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.
7 Known problems in GDB 5.0
8 =========================
10 Below is a list of problems identified during the GDB 5.0 release
11 cycle. People hope to have these problems fixed in a follow-on
14 (The names in paren indicate people that posted the original problem.)
18 GDB requires GCC to build under IRIX
20 IRIX, being more pedantic than GCC reports as errors certain
21 assignments that GCC treats as warnings.
23 This can be worked around by building GDB with the GCC compiler.
27 The BFD directory requires bug-fixed AUTOMAKE et.al.
29 AUTOMAKE 1.4 incorrectly set the TEXINPUTS environment variable. It
30 contained the full path to texinfo.tex when it should have only
31 contained the directory. The bug has been fixed in the current
32 AUTOMAKE sources. Automake snapshots can be found in:
33 ftp://sourceware.cygnus.com/pub/gdb/snapshots
34 and ftp://sourceware.cygnus.com/pub/binutils
38 Generic: lin-thread cannot handle thread exit (Mark Kettenis, Michael
39 Snyder) http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00525.html
41 The thread_db assisted debugging code doesn't handle exiting threads
42 properly, at least in combination with glibc 2.1.3 (the framework is
43 there, just not the actual code). There are at least two problems
44 that prevent this from working.
46 As an additional reference point, the pre thread_db code did not work
51 Java (Anthony Green, David Taylor)
53 Anthony Green has a number of Java patches that did not make it into
57 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00512.html
60 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00515.html
62 Patch: handle N_MAIN stab
63 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00527.html
67 Pascal (Pierre Muller, David Taylor)
69 Pierre Muller has contributed patches for adding Pascal Language
72 2 pascal language patches inserted in database
73 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00521.html
76 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00496.html
80 GNU/Linux/x86 and random thread signals (and Solaris/SPARC but not
82 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00336.html
84 Christopher Blizzard writes:
86 So, I've done some more digging into this and it looks like Jim
87 Kingdon has reported this problem in the past:
89 http://sourceware.cygnus.com/ml/bug-gdb/1999-10/msg00058.html
91 I can reproduce this problem both with and without Tom's patch. Has
92 anyone seen this before? Maybe have a solution for it hanging around?
95 There's a test case for this documented at:
97 when debugging threaded applications you get extra SIGTRAPs
98 http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=9565
100 [There should be a GDB testcase - cagney]
104 Possible regressions with some devel GCCs.
105 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00475.html
107 gcc-2.95.2 outputs a line note *before* the prologue (and one for the
108 closing brace after the epilogue, instead of before it, as it used to
109 be). By disabling the RTL-style prologue generating mechanism
110 (undocumented GCC option -mno-schedule-prologue), you get back the
111 traditional behaviour.
112 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00510.html
114 This should now be fixed.
118 RFD: infrun.c: No bpstat_stop_status call after proceed over break?
120 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00665.html
122 GDB misses watchpoint triggers after proceeding over a breakpoint on
127 x86 linux GDB and SIGALRM (???)
128 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00803.html
130 I know there are problems with single stepping through signal
131 handlers. These problems were present in 4.18. They were just masked
132 because 4.18 failed to recognize signal handlers. Fixing it is not
133 easy, and will require changes to handle_inferior_event(), that I
134 prefer not to make before the 5.0 release.
140 Revised UDP support (was: Re: [Fwd: [patch] UDP transport support])
141 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00000.html
143 (Broken) support for GDB's remote protocol across UDP is to be
144 included in the follow-on release.
148 Can't build IRIX -> arm GDB.
149 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00356.html
152 > Now I'm building for an embedded arm target. If there is a way of turning
153 > remote-rdi off, I couldn't find it. It looks like it gets built by default
154 > in gdb/configure.tgt(line 58) Anyway, the build dies in
155 > gdb/rdi-share/unixcomm.c. SERPORT1 et. al. never get defined because we
156 > aren't one of the architectures supported.
160 Problem with weak functions
161 http://sourceware.cygnus.com/ml/gdb/2000-05/msg00060.html
163 Dan Nicolaescu writes:
164 > It seems that gdb-4.95.1 does not display correctly the function when
165 > stoping in weak functions.
