gdb/TODO

   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.
   5
   6
   7                         GDB 5.1 - Fixes
   8                         ===============
   9
  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.
  12
  13 --
  14
  15 RFD: infrun.c: No bpstat_stop_status call after proceed over break?
  16 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00665.html
  17
  18 GDB misses watchpoint triggers after proceeding over a breakpoint on
  19 x86 targets.
  20
  21 --
  22
  23 x86 linux GDB and SIGALRM (???)
  24 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00803.html
  25
  26 This problem has been fixed, but a regression test still needs to be
  27 added to the testsuite:
  28 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00309.html
  29
  30 Mark
  31
  32 --
  33
  34 Can't build IRIX -> arm GDB.
  35 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00356.html
  36
  37 David Whedon writes:
  38 > Now I'm building for an embedded arm target.  If there is a way of turning
  39 > remote-rdi off, I couldn't find it. It looks like it gets built by default
  40 > in gdb/configure.tgt(line 58) Anyway, the build dies in
  41 > gdb/rdi-share/unixcomm.c.  SERPORT1 et. al. never get defined because we
  42 > aren't one of the architectures supported.
  43
  44 --
  45
  46 Problem with weak functions
  47 http://sourceware.cygnus.com/ml/gdb/2000-05/msg00060.html
  48
  49 Dan Nicolaescu writes:
  50 > It seems that gdb-4.95.1  does not display correctly the function when
  51 > stoping in weak functions.
  52 >
  53 > It stops in a function that is defined as weak, not in the function
  54 > that is actualy run...
  55
  56 --
  57
  58 GDB 5.0 doesn't work on Linux/SPARC
  59
  60 --
  61
  62 Thread support.  Right now, as soon as a thread finishes and exits,
  63 you're hosed.  This problem is reported once a week or so.
  64
  65 --
  66
  67 Wow, three bug reports for the same problem in one day!  We should
  68 probably make fixing this a real priority :-).
  69
  70 Anyway, thanks for reporting.
  71
  72 The following patch will fix the problems with setting breakpoints in
  73 dynamically loaded objects:
  74
  75    http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00230.html
  76
  77 This patch isn't checked in yet (ping Michael/JimB), but I hope this
  78 will be in the next GDB release.
  79
  80 There should really be a test in the testsuite for this problem, since
  81 it keeps coming up :-(.  Any volunteers?
  82
  83 Mark
  84
  85 --
  86
  87 Re: GDB 5.0.1?
  88 http://sources.redhat.com/ml/gdb/2000-07/msg00038.html
  89
  90 Is the Solaris 8 x86 problem fixed?  When you configure it, configure
  91 incorrectly determines that I have no curses.h.  This causes mucho
  92 compilation errors later on.
  93
  94 Simply editing the config.h to define CURSES_H fixes the problem, and
  95 then the build works fine.
  96
  97 The status for this problem:
  98
  99 Solaris 8 x86 (PIII-560)
 100 gcc 2.95.2
 101
 102 I had the same problem with several of the snapshots shortly before
 103 5.0 became official, and 5.0 has the same problem.
 104
 105 I sent some mail in about it long ago, and never saw a reply.
 106
 107 I haven't had time to figure it out myself, especially since I get all
 108 confused trying to figure out what configure does, I was happy to find
 109 the workaround.
 110
 111 Mike
 112
 113 --
 114
 115                 GDB 5.1 - New features
 116                 ======================
 117
 118 The following new features should be included in 5.1.
 119
 120 --
 121
 122 Enable MI by default.  Old code can be deleted after 5.1 is out.
 123
 124 --
 125
 126 Pascal (Pierre Muller, David Taylor)
 127
 128 Pierre Muller has contributed patches for adding Pascal Language
 129 support to GDB.
 130
 131 2 pascal language patches inserted in database
 132 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00521.html
 133
 134 Indent -gnu ?
 135 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00496.html
 136
 137 --
 138
 139 Java (Anthony Green, David Taylor)
 140
 141 Anthony Green has a number of Java patches that did not make it into
 142 the 5.0 release.  The first two are in cvs now, but the third needs
 143 some fixing up before it can go in.
 144
 145 Patch: java tests
 146 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00512.html
 147
 148 Patch: java booleans
 149 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00515.html
 150
 151 Patch: handle N_MAIN stab
 152 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00527.html
 153
 154 --
 155
 156 [Comming...]
 157
 158 Modify gdb to work correctly with Pascal.
 159
 160 --
 161
 162 Revised UDP support (was: Re: [Fwd: [patch] UDP transport support])
 163 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00000.html
 164
 165 (Broken) support for GDB's remote protocol across UDP is to be
 166 included in the follow-on release.
 167
 168 It should be noted that UDP can only work when the [Gg] packet fits in
 169 a single UDP packet.
 170
 171 There is also much debate over the merit of this.
 172
 173 --
 174
 175                 GDB 5.1 - Cleanups
 176                 ==================
 177
 178 The following code cleanups will hopefully be applied to GDB 5.1.
 179
 180 --
 181
 182 Delete macro TARGET_BYTE_ORDER_SELECTABLE.
 183
 184 Patches in the database.
 185
 186 --
 187
 188 Fix copyright notices.
 189
 190 Turns out that ``1998-2000'' isn't considered valid :-(
 191
 192 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00467.html
 193
 194 --
 195
 196 Purge PARAMS.
 197
 198 Eliminate all uses of PARAMS in GDB's source code.
 199
 200 --
 201
 202 printcmd.c (print_address_numeric):
 203
 204 NOTE: This assumes that the significant address information is kept in
 205 the least significant bits of ADDR - the upper bits were either zero
 206 or sign extended.  Should ADDRESS_TO_POINTER() or some
 207 ADDRESS_TO_PRINTABLE() be used to do the conversion?
 208
 209 --
 210
 211 Compiler warnings.
 212
 213 Eliminate all warnings for at least one host/target for the flags:
 214 -Wimplicit -Wreturn-type -Wcomment -Wtrigraphs -Wformat -Wparentheses
 215 -Wpointer-arith -Wuninitialized
 216
 217 --
 218
 219 Follow through `make check' with --enable-shared.
 220
 221 When the srcware tree is configured with --enable-shared, the `expect'
 222 program won't run properly.  Jim Wilson found out gdb has a local hack
 223 to set LD_LIBRARY_PATH, but, AFAIK, no other project has been hacked
 224 similarly.
