_command contains a 1024 byte buffer.
1497 The function assumes that there will never be any more than 1024 bytes
1498 of enum. It should use mem_file.
1502 Should struct cmd_list_element . completer take the command as an
1507 Should the bulk of top.c:line_completion_function() be moved to
1508 command.[hc]? complete_on_cmdlist() and complete_on_enums() could
1509 then be made private.
1513 top.c (execute_command): Should a command being valid when the target
1514 is running be made an attribute (predicate) to the command rather than
1515 an explicit set of tests.
1519 top.c (execute_command): Should the bulk of this function be moved
1520 into command.[hc] so that top.c doesn't grub around in the command
1525 Architectural Change: Async
1526 ===========================
1528 While GDB uses an event loop when prompting the user for input. That
1529 event loop is not exploited by targets when they allow the target
1530 program to continue. Typically targets still block in (target_wait())
1531 until the program again halts.
1533 The closest a target comes to supporting full asynchronous mode are
1534 the remote targets ``async'' and ``extended-async''.
1538 Asynchronous expression evaluator
1540 Inferior function calls hang GDB.
1544 Fix implementation of ``target xxx''.
1546 At present when the user specifies ``target xxxx'', the CLI maps that
1547 directly onto a target open method. It is then assumed that the
1548 target open method should do all sorts of complicated things as this
1549 is the only chance it has. Check how the various remote targets
1550 duplicate the target operations. Check also how the various targets
1551 behave differently for purely arbitrary reasons.
1553 What should happen is that ``target xxxx'' should call a generic
1554 ``target'' function and that should then co-ordinate the opening of
1555 ``xxxx''. This becomes especially important when you're trying to
1556 open an asynchronous target that may need to perform background tasks
1557 as part of the ``attach'' phase.
1559 Unfortunately, due to limitations in the old/creaking command.h
1560 interface, that isn't possible. The function being called isn't told
1561 of the ``xxx'' or any other context information.
1563 Consequently a precursor to fixing ``target xxxx'' is to clean up the
1564 CLI code so that it passes to the callback function (attatched to a
1565 command) useful information such as the actual command and a context
1566 for that command. Other changes such as making ``struct command''
1567 opaque may also help.
1570 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
1574 Make "target xxx" command interruptible.
1576 As things become async this becomes possible. A target would start
1577 the connect and then return control to the event loop. A cntrl-c
1578 would notify the target that the operation is to be abandoned and the
1579 target code could respond.
1583 Add a "suspend" subcommand of the "continue" command to suspend gdb
1584 while continuing execution of the subprocess. Useful when you are
1585 debugging servers and you want to dodge out and initiate a connection
1586 to a server running under gdb.
1595 Frequently requested but not approved requests.
1599 Eliminate unused argument warnings using ATTRIBUTE_UNUSED.
1601 The benefits on this one are thought to be marginal - GDBs design
1602 means that unused parameters are very common. GCC 3.0 will also
1603 include the option -Wno-unused-parameter which means that ``-Wall
1604 -Wno-unused-parameters -Werror'' can be specified.
1613 This list is not up to date, and opinions vary about the importance or
1614 even desirability of some of the items. If you do fix something, it
1615 always pays to check the below.
1619 @c This does not work (yet if ever). FIXME.
1620 @c @item --parse=@var{lang} @dots{}
1621 @c Configure the @value{GDBN} expression parser to parse the listed languages.
1622 @c @samp{all} configures @value{GDBN} for all supported languages. To get a
1623 @c list of all supported languages, omit the argument. Without this
1624 @c option, @value{GDBN} is configured to parse all supported languages.
1628 START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that
1629 is its default value. Clean this up.
1633 It should be possible to use symbols from shared libraries before we know
1634 exactly where the libraries will be loaded. E.g. "b perror" before running
1635 the program. This could maybe be done as an extension of the "breakpoint
1636 re-evaluation" after new symbols are loaded.
1640 Make single_step() insert and remove breakpoints in one operation.
1642 [If this is talking about having single_step() insert the breakpoints,
1643 run the target then pull the breakpoints then it is wrong. The
1644 function has to return as control has to eventually be passed back to
1645 the main event loop.]
1649 Speed up single stepping by avoiding extraneous ptrace calls.
1653 Speed up single stepping by not inserting and removing breakpoints
1654 each time the inferior starts and stops.