167 > It stops in a function that is defined as weak, not in the function
168 > that is actualy run...
172 GDB5 TOT on unixware 7
173 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00119.html
176 > I just spun the top of tree of the GDB5 branch on UnixWare 7. As a
177 > practical matter, the current thread support is somewhat more annoying
178 > than when GDB was thread-unaware.
185 The following are small cleanups that will hopefully be completed by
186 the follow on to 5.0.
192 The need for this as almost been eliminated. The next version of GCC
193 (assuming cagney gets the relevant patch committed) will be able to
194 supress unused parameter warnings.
198 Eliminate more compiler warnings.
200 Of course there also needs to be the usual debate over which warnings
201 are valid and how to best go about this.
203 One method: choose a single option; get agreement that it is
204 reasonable; try it out to see if there isn't anything silly about it
205 (-Wunused-parameters is an example of that) then incrementally hack
208 The other method is to enable all warnings and eliminate them from one
213 Delete macro TARGET_BYTE_ORDER_SELECTABLE.
215 Patches in the database.
221 Readline 4.? is out. A merge wouldn't hurt.
227 Eliminate all uses of PARAMS in GDB's source code.
231 Elimination of make_cleanup_func. (Andrew Cagney)
233 make_cleanup_func elimination
234 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00791.html
235 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00814.html
239 Elimination of ``(catch_errors_ftype *) func''.
241 Like make_cleanup_func it isn't portable.
245 Re: Various C++ things
247 value_headof/value_from_vtable_info are worthless, and should be removed.
248 The one place in printcmd.c that uses it should use the RTTI functions.
250 RTTI for g++ should be using the typeinfo functions rather than the vtables.
251 The typeinfo functions are always at offset 4 from the beginning of the vtable,
252 and are always right. The vtables will have weird names like E::VB sometimes.
253 The typeinfo function will always be "E type_info function", or somesuch.
255 value_virtual_fn_field needs to be fixed so there are no failures for virtual
256 functions for C++ using g++.
258 Testsuite cases are the major priority right now for C++ support, since i have
259 to make a lot of changes that could potentially break each other.
263 Fix ``set architecture <tab>''
265 This command should expand to a list of all supported architectures.
266 At present ``info architecture'' needs to be used. That is simply
267 wrong. It involves the use of add_set_enum_cmd().
271 GDBARCH cleanup (Andrew Cagney)
273 The non-generated parts of gdbarch.{sh,h,c} should be separated out
274 into arch-utils.[hc].
276 Document that gdbarch_init_ftype could easily fail because it didn't
277 identify an architecture.
281 Migrate qfThreadInfo packet -> qThreadInfo. (Andrew Cagney)
283 Add support for packet enable/disable commands with these thread
284 packets. General cleanup.
286 [PATCH] Document the ThreadInfo remote protocol queries
287 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00832.html
289 [PATCH] "info threads" queries for remote.c
290 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00831.html
294 General Cleanups / Fixes
295 ========================
297 The following are more general cleanups and fixes. They are not tied
298 to any specific release.
306 [PATCH/5] src/intl/Makefile.in:distclean additions
307 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00363.html
309 Do not forget to merge the patch back into the trunk.
313 Update ALPHA so that it uses ``struct frame_extra_info'' instead of
316 This is a barrier to replacing mips_extra_func_info with something
317 that works with multi-arch.