 225
 226 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00845.html
 227
 228 --
 229
 230                 GDB 5.2 - Fixes
 231                 ===============
 232
 233 --
 234
 235 Fix at least one thread bug.
 236
 237 --
 238
 239                 GDB 5.2 - New features
 240                 ======================
 241
 242 --
 243
 244 Objective C/C++ Support.  Bu hopefully sooner...
 245
 246 --
 247
 248                 GDB 5.2 - Cleanups
 249                 ==================
 250
 251 The following cleanups have been identified as part of GDB 5.2.
 252
 253 --
 254
 255 Eliminate more compiler warnings.
 256
 257 --
 258
 259 Restructure gdb directory tree so that it avoids any 8.3 and 14
 260 filename problems.
 261
 262 --
 263
 264 Convert GDB build process to AUTOMAKE.
 265
 266 See also sub-directory configure below.
 267
 268 The current convention is (kind of) to use $(<header>_h) in all
 269 dependency lists.  It isn't done in a consistent way.
 270
 271 --
 272
 273                 Code Cleanups: General
 274                 ======================
 275
 276 The following are more general cleanups and fixes.  They are not tied
 277 to any specific release.
 278
 279 --
 280
 281 The BFD directory requires bug-fixed AUTOMAKE et.al.
 282
 283 AUTOMAKE 1.4 incorrectly set the TEXINPUTS environment variable.  It
 284 contained the full path to texinfo.tex when it should have only
 285 contained the directory.  The bug has been fixed in the current
 286 AUTOMAKE sources.  Automake snapshots can be found in:
 287         ftp://sourceware.cygnus.com/pub/gdb/snapshots
 288 and     ftp://sourceware.cygnus.com/pub/binutils
 289
 290 --
 291
 292 Find something better than DEFAULT_BFD_ARCH, DEFAULT_BFD_VEC to
 293 determine the default isa/byte-order.
 294
 295 --
 296
 297 Rely on BFD_BIG_ENDIAN and BFD_LITTLE_ENDIAN instead of host dependant
 298 BIG_ENDIAN and LITTLE_ENDIAN.
 299
 300 --
 301
 302 Eliminate more compiler warnings.
 303
 304 Of course there also needs to be the usual debate over which warnings
 305 are valid and how to best go about this.
 306
 307 One method: choose a single option; get agreement that it is
 308 reasonable; try it out to see if there isn't anything silly about it
 309 (-Wunused-parameters is an example of that) then incrementally hack
 310 away.
 311
 312 The other method is to enable all warnings and eliminate them from one
 313 file at a time.
 314
 315 --
 316
 317 Elimination of ``(catch_errors_ftype *) func''.
 318
 319 Like make_cleanup_func it isn't portable.
 320 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00791.html
 321 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00814.html
 322
 323 --
 324
 325 Nuke #define CONST_PTR.
 326
 327 --
 328
 329 Nuke USG define.
 330
 331 --
 332
 333 [PATCH/5] src/intl/Makefile.in:distclean additions
 334 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00363.html
 335
 336 Do not forget to merge the patch back into the trunk.
 337
 338 --
 339
 340 Rationalize the host-endian code (grep for HOST_BYTE_ORDER).
 341
 342 At present defs.h includes <endian.h> (which is linux specific) yet
 343 almost nothing depends on it.  Suggest "gdb_endian.h" which can also
 344 handle <machine/endian.h> and only include that where it is really
 345 needed.
 346
 347 --
 348
 349 Replace asprintf() calls with xasprintf() calls.
 350
 351 As with things like strdup() most calls to asprintf() don't check the
 352 return value.
 353
 354 --
 355
 356 Replace strsave() + mstrsave() with libiberty:xstrdup().
 357
 358 --
 359
 360 Replace savestring() with something from libiberty.
 361
 362 An xstrldup()? but that would have different semantics.
 363
 364 --
 365
 366 Rationalize use of floatformat_unknown in GDB sources.
 367
 368 Instead of defaulting to floatformat_unknown, should hosts/targets
 369 specify the value explicitly?
 370
 371 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
 372
 373 --
 374
 375 Add a ``name'' member to include/floatformat.h:struct floatformat.
 376 Print that name in gdbarch.c.
 377
 378 --
 379
 380 Sort out the harris mess in include/floatformat.h (it hardwires two
 381 different floating point formats).
 382
 383 --
 384
 385 See of the GDB local floatformat_do_doublest() and libiberty's
 386 floatformat_to_double (which was once GDB's ...) can be merged some
 387 how.
 388
 389 --
 390
 391 Eliminate mmalloc() from GDB.
 392
 393 Also eliminate it from defs.h.
 394
 395 --
 396
 397 Eliminate PTR.  ISO-C allows ``void *''.
 398
 399 --
 400
 401 Eliminate abort ().
 402
 403 GDB should never abort.  GDB should either throw ``error ()'' or
 404 ``internal_error ()''.  Better still GDB should naturally unwind with
 405 an error status.
 406
 407 --
 408
 409 Add __LINE__ and __FILE__ to internal_error().
 410
 411 --
 412
 413 GDB probably doesn't build on FreeBSD pre 2.2.x
 414 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00378.html
 415
 416 Fixes to get FreeBSD working on 2.2.x, 3.x and 4.x caused the code to
 417 suffer bit rot.
 418
 419 --
 420
 421 Deprecate "fg".  Apparently ``fg'' is actually continue.
 422
 423 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00417.html
 424
 425 --
 426
 427 Deprecate current use of ``floatformat_unknown''.
 428
 429 Require all targets to explicitly provide their float format instead
 430 of defaulting to floatformat unknown.  Doing the latter leads to nasty
 431 bugs.
 432
 433 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
 434
 435 --
 436
 437 Rationalize floatformat_to_double() vs floatformat_to_doublest().
 438
 439 Looks like GDB migrated floatformat_to_double() to libiberty but then
 440 turned around and created a ..._to_doublest() the latter containing
 441 several bug fixes.
 442
 443 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00472.html
 444
 445 --
 446
 447 Move floatformat_ia64_ext to libiberty/include floatformat.[ch].
 448
 449 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00466.html
 450
 451 --
 452
 453 The ``maintenance deprecate set endian big'' command doesn't notice
 454 that it is deprecating ``set endian'' and not ``set endian big'' (big
 455 is implemented using an enum).  Is anyone going to notice this?
 456
 457 --
 458
 459 When tab expanding something like ``set arch<tab>'' ignore the
 460 deprecated ``set archdebug'' and expand to ``set architecture''.
 461
 462 --
 463
 464 Eliminate ``arm_register_names[j] = (char *) regnames[j]'' and the
 465 like from arm-tdep.c.