1656 Breakpoints should not be inserted and deleted all the time. Only the
1657 one(s) there should be removed when we have to step over one. Support
1658 breakpoints that don't have to be removed to step over them.
1660 [this has resulted in numerous debates. The issue isn't clear cut]
1664 Provide "voodoo" debugging of core files. This creates a zombie
1665 process as a child of the debugger, and loads it up with the data,
1666 stack, and regs of the core file. This allows you to call functions
1667 in the executable, to manipulate the data in the core file.
1673 GDB reopens the source file on every line, as you "next" through it.
1675 [still true? I've a memory of this being fixed]
1679 Perhaps "i source" should take an argument like that of "list".
1683 Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if
1684 it matches the source line indicated.
1688 The prompt at end of screen should accept space as well as CR.
1692 Backtrace should point out what the currently selected frame is, in
1693 its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar,
1694 ...)" rather than "#3 foo (bar, ...)".
1698 "i program" should work for core files, and display more info, like what
1699 actually caused it to die.
1703 "x/10i" should shorten the long name, if any, on subsequent lines.
1707 "next" over a function that longjumps, never stops until next time you happen
1708 to get to that spot by accident. E.g. "n" over execute_command which has
1713 "set zeroprint off", don't bother printing members of structs which
1714 are entirely zero. Useful for those big structs with few useful
1719 GDB does four ioctl's for every command, probably switching terminal modes
1720 to/from inferior or for readline or something.
1724 terminal_ours versus terminal_inferior: cache state. Switch should be a noop
1725 if the state is the same, too.
1729 "i frame" shows wrong "arglist at" location, doesn't show where the args
1730 should be found, only their actual values.
1734 There should be a way for "set" commands to validate the new setting
1735 before it takes effect.
1739 "ena d" is ambiguous, why? "ena delete" seems to think it is a command!
1743 i line VAR produces "Line number not known for symbol ``var''.". I
1744 thought we were stashing that info now!
1748 We should be able to write to random files at hex offsets like adb.
1752 [elena - delete this]
1754 Handle add_file with separate text, data, and bss addresses. Maybe
1755 handle separate addresses for each segment in the object file?
1759 [Jimb/Elena delete this one]
1761 Handle free_named_symtab to cope with multiply-loaded object files
1762 in a dynamic linking environment. Should remember the last copy loaded,
1763 but not get too snowed if it finds references to the older copy.
1767 [elena delete this also]
1769 Remove all references to:
1776 now that we have BFD. All remaining are in machine dependent files.
1780 Re-organize help categories into things that tend to fit on a screen
1785 Add in commands like ADB's for searching for patterns, etc. We should
1786 be able to examine and patch raw unsymboled binaries as well in gdb as
1787 we can in adb. (E.g. increase the timeout in /bin/login without source).
1789 [actually, add ADB interface :-]
1793 When doing "step" or "next", if a few lines of source are skipped between
1794 the previous line and the current one, print those lines, not just the
1795 last line of a multiline statement.
1799 Handling of "&" address-of operator needs some serious overhaul
1800 for ANSI C and consistency on arrays and functions.
1801 For "float point[15];":
1802 ptype &point[4] ==> Attempt to take address of non-lvalue.
1803 For "char *malloc();":
1804 ptype malloc ==> "char *()"; should be same as
1805 ptype &malloc ==> "char *(*)()"
1806 call printf ("%x\n", malloc) ==> weird value, should be same as
1807 call printf ("%x\n", &malloc) ==> correct value
1811 Fix dbxread.c symbol reading in the presence of interrupts. It
1812 currently leaves a cleanup to blow away the entire symbol table when a
1813 QUIT occurs. (What's wrong with that? -kingdon, 28 Oct 1993).
1815 [I suspect that the grype was that, on a slow system, you might want
1816 to cntrl-c and get just half the symbols and then load the rest later
1817 - scary to be honest]
1821 Mipsread.c reads include files depth-first, because the dependencies
1822 in the psymtabs are way too inclusive (it seems to me). Figure out what
1823 really depends on what, to avoid recursing 20 or 30 times while reading
1828 value_add() should be subtracting the lower bound of arrays, if known,
1829 and possibly checking against the upper bound for error reporting.