321 Multi-arch mips_extra_func_info.
323 This first needs the alpha to be updated so that it uses ``struct
328 Send normal output to gdb_stdout.
329 Send error messages to gdb_stderror.
330 Send debug and log output log gdb_stdlog.
332 GDB still contains many cases where (f)printf or printf_filtered () is
333 used when it should be sending the messages to gdb_stderror or
334 gdb_stdlog. The thought of #defining printf to something has crossed
339 Rationalize the host-endian code (grep for HOST_BYTE_ORDER).
341 At present defs.h includes <endian.h> (which is linux specific) yet
342 almost nothing depends on it. Suggest "gdb_endian.h" which can also
343 handle <machine/endian.h> and only include that where it is really
348 Replace asprintf() calls with xasprintf() calls.
350 As with things like strdup() most calls to asprintf() don't check the
355 Rationaize savestring(), msavestring() and mstrsave().
357 In general libiberty's xstrdup () can be used.
361 Eliminate mmalloc() from GDB.
363 Also eliminate it from defs.h.
367 Check/cleanup MI documentation.
369 The list of commands specified in the documentation needs to be
370 checked against the mi-cmds.c table in a mechanical way (so that they
371 two can be kept up-to-date).
375 Eliminate error_begin().
377 With ui_file, there is no need for the statefull error_begin ()
382 Add built-by, build-date, tm, xm, nm and anything else into gdb binary
383 so that you can see how the GDB was created.
385 Some of these (*m.h) would be added to the generated config.h. That
386 in turn would fix a long standing bug where by the build process many
387 not notice a changed tm.h file. Since everything depends on config.h,
388 a change to *m.h forces a change to config.h and, consequently forces
393 Replace gdb_stdtarg with gdb_targout (and possibly gdb_targerr).
395 gdb_stdtarg is easily confused with gdb_stdarg.
399 Remote protocol doco feedback.
401 Too much feedback to mention needs to be merged in (901660). Search
402 for the word ``remote''.
406 set/show remote X-packet ...
408 ``(gdb) help set remote X-packet'' doesn't list the applicable
409 responses. The help message needs to be expanded.
413 Extra ui_file methods - dump.
415 These are for debugging / testing. An aside is to set up a whitebox
416 testsuite for key internals such as ui_file.
420 Add an "info bfd" command that displays supported object formats,
421 similarly to objdump -i.
423 Is there a command already?
427 Eliminate PTR. ISO-C allows ``void *''.
433 GDB should never abort. GDB should either throw ``error ()'' or
434 ``internal_error ()''. Better still GDB should naturally unwind with
439 GDB probably doesn't build on FreeBSD pre 2.2.x
440 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00378.html
442 Fixes to get FreeBSD working on 2.2.x, 3.x and 4.x caused the code to
447 Architectural Changes
448 =====================
450 These are harder than simple cleanups / fixes and, consequently
451 involve more work. Typically an Architectural Change will be broken
452 down into a more digestible set of cleanups and fixes.
456 Replace READ_FP() with FRAME_HANDLE().
458 READ_FP() is a hangover from the days of the vax when the ABI really
459 did have a frame pointer register. Modern architectures typically
460 construct a virtual frame-handle from the stack pointer and various
461 other bits of string.
463 Unfortunatly GDB still treats this synthetic FP register as though it
464 is real. That in turn really confuses users (arm and ``print $fp'' VS
465 ``info registers fp''). The synthetic FP should be separated out of
466 the true register set presented to the user.
470 MI's input does not use buffering.
472 At present the MI interface reads raw characters of from an unbuffered
473 FD. This is to avoid several nasty buffer/race conditions. That code
474 should be changed so that it registers its self with the event loop
475 (on the input FD) and then push commands up to MI as they arrive.
477 The serial code already does this.
481 Register Cache Cleanup (below from Andrew Cagney)
483 I would depict the current register architecture as something like:
490 register + REGISTER_BYTE(reg_nr)
493 -------------------------
494 | extern register[] |
495 -------------------------
497 where neither the high (valops.c et.al.) or low gdb (*-tdep.c) are
498 really clear on what mechanisms they should be using to manipulate that
499 buffer. Further, much code assumes, dangerously, that registers are
500 contigious. Having got mips-tdep.c to support multiple ABIs, believe
501 me, that is a bad assumption. Finally, that register cache layout is
502 determined by the current remote/local target and _not_ the less
503 specific target ISA. In fact, in many cases it is determined by the
504 somewhat arbitrary layout of the [gG] packets!