 466
 467 --
 468
 469 Fix uses of ->function.cfunc = set_function().
 470
 471 The command.c code calls sfunc() when a set command.  Rather than
 472 change it suggest fixing the callback function so that it is more
 473 useful.  See:
 474
 475 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
 476
 477 See also ``Fix implementation of  ``target xxx''.'' below.
 478
 479 --
 480
 481 IRIX 3.x support is probably broken.
 482
 483 --
 484
 485 Delete sim/SIM_HAVE_BREAKPOINTS and gdb/SIM_HAS_BREAKPOINTS.
 486 http://sourceware.cygnus.com/ml/gdb-patches/2000-07/msg00042.html
 487
 488 Apart from the d30v, are there any sim/common simulators that make use
 489 of this?
 490
 491 A brief summary of what happended is that sim/common/sim-break.c was
 492 created as a good idea.  It turned out a better idea was to use
 493 SIM_SIGBREAK and have GDB pass back sim_resume (..., SIGBREAK).
 494
 495 --
 496
 497 Move remote_remove_hw_breakpoint, remote_insert_hw_breakpoint,
 498 remote_remove_watchpoint, remote_insert_watchpoint into target vector.
 499
 500 --
 501
 502                         New Features and Fixes
 503                         ======================
 504
 505 These are harder than cleanups but easier than work involving
 506 fundamental architectural change.
 507
 508 --
 509
 510 Add built-by, build-date, tm, xm, nm and anything else into gdb binary
 511 so that you can see how the GDB was created.
 512
 513 --
 514
 515 Add an "info bfd" command that displays supported object formats,
 516 similarly to objdump -i.
 517
 518 Is there a command already?
 519
 520 --
 521
 522 Fix ``I'm sorry, Dave, I can't do that.'' from symfile.c.
 523
 524 This requires internationalization.
 525
 526 --
 527
 528 Add support for:
 529
 530 (gdb) p fwprintf(stdout,L"%S\n", f)
 531 No symbol "L" in current context.
 532
 533 --
 534
 535 Cleanup configury support for optional sub-directories.
 536
 537 Check how GCC handles multiple front ends for an example of how things
 538 could work.  A tentative first step is to rationalize things so that
 539 all sub directories are handled in a fashion similar to gdb/mi.
 540
 541 See also automake above.
 542
 543 --
 544
 545 Add a transcript mechanism to GDB.
 546
 547 Such a mechanism might log all gdb input and output to a file in a
 548 form that would allow it to be replayed.  It could involve ``gdb
 549 --transcript=FILE'' or it could involve ``(gdb) transcript file''.
 550
 551 --
 552
 553 Can the xdep files be replaced by autoconf?
 554
 555 --
 556
 557 Document trace machinery
 558
 559 --
 560
 561 Document ui-out and ui-file.
 562
 563 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00121.html
 564
 565 --
 566
 567 Update texinfo.tex to latest?
 568
 569 --
 570
 571 Incorporate agentexpr.texi into gdb.texinfo
 572
 573 agentexpr.texi mostly describes the details of the byte code used for
 574 tracepoints, not the internals of the support for this in GDB.  So it
 575 looks like gdb.texinfo is a better place for this information.
 576
 577 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00566.html
 578
 579 --
 580
 581 Document overlay machinery.
 582
 583 --
 584
 585 ``(gdb) catch signal SIGNAL''
 586
 587 Overlaps with ``handle SIGNAL'' but the implied behavour is different.
 588 You can attach commands to a catch but not a handle.  A handle has a
 589 limited number of hardwired actions.
 590
 591 --
 592
 593 Get the TUI working on all platforms.
 594
 595 --
 596
 597 Add support for ``gdb --- PROGRAM ARGS ...''.
 598 Add support for ``gdb -cmd=...''
 599
 600 Along with many variations.  Check:
 601
 602 ????? for a full discussion.
 603
 604 for a discussion.
 605
 606 --
 607
 608 Implement ``(gdb) !ls''.
 609
 610 Which is very different from ``(gdb) ! ls''.  Implementing the latter
 611 is trivial.
 612
 613 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00034.html
 614
 615 --
 616
 617 Change the (char *list[]) to (const char (*)[]) so that dynamic lists can
 618 be passed.
 619
 620 --
 621
 622 When tab expanding something like ``set arch<tab>'' ignore the
 623 deprecated ``set archdebug'' and expand to ``set architecture''.
 624
 625 --
 626
 627 Replace the code that uses the host FPU with an emulator of the target
 628 FPU.
 629
 630 --
 631
 632                         Thread Support
 633                         ==============
 634
 635 --
 636
 637 Generic: lin-thread cannot handle thread exit (Mark Kettenis, Michael
 638 Snyder) http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00525.html
 639
 640 The thread_db assisted debugging code doesn't handle exiting threads
 641 properly, at least in combination with glibc 2.1.3 (the framework is
 642 there, just not the actual code).  There are at least two problems
 643 that prevent this from working.
 644
 645 As an additional reference point, the pre thread_db code did not work
 646 either.
 647
 648 --
 649
 650 GNU/Linux/x86 and random thread signals (and Solaris/SPARC but not
 651 Solaris/x86).
 652 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00336.html
 653
 654 Christopher Blizzard writes:
 655
 656 So, I've done some more digging into this and it looks like Jim
 657 Kingdon has reported this problem in the past:
 658
 659 http://sourceware.cygnus.com/ml/bug-gdb/1999-10/msg00058.html
 660
 661 I can reproduce this problem both with and without Tom's patch.  Has
 662 anyone seen this before?  Maybe have a solution for it hanging around?
 663 :)
 664
 665 There's a test case for this documented at:
 666
 667 when debugging threaded applications you get extra SIGTRAPs
 668 http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=9565
 669
 670 [There should be a GDB testcase - cagney]
 671
 672 --
 673
 674 GDB5 TOT on unixware 7
 675 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00119.html
 676
 677 Robert Lipe writes:
 678 > I just spun the top of tree of the GDB5 branch on UnixWare 7.  As a
 679 > practical matter, the current thread support is somewhat more annoying
 680 > than when GDB was thread-unaware.
 681
 682 --
 683
 684 Migrate qfThreadInfo packet -> qThreadInfo. (Andrew Cagney)
 685
 686 Add support for packet enable/disable commands with these thread
 687 packets.  General cleanup.
 688
 689 [PATCH] Document the ThreadInfo remote protocol queries
 690 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00832.html
 691
 692 [PATCH] "info threads" queries for remote.c
 693 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00831.html
 694
 695 --
 696
 697                         Language Support
 698                         ================
 699
 700 New languages come onto the scene all the time.
 701
 702 --
 703
 704 Re: Various C++ things
 705
 706 value_headof/value_from_vtable_info are worthless, and should be
 707 removed.  The one place in printcmd.c that uses it should use the RTTI
 708 functions.