1833 When listing source lines, check for a preceding \n, to verify that
1834 the file hasn't changed out from under us.
1836 [fixed by some other means I think. That hack wouldn't actually work
1837 reliably - the file might move such that another \n appears. ]
1841 Get all the remote systems (where the protocol allows it) to be able to
1842 stop the remote system when the GDB user types ^C (like remote.c
1843 does). For ebmon, use ^Ak.
1847 Possible feature: A version of the "disassemble" command which shows
1848 both source and assembly code ("set symbol-filename on" is a partial
1851 [has this been done? It was certainly done for MI and GDBtk]
1855 investigate "x/s 0" (right now stops early) (I think maybe GDB is
1856 using a 0 address for bad purposes internally).
1860 Make "info path" and path_command work again (but independent of the
1861 environment either of gdb or that we'll pass to the inferior).
1865 Make GDB understand the GCC feature for putting octal constants in
1866 enums. Make it so overflow on an enum constant does not error_type
1867 the whole type. Allow arbitrarily large enums with type attributes.
1868 Put all this stuff in the testsuite.
1872 Make TYPE_CODE_ERROR with a non-zero TYPE_LENGTH more useful (print
1873 the value in hex; process type attributes). Add this to the
1874 testsuite. This way future compilers can add new types and old
1875 versions of GDB can do something halfway reasonable.
1879 Fix mdebugread.c:parse_type to do fundamental types right (see
1880 rs6000_builtin_type in stabsread.c for what "right" is--the point is
1881 that the debug format fixes the sizes of these things and it shouldn't
1882 depend on stuff like TARGET_PTR_BIT and so on. For mdebug, there seem
1883 to be separate bt* codes for 64 bit and 32 bit things, and GDB should
1884 be aware of that). Also use a switch statement for clarity and speed.
1888 Investigate adding symbols in target_load--some targets do, some
1893 Put dirname in psymtabs and change lookup*symtab to use dirname (so
1894 /foo/bar.c works whether compiled by cc /foo/bar.c, or cd /foo; cc
1899 Merge xcoffread.c and coffread.c. Use breakpoint_re_set instead of
1904 Make a watchpoint which contains a function call an error (it is
1905 broken now, making it work is probably not worth the effort).
1909 New test case based on weird.exp but in which type numbers are not
1910 renumbered (thus multiply defining a type). This currently causes an
1911 infinite loop on "p v_comb".
1915 [Hey! Hint Hint Delete Delete!!!]
1917 Fix 386 floating point so that floating point registers are real
1918 registers (but code can deal at run-time if they are missing, like
1919 mips and 68k). This would clean up "info float" and related stuff.
1923 gcc -g -c enummask.c then gdb enummask.o, then "p v". GDB complains
1924 about not being able to access memory location 0.
1926 -------------------- enummask.c
1947 If try to modify value in file with "set write off" should give
1948 appropriate error not "cannot access memory at address 0x65e0".
1952 Allow core file without exec file on RS/6000.
1956 Make sure "shell" with no arguments works right on DOS.
1960 Make gdb.ini (as well as .gdbinit) be checked on all platforms, so
1961 the same directory can be NFS-mounted on unix or DOS, and work the
1966 [Is this another delete???]
1968 Get SECT_OFF_TEXT stuff out of objfile_relocate (might be needed to
1969 get RS/6000 to work right, might not be immediately relevant).
1973 Work out some kind of way to allow running the inferior to be done as
1974 a sub-execution of, eg. breakpoint command lists. Currently running
1975 the inferior interupts any command list execution. This would require
1976 some rewriting of wait_for_inferior & friends, and hence should
1977 probably be done in concert with the above.
1981 Add function arguments to gdb user defined functions.
1985 Add convenience variables that refer to exec file, symbol file,
1986 selected frame source file, selected frame function, selected frame
1991 Modify the handling of symbols grouped through BINCL/EINCL stabs to
1992 allocate a partial symtab for each BINCL/EINCL grouping. This will
1993 seriously decrease the size of inter-psymtab dependencies and hence
1994 lessen the amount that needs to be read in when a new source file is
1999 Add a command for searching memory, a la adb. It specifies size,
2000 mask, value, start address. ADB searches until it finds it or hits
2001 an error (or is interrupted).
2005 Remove the range and type checking code and documentation, if not