507 How I would like the register file to work is more like:
535 The main objectives being:
537 o a clear separation between the low
538 level target and the high level GDB
540 o a mechanism that solves the general
541 problem of register aliases, overlaps
542 etc instead of treating them as optional
543 extras that can be wedged in as an after
544 thought (that is a reasonable description
545 of the current code).
547 Identify then solve the hard case and the
548 rest just falls out. GDB solved the easy
549 case and then tried to ignore the real
552 o a removal of the assumption that the
553 mapping between the register cache
554 and virtual registers is largely static.
555 If you flip the USR/SSR stack register
556 select bit in the status-register then
557 the corresponding stack registers should
560 o a mechanism that clearly separates the
561 gdb internal register cache from any
562 target (not architecture) dependant
563 specifics such as [gG] packets.
565 Of course, like anything, it sounds good in theory. In reality, it
566 would have to contend with many<->many relationships at both the
567 virt<->cache and cache<->target level. For instance:
570 Modifying an mmx register may involve
571 scattering values across both FP and
572 mmpx specific parts of a buffer
575 When writing back a SP it may need to
576 both be written to both SP and USP.
581 Rather than let this like the last time it was discussed, just slip, I'm
582 first going to add this e-mail (+ references) to TODO. I'd then like to
583 sketch out a broad strategy I think could get us there.
586 First thing I'd suggest is separating out the ``extern registers[]''
587 code so that we can at least identify what is using it. At present
588 things are scattered across many files. That way we can at least
589 pretend that there is a cache instead of a global array :-)
591 I'd then suggest someone putting up a proposal for the pseudo-reg /
592 high-level side interface so that code can be adopted to it. For old
593 code, initially a blanket rename of write_register_bytes() to
594 deprecated_write_register_bytes() would help.
596 Following that would, finaly be the corresponding changes to the target.
600 Fix ``I'm sorry, Dave, I can't do that.'' from symfile.c.
602 This requires internationalization.
606 Check that GDB can handle all BFD architectures (Andrew Cagney)
608 There should be a test that checks that BFD/GDB are in sync with
609 regard to architecture changes. Something like a test that first
610 queries GDB for all supported architectures and then feeds each back
611 to GDB.. Anyone interested in learning how to write tests? :-)
615 Add support for Modula3
617 Get DEC/Compaq to contribute their Modula-3 support.
621 Convert ALL architectures to MULTI-ARCH.
625 Convert GDB build process to AUTOMAKE.
627 See also sub-directory configure below.
631 Cleanup configury support for optional sub-directories.
633 Check how GCC handles multiple front ends for an example of how things
634 could work. A tentative first step is to rationalize things so that
635 all sub directories are handled in a fashion similar to gdb/mi.
637 See also automake above.
641 Restructure gdb directory tree so that it avoids any 8.3 and 14
646 Can the xdep files be replaced by autoconf?
647 Can the tm.h and nm.h files be eliminated by multi-arch.
651 Add a transcript mechanism to GDB.
653 Such a mechanism might log all gdb input and output to a file in a
654 form that would allow it to be replayed. It could involve ``gdb
655 --transcript=FILE'' or it could involve ``(gdb) transcript file''.
659 Make MI interface accessable from existing CLI.
663 Select the initial multi-arch ISA / ABI based on --target or similar.
665 At present the default is based on what ever is first in the BFD
666 archures table. It should be determined based on the ``--target=...''
673 Enable the code to recognize --enable-targets=.... like BINUTILS does.
677 Add a breakpoint-edit command to MI.
679 It would be similar to MI's breakpoint create but would apply to an
680 existing breakpoint. It saves the need to delete/create breakpoints
681 when ever they are changed.
685 Add directory path to MI breakpoint.
687 That way the GUI's task of finding the file within which the
688 breakpoint was set is simplified.
692 Re-do GDB's output pager.
694 GDB's output pager still relies on people correctly using *_filtered
695 for gdb_stdout and *_unfiltered for gdb_stdlog / gdb_stderr.
696 Hopefully, with all normal output going to gdb_stdout, the pager can
697 just look at the ui_file that the output is on and then use that to
698 decide what to do about paging. Sounds good in theory.