 709
 710 RTTI for g++ should be using the typeinfo functions rather than the
 711 vtables.  The typeinfo functions are always at offset 4 from the
 712 beginning of the vtable, and are always right. The vtables will have
 713 weird names like E::VB sometimes.  The typeinfo function will always
 714 be "E type_info function", or somesuch.
 715
 716 value_virtual_fn_field needs to be fixed so there are no failures for
 717 virtual functions for C++ using g++.
 718
 719 Testsuite cases are the major priority right now for C++ support,
 720 since i have to make a lot of changes that could potentially break
 721 each other.
 722
 723 --
 724
 725 Add support for Modula3
 726
 727 Get DEC/Compaq to contribute their Modula-3 support.
 728
 729 --
 730
 731                 Remote Protocol Support
 732                 =======================
 733
 734 --
 735
 736 Remote protocol doco feedback.
 737
 738 Too much feedback to mention needs to be merged in (901660).  Search
 739 for the word ``remote''.
 740
 741
 742 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00023.html
 743 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00056.html
 744 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00382.html
 745
 746 --
 747
 748 GDB doesn't recover gracefully from remote protocol errors.
 749
 750 GDB wasn't checking for NAKs from the remote target.  Instead a NAK is
 751 ignored and a timeout is required before GDB retries.  A pre-cursor to
 752 fixing this this is making GDB's remote protocol packet more robust.
 753
 754 While downloading to a remote protocol target, gdb ignores packet
 755 errors in so far as it will continue to download with chunk N+1 even
 756 if chunk N was not correctly sent.  This causes gdb.base/remote.exp to
 757 take a painfully long time to run.  As a PS that test needs to be
 758 fixed so that it builds on 16 bit machines.
 759
 760 --
 761
 762 Add the cycle step command.
 763
 764 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00237.html
 765
 766 --
 767
 768 Resolve how to scale things to support very large packets.
 769
 770 --
 771
 772 Resolve how to handle a target that changes things like its endianess
 773 on the fly - should it be returned in the ``T'' packet?
 774
 775 Underlying problem is that the register file is target endian.  If the
 776 target endianess changes gdb doesn't know.
 777
 778 --
 779
 780 Rename read_register{,_pid}() to read_unsigned_register{,_pid}().
 781
 782 --
 783
 784                         Symbol Support
 785                         ==============
 786
 787 If / when GDB starts to support the debugging of multi-processor
 788 (rather than multi-thread) applications the symtab code will need to
 789 be updated a little so that several independant symbol tables are
 790 active at a given time.
 791
 792 The other interesting change is a clarification of the exact meaning
 793 of CORE_ADDR and that has had consequences for a few targets (that
 794 were abusing that data type).
 795
 796 --
 797
 798 Investiagate ways of reducing memory.
 799
 800 --
 801
 802 Investigate ways of improving load time.
 803
 804 --
 805
 806 Get the d10v to use POINTER_TO_ADDRESS and ADDRESS_TO_POINTER.
 807
 808 Consequence of recent symtab clarification.  No marks for figuring out
 809 who maintains the d10v.
 810
 811 --
 812
 813 Get the MIPS to correctly sign extend all address <-> pointer
 814 conversions.
 815
 816 Consequence of recent symtab clarification.  No marks for figuring out
 817 who maintains the MIPS.
 818
 819 --
 820
 821 GDB truncates 64 bit enums.
 822
 823 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00290.html
 824
 825 --
 826
 827                         Testsuite Support
 828                         =================
 829
 830 There are never to many testcases.
 831
 832 --
 833
 834 Better thread testsuite.
 835
 836 --
 837
 838 Better C++ testsuite.
 839
 840 --
 841
 842 Look at adding a GDB specific testsuite directory so that white box
 843 tests of key internals can be added (eg ui_file).
 844
 845 --
 846
 847 Separate out tests that involve the floating point (FP).
 848
 849 (Something for people brining up new targets).  FP and non-fp tests
 850 are combined.  I think there should be set of basic tests that
 851 exercise pure integer support and then a more expanded set that
 852 exercise FP and FP/integer interactions.
 853
 854 As an example, the MIPS, for n32 as problems with passing FP's and
 855 structs.  Since most inferior call tests include FP it is difficult to
 856 determine of the integer tests are ok.
 857
 858 --
 859
 860                 Architectural Changes: General
 861                 ==============================
 862
 863 These are harder than simple cleanups / fixes and, consequently
 864 involve more work.  Typically an Architectural Change will be broken
 865 down into a more digestible set of cleanups and fixes.
 866
 867 --
 868
 869 Cleanup software single step.
 870
 871 At present many targets implement software single step by directly
 872 blatting memory (see rs6000-tdep.c).  Those targets should register
 873 the applicable breakpoints using the breakpoint framework.  Perhaphs a
 874 new internal breakpoint class ``step'' is needed.
 875
 876 --
 877
 878 Replace READ_FP() with FRAME_HANDLE().
 879
 880 READ_FP() is a hangover from the days of the vax when the ABI really
 881 did have a frame pointer register.  Modern architectures typically
 882 construct a virtual frame-handle from the stack pointer and various
 883 other bits of string.
 884
 885 Unfortunatly GDB still treats this synthetic FP register as though it
 886 is real.  That in turn really confuses users (arm and ``print $fp'' VS
 887 ``info registers fp'').  The synthetic FP should be separated out of
 888 the true register set presented to the user.
 889
 890 --
 891
 892 Register Cache Cleanup (below from Andrew Cagney)
 893
 894 I would depict the current register architecture as something like:
 895
 896         High GDB  -->   Low GDB
 897             |              |
 898            \|/            \|/
 899             --- REG NR -----
 900                    |
 901     register + REGISTER_BYTE(reg_nr)
 902                    |
 903                   \|/
 904         -------------------------
 905         |  extern register[]    |
 906         -------------------------
 907
 908 where neither the high (valops.c et.al.) or low gdb (*-tdep.c) are
 909 really clear on what mechanisms they should be using to manipulate that
 910 buffer.   Further, much code assumes, dangerously, that registers are
 911 contigious.  Having got mips-tdep.c to support multiple ABIs, believe
 912 me, that is a bad assumption.  Finally, that register cache layout is
 913 determined by the current remote/local target and _not_ the less
 914 specific target ISA.  In fact, in many cases it is determined by the
 915 somewhat arbitrary layout of the [gG] packets!
 916
 917
 918 How I would like the register file to work is more like:
 919
 920
 921           High GDB
 922              |
 923             \|/
 924        pseudo reg-nr
 925              |
 926         map pseudo <->
 927         random cache
 928         bytes
 929              |
 930             \|/
 931         ------------
 932         | register |
 933         |  cache   |
 934         ------------
 935             /|\
 936              |
 937         map random cache
 938         bytes to target
 939         dependant i-face
 940             /|\
 941              |
 942        target dependant
 943         such as [gG] packet
 944         or ptrace buffer
 945
 946 The main objectives being:
 947
 948         o       a clear separation between the low
 949                 level target and the high level GDB
 950
 951         o       a mechanism that solves the general
 952                 problem of register aliases, overlaps
 953                 etc instead of treating them as optional
 954                 extras that can be wedged in as an after
 955                 thought (that is a reasonable description
 956                 of the current code).
 957
 958                 Identify then solve the hard case and the
 959                 rest just falls out.  GDB solved the easy
 960                 case and then tried to ignore the real
 961                 world :-)
 962
 963         o       a removal of the assumption that the
 964                 mapping between the register cache
 965                 and virtual registers is largely static.
 966                 If you flip the USR/SSR stack register
 967                 select bit in the status-register then
 968                 the corresponding stack registers should
 969                 reflect the change.
 970
 971         o       a mechanism that clearly separates the
 972                 gdb internal register cache from any
 973                 target (not architecture) dependant
 974                 specifics such as [gG] packets.
 975
 976 Of course, like anything, it sounds good in theory.  In reality, it
 977 would have to contend with many<->many relationships at both the
 978 virt<->cache and cache<->target level.  For instance:
 979
 980         virt<->cache
 981                 Modifying an mmx register may involve
 982                 scattering values across both FP and
 983                 mmpx specific parts of a buffer
 984
 985         cache<->target
 986                 When writing back a SP it may need to
 987                 both be written to both SP and USP.
 988
 989
 990 Hmm,
 991
 992 Rather than let this like the last time it was discussed, just slip, I'm
 993 first going to add this e-mail (+ references) to TODO.  I'd then like to
 994 sketch out a broad strategy I think could get us there.
 995
 996
 997 First thing I'd suggest is separating out the ``extern registers[]''
 998 code so that we can at least identify what is using it.  At present
 999 things are scattered across many files.  That way we can at least
1000 pretend that there is a cache instead of a global array :-)
1001
1002 I'd then suggest someone putting up a proposal for the pseudo-reg /
1003 high-level side interface so that code can be adopted to it.  For old
1004 code, initially a blanket rename of write_register_bytes() to
1005 deprecated_write_register_bytes() would help.
1006
1007 Following that would, finaly be the corresponding changes to the target.
1008
1009 --
1010
1011 Check that GDB can handle all BFD architectures (Andrew Cagney)
1012
1013 There should be a test that checks that BFD/GDB are in sync with
1014 regard to architecture changes.  Something like a test that first
1015 queries GDB for all supported architectures and then feeds each back
1016 to GDB..  Anyone interested in learning how to write tests?  :-)
1017
1018 --
1019
1020                 Architectural Change: Multi-arch et al.
1021                 =======================================
1022
1023 The long term objective is to remove all assumptions that there is a
1024 single target with a single address space with a single instruction
1025 set architecture and single application binary interface.
1026
1027 This is an ongoing effort.  The first milestone is to enable
1028 ``multi-arch'' where by all architectural decisions are made at
1029 runtime.
1030
1031 It should be noted that ``gdbarch'' is really ``gdbabi'' and
1032 ``gdbisa''.  Once things are multi-arched breaking that down correctly
1033 will become much easier.
1034
1035 --
1036
1037 GDBARCH cleanup (Andrew Cagney)
1038
1039 The non-generated parts of gdbarch.{sh,h,c} should be separated out
1040 into arch-utils.[hc].
1041
1042 Document that gdbarch_init_ftype could easily fail because it didn't
1043 identify an architecture.
1044
1045 --
1046
1047 Fix BELIEVE_PPC_PROMOTION.  Change it to BELIEVE_PPC_PROMOTION_P?
1048
1049 At present there is still #ifdef BELIEVE_PPC_PROMOTION code in the
1050 symtab file.
1051
1052 --
1053
1054 Fix target_signal_from_host() etc.
1055
1056 The name is wrong for starters.  ``target_signal'' should probably be
1057 ``gdb_signal''.  ``from_host'' should be ``from_target_signal''.
1058 After that it needs to be multi-arched and made independant of any
1059 host signal numbering.
1060
1061 --
1062
1063 Update ALPHA so that it uses ``struct frame_extra_info'' instead of
1064 EXTRA_FRAME_INFO.
1065
1066 This is a barrier to replacing mips_extra_func_info with something
1067 that works with multi-arch.
1068
1069 --
1070
1071 Multi-arch  mips_extra_func_info.
1072
1073 This first needs the alpha to be updated so that it uses ``struct
1074 frame_extra_info''.
1075
1076 --
1077
1078 Rationalize TARGET_SINGLE_FORMAT and TARGET_SINGLE_BIT et al.
1079
1080 Surely one of them is redundant.
1081
1082 --
1083
1084 Convert ALL architectures to MULTI-ARCH.
1085
1086 --
1087
1088 Select the initial multi-arch ISA / ABI based on --target or similar.
1089
1090 At present the default is based on what ever is first in the BFD
1091 archures table.  It should be determined based on the ``--target=...''
1092 name.
1093
1094 --
1095
1096 Make MIPS pure multi-arch.
1097
1098 It is only at the multi-arch enabled stage.
1099
1100 --
1101
1102 Truly multi-arch.
1103
1104 Enable the code to recognize --enable-targets=.... like BINUTILS does.
1105
1106 Can the tm.h and nm.h files be eliminated by multi-arch.
1107
1108 --
1109
1110         Architectural Change: MI, LIBGDB and scripting languages
1111         ========================================================
1112
1113 See also architectural changes related to the event loop.  LIBGDB
1114 can't be finished until there is a generic event loop being used by
1115 all targets.
1116
1117 The long term objective is it to be possible to integrate GDB into
1118 scripting languages.
1119
1120 --
1121
1122 Implement generic ``(gdb) commmand > file''
1123
1124 Once everything is going through ui_file it should be come fairly
1125 easy.
1126
1127 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00104.html
1128
1129 --
1130
1131 Replace gdb_stdtarg with gdb_targout (and possibly gdb_targerr).
1132
1133 gdb_stdtarg is easily confused with gdb_stdarg.