702 Add mechanism to reject expression classes to MI
704 There are situtations where you don't want GDB's expression
705 parser/evaluator to perform inferior function calls or variable
710 Remove sideffects from libgdb breakpoint create function.
712 The user can use the CLI to create a breakpoint with partial
713 information - no file (gdb would use the file from the last
716 The libgdb interface currently affects that environment which can lead
717 to confusion when a user is setting breakpoints via both the MI and
720 This is also a good example of how getting the CLI ``right'' will be
725 GDB doesn't recover gracefully from remote protocol errors.
727 GDB wasn't checking for NAKs from the remote target. Instead a NAK is
728 ignored and a timeout is required before GDB retries. A pre-cursor to
729 fixing this this is making GDB's remote protocol packet more robust.
731 While downloading to a remote protocol target, gdb ignores packet
732 errors in so far as it will continue to edownload with chunk N+1 even
733 if chunk N was not correctly sent. This causes gdb.base/remote.exp to
734 take a painfully long time to run. As a PS that test needs to be
735 fixed so that it builds on 16 bit machines.
739 Move gdb_lasterr to ui_out?
741 The way GDB throws errors and records them needs a re-think. ui_out
742 handles the correct output well. It doesn't resolve what to do with
743 output / error-messages when things go wrong.
747 Fix implementation of ``target xxx''.
749 At present when the user specifies ``target xxxx'', the CLI maps that
750 directly onto a target open method. It is then assumed that the
751 target open method should do all sorts of complicated things as this
752 is the only chance it has. Check how the various remote targets
753 duplicate the target operations. Check also how the various targets
754 behave differently for purely arbitrary reasons.
756 What should happen is that ``target xxxx'' should call a generic
757 ``target'' function and that should then co-ordinate the opening of
758 ``xxxx''. This becomes especially important when you're trying to
759 open an asynchronous target that may need to perform background tasks
760 as part of the ``attach'' phase.
762 Unfortunatly, due to limitations in the old/creaking command.h
763 interface, that isn't possible. The function being called isn't told
764 of the ``xxx'' or any other context information.
766 Consequently a precursor to fixing ``target xxxx'' is to clean up the
767 CLI code so that it passes to the callback function (attatched to a
768 command) useful information such as the actual command and a context
769 for that command. Other changes such as making ``struct command''
770 opaque may also help.
774 Document trace machinery
778 Document overlay machinery.
785 This list is not up to date, and opinions vary about the importance or
786 even desirability of some of the items. If you do fix something, it
787 always pays to check the below.
791 @c This does not work (yet if ever). FIXME.
792 @c @item --parse=@var{lang} @dots{}
793 @c Configure the @value{GDBN} expression parser to parse the listed languages.
794 @c @samp{all} configures @value{GDBN} for all supported languages. To get a
795 @c list of all supported languages, omit the argument. Without this
796 @c option, @value{GDBN} is configured to parse all supported languages.
800 START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that
801 is its default value. Clean this up.
805 It should be possible to use symbols from shared libraries before we know
806 exactly where the libraries will be loaded. E.g. "b perror" before running
807 the program. This could maybe be done as an extension of the "breakpoint
808 re-evaluation" after new symbols are loaded.
812 Make single_step() insert and remove breakpoints in one operation.
816 Speed up single stepping by avoiding extraneous ptrace calls.
820 Speed up single stepping by not inserting and removing breakpoints
821 each time the inferior starts and stops.
823 Breakpoints should not be inserted and deleted all the time. Only the
824 one(s) there should be removed when we have to step over one. Support
825 breakpoints that don't have to be removed to step over them.
827 [this has resulted in numerous debates. The issue isn't clear cut]
831 Provide "voodoo" debugging of core files. This creates a zombie
832 process as a child of the debugger, and loads it up with the data,
833 stack, and regs of the core file. This allows you to call functions
834 in the executable, to manipulate the data in the core file.
840 GDB reopens the source file on every line, as you "next" through it.
842 [still true? I've a memory of this being fixed]
846 Perhaps "i source" should take an argument like that of "list".
850 Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if
851 it matches the source line indicated.