1134
1135 --
1136
1137 Extra ui_file methods - dump.
1138
1139 Very useful for whitebox testing.
1140
1141 --
1142
1143 Eliminate error_begin().
1144
1145 With ui_file, there is no need for the statefull error_begin ()
1146 function.
1147
1148 --
1149
1150 Send normal output to gdb_stdout.
1151 Send error messages to gdb_stderror.
1152 Send debug and log output log gdb_stdlog.
1153
1154 GDB still contains many cases where (f)printf or printf_filtered () is
1155 used when it should be sending the messages to gdb_stderror or
1156 gdb_stdlog.  The thought of #defining printf to something has crossed
1157 peoples minds ;-)
1158
1159 --
1160
1161 Re-do GDB's output pager.
1162
1163 GDB's output pager still relies on people correctly using *_filtered
1164 for gdb_stdout and *_unfiltered for gdb_stdlog / gdb_stderr.
1165 Hopefully, with all normal output going to gdb_stdout, the pager can
1166 just look at the ui_file that the output is on and then use that to
1167 decide what to do about paging.  Sounds good in theory.
1168
1169 --
1170
1171 Check/cleanup MI documentation.
1172
1173 The list of commands specified in the documentation needs to be
1174 checked against the mi-cmds.c table in a mechanical way (so that they
1175 two can be kept up-to-date).
1176
1177 --
1178
1179 Convert MI into libgdb
1180
1181 MI provides a text interface into what should be many of the libgdb
1182 functions.  The implementation of those functions should be separated
1183 into the MI interface and the functions proper.  Those functions being
1184 moved to gdb/lib say.
1185
1186 --
1187
1188 Create libgdb.h
1189
1190 The first part can already be found in defs.h.
1191
1192 --
1193
1194 MI's input does not use buffering.
1195
1196 At present the MI interface reads raw characters of from an unbuffered
1197 FD.  This is to avoid several nasty buffer/race conditions.  That code
1198 should be changed so that it registers its self with the event loop
1199 (on the input FD) and then push commands up to MI as they arrive.
1200
1201 The serial code already does this.
1202
1203 --
1204
1205 Make MI interface accessable from existing CLI.
1206
1207 --
1208
1209 Add a breakpoint-edit command to MI.
1210
1211 It would be similar to MI's breakpoint create but would apply to an
1212 existing breakpoint.  It saves the need to delete/create breakpoints
1213 when ever they are changed.
1214
1215 --
1216
1217 Add directory path to MI breakpoint.
1218
1219 That way the GUI's task of finding the file within which the
1220 breakpoint was set is simplified.
1221
1222 --
1223
1224 Add a mechanism to reject certain expression classes to MI
1225
1226 There are situtations where you don't want GDB's expression
1227 parser/evaluator to perform inferior function calls or variable
1228 assignments.  A way of restricting the expression parser so that such
1229 operations are not accepted would be very helpful.
1230
1231 --
1232
1233 Remove sideffects from libgdb breakpoint create function.
1234
1235 The user can use the CLI to create a breakpoint with partial
1236 information - no file (gdb would use the file from the last
1237 breakpoint).
1238
1239 The libgdb interface currently affects that environment which can lead
1240 to confusion when a user is setting breakpoints via both the MI and
1241 the CLI.
1242
1243 This is also a good example of how getting the CLI ``right'' will be
1244 hard.
1245
1246 --
1247
1248 Move gdb_lasterr to ui_out?
1249
1250 The way GDB throws errors and records them needs a re-think.  ui_out
1251 handles the correct output well.  It doesn't resolve what to do with
1252 output / error-messages when things go wrong.
1253
1254 --
1255
1256 do_setshow_command contains a 1024 byte buffer.
1257
1258 The function assumes that there will never be any more than 1024 bytes
1259 of enum.  It should use mem_file.
1260
1261 --
1262
1263 Should struct cmd_list_element . completer take the command as an
1264 argument?
1265
1266 --
1267
1268 Should the bulk of top.c:line_completion_function() be moved to
1269 command.[hc]?  complete_on_cmdlist() and complete_on_enums() could
1270 then be made private.
1271
1272 --
1273
1274 top.c (execute_command): Should a command being valid when the target
1275 is running be made an attribute (predicate) to the command rather than
1276 an explicit set of tests.
1277
1278 --
1279
1280 top.c (execute_command): Should the bulk of this function be moved
1281 into command.[hc] so that top.c doesn't grub around in the command
1282 internals?
1283
1284 --
1285
1286                 Architectural Change: Async
1287                 ===========================
1288
1289 While GDB uses an event loop when prompting the user for input.  That
1290 event loop is not exploited by targets when they allow the target
1291 program to continue.  Typically targets still block in (target_wait())
1292 until the program again halts.
1293
1294 The closest a target comes to supporting full asynchronous mode are
1295 the remote targets ``async'' and ``extended-async''.
1296
1297 --
1298
1299 Asynchronous expression evaluator
1300
1301 Inferior function calls hang GDB.
1302
1303 --
1304
1305 Fix implementation of  ``target xxx''.
1306
1307 At present when the user specifies ``target xxxx'', the CLI maps that
1308 directly onto a target open method.  It is then assumed that the
1309 target open method should do all sorts of complicated things as this
1310 is the only chance it has.  Check how the various remote targets
1311 duplicate the target operations.  Check also how the various targets
1312 behave differently for purely arbitrary reasons.
1313
1314 What should happen is that ``target xxxx'' should call a generic
1315 ``target'' function and that should then co-ordinate the opening of
1316 ``xxxx''.  This becomes especially important when you're trying to
1317 open an asynchronous target that may need to perform background tasks
1318 as part of the ``attach'' phase.
1319
1320 Unfortunatly, due to limitations in the old/creaking command.h
1321 interface, that isn't possible.  The function being called isn't told
1322 of the ``xxx'' or any other context information.
1323
1324 Consequently a precursor to fixing ``target xxxx'' is to clean up the
1325 CLI code so that it passes to the callback function (attatched to a
1326 command) useful information such as the actual command and a context
1327 for that command.  Other changes such as making ``struct command''
1328 opaque may also help.
1329
1330 See also:
1331 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
1332
1333 --
1334
1335 Make "target xxx" command interruptible.
1336
1337 As things become async this becomes possible.  A target would start
1338 the connect and then return control to the event loop.  A cntrl-c
1339 would notify the target that the operation is to be abandoned and the
1340 target code could respond.
1341
1342 --
1343
1344 Add a "suspend" subcommand of the "continue" command to suspend gdb
1345 while continuing  execution of the subprocess.  Useful when you are
1346 debugging servers and you want to dodge out and initiate a connection
1347 to a server running under gdb.