855 The prompt at end of screen should accept space as well as CR.
859 Backtrace should point out what the currently selected frame is, in
860 its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar,
861 ...)" rather than "#3 foo (bar, ...)".
865 "i program" should work for core files, and display more info, like what
866 actually caused it to die.
870 "x/10i" should shorten the long name, if any, on subsequent lines.
874 "next" over a function that longjumps, never stops until next time you happen
875 to get to that spot by accident. E.g. "n" over execute_command which has
880 "set zeroprint off", don't bother printing members of structs which
881 are entirely zero. Useful for those big structs with few useful
886 GDB does four ioctl's for every command, probably switching terminal modes
887 to/from inferior or for readline or something.
891 terminal_ours versus terminal_inferior: cache state. Switch should be a noop
892 if the state is the same, too.
896 "i frame" shows wrong "arglist at" location, doesn't show where the args
897 should be found, only their actual values.
901 There should be a way for "set" commands to validate the new setting
902 before it takes effect.
906 "ena d" is ambiguous, why? "ena delete" seems to think it is a command!
910 i line VAR produces "Line number not known for symbol ``var''.". I
911 thought we were stashing that info now!
915 We should be able to write to random files at hex offsets like adb.
919 Make "target xxx" command interruptible.
923 [elena - delete this]
925 Handle add_file with separate text, data, and bss addresses. Maybe
926 handle separate addresses for each segment in the object file?
930 [Jimb/Elena delete this one]
932 Handle free_named_symtab to cope with multiply-loaded object files
933 in a dynamic linking environment. Should remember the last copy loaded,
934 but not get too snowed if it finds references to the older copy.
938 [elena delete this also]
940 Remove all references to:
947 now that we have BFD. All remaining are in machine dependent files.
951 Re-organize help categories into things that tend to fit on a screen
956 Add in commands like ADB's for searching for patterns, etc. We should
957 be able to examine and patch raw unsymboled binaries as well in gdb as
958 we can in adb. (E.g. increase the timeout in /bin/login without source).
960 [actually, add ADB interface :-]
964 When doing "step" or "next", if a few lines of source are skipped between
965 the previous line and the current one, print those lines, not just the
966 last line of a multiline statement.
970 Handling of "&" address-of operator needs some serious overhaul
971 for ANSI C and consistency on arrays and functions.
972 For "float point[15];":
973 ptype &point[4] ==> Attempt to take address of non-lvalue.
974 For "char *malloc();":
975 ptype malloc ==> "char *()"; should be same as
976 ptype &malloc ==> "char *(*)()"
977 call printf ("%x\n", malloc) ==> weird value, should be same as
978 call printf ("%x\n", &malloc) ==> correct value
982 Fix dbxread.c symbol reading in the presence of interrupts. It
983 currently leaves a cleanup to blow away the entire symbol table when a
984 QUIT occurs. (What's wrong with that? -kingdon, 28 Oct 1993).
986 [I suspect that the grype was that, on a slow system, you might want
987 to cntrl-c and get just half the symbols and then load the rest later
988 - scary to be honest]
992 Mipsread.c reads include files depth-first, because the dependencies
993 in the psymtabs are way too inclusive (it seems to me). Figure out what
994 really depends on what, to avoid recursing 20 or 30 times while reading
999 value_add() should be subtracting the lower bound of arrays, if known,
1000 and possibly checking against the upper bound for error reporting.
1004 When listing source lines, check for a preceding \n, to verify that
1005 the file hasn't changed out from under us.
1007 [fixed by some other means I think. That hack wouldn't actually work
1008 reliably - the file might move such that another \n appears. ]
1012 Get all the remote systems (where the protocol allows it) to be able to
1013 stop the remote system when the GDB user types ^C (like remote.c
1014 does). For ebmon, use ^Ak.
1018 Possible feature: A version of the "disassemble" command which shows
1019 both source and assembly code ("set symbol-filename on" is a partial
1022 [has this been done? It was certainly done for MI and GDBtk]
1026 investigate "x/s 0" (right now stops early) (I think maybe GDB is
1027 using a 0 address for bad purposes internally).
1031 Make "info path" and path_command work again (but independent of the
1032 environment either of gdb or that we'll pass to the inferior).
1036 Make GDB understand the GCC feature for putting octal constants in
1037 enums. Make it so overflow on an enum constant does not error_type
1038 the whole type. Allow arbitrarily large enums with type attributes.
1039 Put all this stuff in the testsuite.
1043 Make TYPE_CODE_ERROR with a non-zero TYPE_LENGTH more useful (print
1044 the value in hex; process type attributes). Add this to the
1045 testsuite. This way future compilers can add new types and old
1046 versions of GDB can do something halfway reasonable.
1050 Fix mdebugread.c:parse_type to do fundamental types right (see
1051 rs6000_builtin_type in stabsread.c for what "right" is--the point is
1052 that the debug format fixes the sizes of these things and it shouldn't
1053 depend on stuff like TARGET_PTR_BIT and so on. For mdebug, there seem
1054 to be separate bt* codes for 64 bit and 32 bit things, and GDB should
1055 be aware of that). Also use a switch statement for clarity and speed.
1059 Investigate adding symbols in target_load--some targets do, some
1064 Put dirname in psymtabs and change lookup*symtab to use dirname (so
1065 /foo/bar.c works whether compiled by cc /foo/bar.c, or cd /foo; cc
1070 Merge xcoffread.c and coffread.c. Use breakpoint_re_set instead of
1075 Make a watchpoint which contains a function call an error (it is
1076 broken now, making it work is probably not worth the effort).
1080 New test case based on weird.exp but in which type numbers are not
1081 renumbered (thus multiply defining a type). This currently causes an
1082 infinite loop on "p v_comb".
1086 [Hey! Hint Hint Delete Delete!!!]
1088 Fix 386 floating point so that floating point registers are real
1089 registers (but code can deal at run-time if they are missing, like
1090 mips and 68k). This would clean up "info float" and related stuff.
1094 gcc -g -c enummask.c then gdb enummask.o, then "p v". GDB complains
1095 about not being able to access memory location 0.
1097 -------------------- enummask.c
1118 If try to modify value in file with "set write off" should give
1119 appropriate error not "cannot access memory at address 0x65e0".
1123 Allow core file without exec file on RS/6000.
1127 Make sure "shell" with no arguments works right on DOS.
1131 Make gdb.ini (as well as .gdbinit) be checked on all platforms, so
1132 the same directory can be NFS-mounted on unix or DOS, and work the
1137 [Is this another delete???]
1139 Get SECT_OFF_TEXT stuff out of objfile_relocate (might be needed to
1140 get RS/6000 to work right, might not be immediately relevant).
1144 Work out some kind of way to allow running the inferior to be done as
1145 a sub-execution of, eg. breakpoint command lists. Currently running
1146 the inferior interupts any command list execution. This would require
1147 some rewriting of wait_for_inferior & friends, and hence should
1148 probably be done in concert with the above.
1152 Add function arguments to gdb user defined functions.
1156 Add convenience variables that refer to exec file, symbol file,
1157 selected frame source file, selected frame function, selected frame
1162 Add a "suspend" subcommand of the "continue" command to suspend gdb
1163 while continuing execution of the subprocess. Useful when you are
1164 debugging servers and you want to dodge out and initiate a connection
1165 to a server running under gdb.
1171 Modify the handling of symbols grouped through BINCL/EINCL stabs to
1172 allocate a partial symtab for each BINCL/EINCL grouping. This will
1173 seriously decrease the size of inter-psymtab dependencies and hence
1174 lessen the amount that needs to be read in when a new source file is
1181 Modify gdb to work correctly with Pascal.
1185 Add a command for searching memory, a la adb. It specifies size,
1186 mask, value, start address. ADB searches until it finds it or hits
1187 an error (or is interrupted).
1191 Remove the range and type checking code and documentation, if not