1348
1349 [hey async!!]
1350
1351 --
1352
1353                         TODO FAQ
1354                         ========
1355
1356 Frequently requested but not approved requests.
1357
1358 --
1359
1360 Eliminate unused argument warnings using ATTRIBUTE_UNUSED.
1361
1362 The benefits on this one are thought to be marginal - GDBs design
1363 means that unused parameters are very common.  GCC 3.0 will also
1364 include the option -Wno-unused-parameter which means that ``-Wall
1365 -Wno-unused-parameters -Werror'' can be specified.
1366
1367 --
1368
1369
1370
1371                         Legacy Wish List
1372                         ================
1373
1374 This list is not up to date, and opinions vary about the importance or
1375 even desirability of some of the items.  If you do fix something, it
1376 always pays to check the below.
1377
1378 --
1379
1380 @c This does not work (yet if ever).  FIXME.
1381 @c @item --parse=@var{lang} @dots{}
1382 @c Configure the @value{GDBN} expression parser to parse the listed languages.
1383 @c @samp{all} configures @value{GDBN} for all supported languages.  To get a
1384 @c list of all supported languages, omit the argument.  Without this
1385 @c option, @value{GDBN} is configured to parse all supported languages.
1386
1387 --
1388
1389 START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that
1390 is its default value.  Clean this up.
1391
1392 --
1393
1394 It should be possible to use symbols from shared libraries before we know
1395 exactly where the libraries will be loaded.  E.g. "b perror" before running
1396 the program.  This could maybe be done as an extension of the "breakpoint
1397 re-evaluation" after new symbols are loaded.
1398
1399 --
1400
1401 Make single_step() insert and remove breakpoints in one operation.
1402
1403 [If this is talking about having single_step() insert the breakpoints,
1404 run the target then pull the breakpoints then it is wrong.  The
1405 function has to return as control has to eventually be passed back to
1406 the main event loop.]
1407
1408 --
1409
1410 Speed up single stepping by avoiding extraneous ptrace calls.
1411
1412 --
1413
1414 Speed up single stepping by not inserting and removing breakpoints
1415 each time the inferior starts and stops.
1416
1417 Breakpoints should not be inserted and deleted all the time.  Only the
1418 one(s) there should be removed when we have to step over one.  Support
1419 breakpoints that don't have to be removed to step over them.
1420
1421 [this has resulted in numerous debates.  The issue isn't clear cut]
1422
1423 --
1424
1425 Provide "voodoo" debugging of core files.  This creates a zombie
1426 process as a child of the debugger, and loads it up with the data,
1427 stack, and regs of the core file.  This allows you to call functions
1428 in the executable, to manipulate the data in the core file.
1429
1430 [you wish]
1431
1432 --
1433
1434 GDB reopens the source file on every line, as you "next" through it.
1435
1436 [still true? I've a memory of this being fixed]
1437
1438 --
1439
1440 Perhaps "i source" should take an argument like that of "list".
1441
1442 --
1443
1444 Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if
1445 it matches the source line indicated.
1446
1447 --
1448
1449 The prompt at end of screen should accept space as well as CR.
1450
1451 --
1452
1453 Backtrace should point out what the currently selected frame is, in
1454 its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar,
1455 ...)" rather than "#3 foo (bar, ...)".
1456
1457 --
1458
1459 "i program" should work for core files, and display more info, like what
1460 actually caused it to die.
1461
1462 --
1463
1464 "x/10i" should shorten the long name, if any, on subsequent lines.
1465
1466 --
1467
1468 "next" over a function that longjumps, never stops until next time you happen
1469 to get to that spot by accident.  E.g. "n" over execute_command which has
1470 an error.
1471
1472 --
1473
1474 "set zeroprint off", don't bother printing members of structs which
1475 are entirely zero.  Useful for those big structs with few useful
1476 members.
1477
1478 --
1479
1480 GDB does four ioctl's for every command, probably switching terminal modes
1481 to/from inferior or for readline or something.
1482
1483 --
1484
1485 terminal_ours versus terminal_inferior: cache state.  Switch should be a noop
1486 if the state is the same, too.
1487
1488 --
1489
1490 "i frame" shows wrong "arglist at" location, doesn't show where the args
1491 should be found, only their actual values.
1492
1493 --
1494
1495 There should be a way for "set" commands to validate the new setting
1496 before it takes effect.
1497
1498 --
1499
1500 "ena d" is ambiguous, why?  "ena delete" seems to think it is a command!
1501
1502 --
1503
1504 i line VAR produces "Line number not known for symbol ``var''.".  I
1505 thought we were stashing that info now!
1506
1507 --
1508
1509 We should be able to write to random files at hex offsets like adb.
1510
1511 --
1512
1513 [elena - delete this]
1514
1515 Handle add_file with separate text, data, and bss addresses.  Maybe
1516 handle separate addresses for each segment in the object file?
1517
1518 --
1519
1520 [Jimb/Elena delete this one]
1521
1522 Handle free_named_symtab to cope with multiply-loaded object files
1523 in a dynamic linking environment.  Should remember the last copy loaded,
1524 but not get too snowed if it finds references to the older copy.
1525
1526 --
1527
1528 [elena delete this also]
1529
1530 Remove all references to:
1531         text_offset
1532         data_offset
1533         text_data_start
1534         text_end
1535         exec_data_offset
1536         ...
1537 now that we have BFD.  All remaining are in machine dependent files.
1538
1539 --
1540
1541 Re-organize help categories into things that tend to fit on a screen
1542 and hang together.
1543
1544 --
1545
1546 Add in commands like ADB's for searching for patterns, etc.  We should
1547 be able to examine and patch raw unsymboled binaries as well in gdb as
1548 we can in adb.  (E.g. increase the timeout in /bin/login without source).
1549
1550 [actually, add ADB interface :-]
1551
1552 --
1553
1554 When doing "step" or "next", if a few lines of source are skipped between
1555 the previous line and the current one, print those lines, not just the
1556 last line of a multiline statement.
1557
1558 --
1559
1560 Handling of "&" address-of operator needs some serious overhaul
1561 for ANSI C and consistency on arrays and functions.
1562   For "float point[15];":
1563 ptype &point[4]   ==> Attempt to take address of non-lvalue.
1564   For "char *malloc();":
1565 ptype malloc      ==> "char *()";  should be same as
1566 ptype &malloc     ==> "char *(*)()"
1567 call printf ("%x\n", malloc)   ==> weird value, should be same as
1568 call printf ("%x\n", &malloc)  ==> correct value
1569
1570 --
1571
1572 Fix dbxread.c symbol reading in the presence of interrupts.  It
1573 currently leaves a cleanup to blow away the entire symbol table when a
1574 QUIT occurs.  (What's wrong with that? -kingdon, 28 Oct 1993).
1575
1576 [I suspect that the grype was that, on a slow system, you might want
1577 to cntrl-c and get just half the symbols and then load the rest later
1578 - scary to be honest]
1579
1580 --
1581
1582 Mipsread.c reads include files depth-first, because the dependencies
1583 in the psymtabs are way too inclusive (it seems to me).  Figure out what
1584 really depends on what, to avoid recursing 20 or 30 times while reading
1585 real symtabs.
1586
1587 --
1588
1589 value_add() should be subtracting the lower bound of arrays, if known,
1590 and possibly checking against the upper bound for error reporting.
1591
1592 --
1593
1594 When listing source lines, check for a preceding \n, to verify that
1595 the file hasn't changed out from under us.
1596
1597 [fixed by some other means I think.  That hack wouldn't actually work
1598 reliably - the file might move such that another \n appears. ]
1599
1600 --
1601
1602 Get all the remote systems (where the protocol allows it) to be able to
1603 stop the remote system when the GDB user types ^C (like remote.c
1604 does).  For ebmon, use ^Ak.
1605
1606 --
1607
1608 Possible feature: A version of the "disassemble" command which shows
1609 both source and assembly code ("set symbol-filename on" is a partial
1610 solution).
1611
1612 [has this been done?  It was certainly done for MI and GDBtk]
1613
1614 --
1615
1616 investigate "x/s 0" (right now stops early) (I think maybe GDB is
1617 using a 0 address for bad purposes internally).
1618
1619 --
1620
1621 Make "info path" and path_command work again (but independent of the
1622 environment either of gdb or that we'll pass to the inferior).
1623
1624 --
1625
1626 Make GDB understand the GCC feature for putting octal constants in
1627 enums.  Make it so overflow on an enum constant does not error_type
1628 the whole type.  Allow arbitrarily large enums with type attributes.
1629 Put all this stuff in the testsuite.
1630
1631 --
1632
1633 Make TYPE_CODE_ERROR with a non-zero TYPE_LENGTH more useful (print
1634 the value in hex; process type attributes).  Add this to the
1635 testsuite.  This way future compilers can add new types and old
1636 versions of GDB can do something halfway reasonable.
1637
1638 --
1639
1640 Fix mdebugread.c:parse_type to do fundamental types right (see
1641 rs6000_builtin_type in stabsread.c for what "right" is--the point is
1642 that the debug format fixes the sizes of these things and it shouldn't
1643 depend on stuff like TARGET_PTR_BIT and so on.  For mdebug, there seem
1644 to be separate bt* codes for 64 bit and 32 bit things, and GDB should
1645 be aware of that).  Also use a switch statement for clarity and speed.
1646
1647 --
1648
1649 Investigate adding symbols in target_load--some targets do, some
1650 don't.
1651
1652 --
1653
1654 Put dirname in psymtabs and change lookup*symtab to use dirname (so
1655 /foo/bar.c works whether compiled by cc /foo/bar.c, or cd /foo; cc
1656 bar.c).
1657
1658 --
1659
1660 Merge xcoffread.c and coffread.c.  Use breakpoint_re_set instead of
1661 fixup_breakpoints.
1662
1663 --
1664
1665 Make a watchpoint which contains a function call an error (it is
1666 broken now, making it work is probably not worth the effort).
1667
1668 --
1669
1670 New test case based on weird.exp but in which type numbers are not
1671 renumbered (thus multiply defining a type).  This currently causes an
1672 infinite loop on "p v_comb".
1673
1674 --
1675
1676 [Hey! Hint Hint Delete Delete!!!]
1677
1678 Fix 386 floating point so that floating point registers are real
1679 registers (but code can deal at run-time if they are missing, like
1680 mips and 68k).  This would clean up "info float" and related stuff.
1681
1682 --
1683
1684 gcc -g -c enummask.c then gdb enummask.o, then "p v".  GDB complains
1685 about not being able to access memory location 0.
1686
1687 -------------------- enummask.c
1688 enum mask
1689 {
1690   ANIMAL = 0,
1691   VEGETABLE = 1,
1692   MINERAL = 2,
1693   BASIC_CATEGORY = 3,
1694
1695   WHITE = 0,
1696   BLUE = 4,
1697   GREEN = 8,
1698   BLACK = 0xc,
1699   COLOR = 0xc,
1700
1701   ALIVE = 0x10,
1702
1703   LARGE = 0x20
1704 } v;
1705
1706 --
1707
1708 If try to modify value in file with "set write off" should give
1709 appropriate error not "cannot access memory at address 0x65e0".
1710
1711 --
1712
1713 Allow core file without exec file on RS/6000.
1714
1715 --
1716
1717 Make sure "shell" with no arguments works right on DOS.
1718
1719 --
1720
1721 Make gdb.ini (as well as .gdbinit) be checked on all platforms, so
1722 the same directory can be NFS-mounted on unix or DOS, and work the
1723 same way.
1724
1725 --
1726
1727 [Is this another delete???]
1728
1729 Get SECT_OFF_TEXT stuff out of objfile_relocate (might be needed to
1730 get RS/6000 to work right, might not be immediately relevant).
1731
1732 --
1733
1734 Work out some kind of way to allow running the inferior to be done as
1735 a sub-execution of, eg. breakpoint command lists.  Currently running
1736 the inferior interupts any command list execution.  This would require
1737 some rewriting of wait_for_inferior & friends, and hence should
1738 probably be done in concert with the above.
1739
1740 --
1741
1742 Add function arguments to gdb user defined functions.
1743
1744 --
1745
1746 Add convenience variables that refer to exec file, symbol file,
1747 selected frame source file, selected frame function, selected frame
1748 line number, etc.
1749
1750 --
1751
1752 Modify the handling of symbols grouped through BINCL/EINCL stabs to
1753 allocate a partial symtab for each BINCL/EINCL grouping.  This will
1754 seriously decrease the size of inter-psymtab dependencies and hence
1755 lessen the amount that needs to be read in when a new source file is
1756 accessed.
1757
1758 --
1759
1760 Add a command for searching memory, a la adb.  It specifies size,
1761 mask, value, start address.  ADB searches until it finds it or hits
1762 an error (or is interrupted).
1763
1764 --
1765
1766 Remove the range and type checking code and documentation, if not
1767 going to implement.
1768
1769 # Local Variables:
1770 # mode: text
1771 # End: