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 Hardware watchpint problems on x86 OSes, including Linux:
  16
  17 1. Delete/disable hardware watchpoints should free hardware debug
  18 registers.
  19 2. Watch for different values on a viariable with one hardware debug
  20 register.
  21
  22 According to Eli Zaretskii <eliz@delorie.com>:
  23
  24 These are not GDB/ia32 issues per se: the above features are all
  25 implemented in the DJGPP port of GDB and work in v5.0.  Every
  26 x86-based target should be able to lift the relevant parts of
  27 go32-nat.c and use them almost verbatim.  You get debug register
  28 sharing through reference counts, and the ability to watch large
  29 regions (up to 16 bytes) using multiple registers.  (The required
  30 infrastructure in high-level GDB application code, mostly in
  31 breakpoint.c, is also working since v5.0.)
  32
  33 --
  34
  35 RFD: infrun.c: No bpstat_stop_status call after proceed over break?
  36 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00665.html
  37
  38 GDB misses watchpoint triggers after proceeding over a breakpoint on
  39 x86 targets.
  40
  41 --
  42
  43 x86 linux GDB and SIGALRM (???)
  44 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00803.html
  45
  46 This problem has been fixed, but a regression test still needs to be
  47 added to the testsuite:
  48 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00309.html
  49
  50 Mark
  51
  52 --
  53
  54 Can't build IRIX -> arm GDB.
  55 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00356.html
  56
  57 David Whedon writes:
  58 > Now I'm building for an embedded arm target.  If there is a way of turning
  59 > remote-rdi off, I couldn't find it. It looks like it gets built by default
  60 > in gdb/configure.tgt(line 58) Anyway, the build dies in
  61 > gdb/rdi-share/unixcomm.c.  SERPORT1 et. al. never get defined because we
  62 > aren't one of the architectures supported.
  63
  64 --
  65
  66 Problem with weak functions
  67 http://sourceware.cygnus.com/ml/gdb/2000-05/msg00060.html
  68
  69 Dan Nicolaescu writes:
  70 > It seems that gdb-4.95.1  does not display correctly the function when
  71 > stoping in weak functions.
  72 >
  73 > It stops in a function that is defined as weak, not in the function
  74 > that is actualy run...
  75
  76 --
  77
  78 GDB 5.0 doesn't work on Linux/SPARC
  79
  80 --
  81
  82 Thread support.  Right now, as soon as a thread finishes and exits,
  83 you're hosed.  This problem is reported once a week or so.
  84
  85 --
  86
  87 Wow, three bug reports for the same problem in one day!  We should
  88 probably make fixing this a real priority :-).
  89
  90 Anyway, thanks for reporting.
  91
  92 The following patch will fix the problems with setting breakpoints in
  93 dynamically loaded objects:
  94
  95    http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00230.html
  96
  97 This patch isn't checked in yet (ping Michael/JimB), but I hope this
  98 will be in the next GDB release.
  99
 100 There should really be a test in the testsuite for this problem, since
 101 it keeps coming up :-(.  Any volunteers?
 102
 103 Mark
 104
 105 --
 106
 107 Re: GDB 5.0.1?
 108 http://sources.redhat.com/ml/gdb/2000-07/msg00038.html
 109
 110 Is the Solaris 8 x86 problem fixed?  When you configure it, configure
 111 incorrectly determines that I have no curses.h.  This causes mucho
 112 compilation errors later on.
 113
 114 Simply editing the config.h to define CURSES_H fixes the problem, and
 115 then the build works fine.
 116
 117 The status for this problem:
 118
 119 Solaris 8 x86 (PIII-560)
 120 gcc 2.95.2
 121
 122 I had the same problem with several of the snapshots shortly before
 123 5.0 became official, and 5.0 has the same problem.
 124
 125 I sent some mail in about it long ago, and never saw a reply.
 126
 127 I haven't had time to figure it out myself, especially since I get all
 128 confused trying to figure out what configure does, I was happy to find
 129 the workaround.
 130
 131 Mike
 132
 133 --
 134
 135                 GDB 5.1 - New features
 136                 ======================
 137
 138 The following new features should be included in 5.1.
 139
 140 --
 141
 142 Enable MI by default.  Old code can be deleted after 5.1 is out.
 143
 144 --
 145
 146 Pascal (Pierre Muller, David Taylor)
 147
 148 Pierre Muller has contributed patches for adding Pascal Language
 149 support to GDB.
 150
 151 2 pascal language patches inserted in database
 152 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00521.html
 153
 154 Indent -gnu ?
 155 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00496.html
 156
 157 --
 158
 159 Java (Anthony Green, David Taylor)
 160
 161 Anthony Green has a number of Java patches that did not make it into
 162 the 5.0 release.  The first two are in cvs now, but the third needs
 163 some fixing up before it can go in.
 164
 165 Patch: java tests
 166 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00512.html
 167
 168 Patch: java booleans
 169 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00515.html
 170
 171 Patch: handle N_MAIN stab
 172 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00527.html
 173
 174 --
 175
 176 [Comming...]
 177
 178 Modify gdb to work correctly with Pascal.
 179
 180 --
 181
 182 Revised UDP support (was: Re: [Fwd: [patch] UDP transport support])
 183 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00000.html
 184
 185 (Broken) support for GDB's remote protocol across UDP is to be
 186 included in the follow-on release.
 187
 188 It should be noted that UDP can only work when the [Gg] packet fits in
 189 a single UDP packet.
 190
 191 There is also much debate over the merit of this.
 192
 193 --
 194
 195                 GDB 5.1 - Cleanups
 196                 ==================
 197
 198 The following code cleanups will hopefully be applied to GDB 5.1.
 199
 200 --
 201
 202 Delete macro TARGET_BYTE_ORDER_SELECTABLE.
 203
 204 Patches in the database.
 205
 206 --
 207
 208 Fix copyright notices.
 209
 210 Turns out that ``1998-2000'' isn't considered valid :-(
 211
 212 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00467.html
 213
 214 --
 215
 216 Purge PARAMS.
 217
 218 Eliminate all uses of PARAMS in GDB's source code.
 219
 220 --
 221
 222 printcmd.c (print_address_numeric):
 223
 224 NOTE: This assumes that the significant address information is kept in
 225 the least significant bits of ADDR - the upper bits were either zero
 226 or sign extended.  Should ADDRESS_TO_POINTER() or some
 227 ADDRESS_TO_PRINTABLE() be used to do the conversion?
 228
 229 --
 230
 231 Compiler warnings.
 232
 233 Eliminate all warnings for at least one host/target for the flags:
 234 -Wimplicit -Wreturn-type -Wcomment -Wtrigraphs -Wformat -Wparentheses
 235 -Wpointer-arith -Wuninitialized
 236
 237 --
 238
 239 Follow through `make check' with --enable-shared.
 240
 241 When the srcware tree is configured with --enable-shared, the `expect'
 242 program won't run properly.  Jim Wilson found out gdb has a local hack
 243 to set LD_LIBRARY_PATH, but, AFAIK, no other project has been hacked
 244 similarly.
 245
 246 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00845.html
 247
 248 --
 249
 250                 GDB 5.2 - Fixes
 251                 ===============
 252
 253 --
 254
 255 Fix at least one thread bug.
 256
 257 --
 258
 259                 GDB 5.2 - New features
 260                 ======================
 261
 262 --
 263
 264 Objective C/C++ Support.  Bu hopefully sooner...
 265
 266 --
 267
 268                 GDB 5.2 - Cleanups
 269                 ==================
 270
 271 The following cleanups have been identified as part of GDB 5.2.
 272
 273 --
 274
 275 Eliminate more compiler warnings.
 276
 277 --
 278
 279 Restructure gdb directory tree so that it avoids any 8.3 and 14
 280 filename problems.
 281
 282 --
 283
 284 Convert GDB build process to AUTOMAKE.
 285
 286 See also sub-directory configure below.
 287
 288 The current convention is (kind of) to use $(<header>_h) in all
 289 dependency lists.  It isn't done in a consistent way.
 290
 291 --
 292
 293                 Code Cleanups: General
 294                 ======================
 295
 296 The following are more general cleanups and fixes.  They are not tied
 297 to any specific release.
 298
 299 --
 300
 301 The BFD directory requires bug-fixed AUTOMAKE et.al.
 302
 303 AUTOMAKE 1.4 incorrectly set the TEXINPUTS environment variable.  It
 304 contained the full path to texinfo.tex when it should have only
 305 contained the directory.  The bug has been fixed in the current
 306 AUTOMAKE sources.  Automake snapshots can be found in:
 307         ftp://sourceware.cygnus.com/pub/gdb/snapshots
 308 and     ftp://sourceware.cygnus.com/pub/binutils
 309
 310 --
 311
 312 Find something better than DEFAULT_BFD_ARCH, DEFAULT_BFD_VEC to
 313 determine the default isa/byte-order.
 314
 315 --
 316
 317 Rely on BFD_BIG_ENDIAN and BFD_LITTLE_ENDIAN instead of host dependant
 318 BIG_ENDIAN and LITTLE_ENDIAN.
 319
 320 --
 321
 322 Eliminate more compiler warnings.
 323
 324 Of course there also needs to be the usual debate over which warnings
 325 are valid and how to best go about this.
 326
 327 One method: choose a single option; get agreement that it is
 328 reasonable; try it out to see if there isn't anything silly about it
 329 (-Wunused-parameters is an example of that) then incrementally hack
 330 away.
 331
 332 The other method is to enable all warnings and eliminate them from one
 333 file at a time.
 334
 335 --
 336
 337 Elimination of ``(catch_errors_ftype *) func''.
 338
 339 Like make_cleanup_func it isn't portable.
 340 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00791.html
 341 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00814.html
 342
 343 --
 344
 345 Nuke #define CONST_PTR.
 346
 347 --
 348
 349 Nuke USG define.
 350
 351 --
 352
 353 [PATCH/5] src/intl/Makefile.in:distclean additions
 354 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00363.html
 355
 356 Do not forget to merge the patch back into the trunk.
 357
 358 --
 359
 360 Rationalize the host-endian code (grep for HOST_BYTE_ORDER).
 361
 362 At present defs.h includes <endian.h> (which is linux specific) yet
 363 almost nothing depends on it.  Suggest "gdb_endian.h" which can also
 364 handle <machine/endian.h> and only include that where it is really
 365 needed.
 366
 367 --
 368
 369 Replace asprintf() calls with xasprintf() calls.
 370
 371 As with things like strdup() most calls to asprintf() don't check the
 372 return value.
 373
 374 --
 375
 376 Replace strsave() + mstrsave() with libiberty:xstrdup().
 377
 378 --
 379
 380 Replace savestring() with something from libiberty.
 381
 382 An xstrldup()? but that would have different semantics.
 383
 384 --
 385
 386 Rationalize use of floatformat_unknown in GDB sources.
 387
 388 Instead of defaulting to floatformat_unknown, should hosts/targets
 389 specify the value explicitly?
 390
 391 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
 392
 393 --
 394
 395 Add a ``name'' member to include/floatformat.h:struct floatformat.
 396 Print that name in gdbarch.c.
 397
 398 --
 399
 400 Sort out the harris mess in include/floatformat.h (it hardwires two
 401 different floating point formats).
 402
 403 --
 404
 405 See of the GDB local floatformat_do_doublest() and libiberty's
 406 floatformat_to_double (which was once GDB's ...) can be merged some
 407 how.
 408
 409 --
 410
 411 Eliminate mmalloc() from GDB.
 412
 413 Also eliminate it from defs.h.
 414
 415 --
 416
 417 Eliminate PTR.  ISO-C allows ``void *''.
 418
 419 --
 420
 421 Eliminate abort ().
 422
 423 GDB should never abort.  GDB should either throw ``error ()'' or
 424 ``internal_error ()''.  Better still GDB should naturally unwind with
 425 an error status.
 426
 427 --
 428
 429 Add __LINE__ and __FILE__ to internal_error().
 430
 431 --
 432
 433 GDB probably doesn't build on FreeBSD pre 2.2.x
 434 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00378.html
 435
 436 Fixes to get FreeBSD working on 2.2.x, 3.x and 4.x caused the code to
 437 suffer bit rot.
 438
 439 --
 440
 441 Deprecate "fg".  Apparently ``fg'' is actually continue.
 442
 443 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00417.html
 444
 445 --
 446
 447 Deprecate current use of ``floatformat_unknown''.
 448
 449 Require all targets to explicitly provide their float format instead
 450 of defaulting to floatformat unknown.  Doing the latter leads to nasty
 451 bugs.
 452
 453 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
 454
 455 --
 456
 457 Rationalize floatformat_to_double() vs floatformat_to_doublest().
 458
 459 Looks like GDB migrated floatformat_to_double() to libiberty but then
 460 turned around and created a ..._to_doublest() the latter containing
 461 several bug fixes.
 462
 463 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00472.html
 464
 465 --
 466
 467 Move floatformat_ia64_ext to libiberty/include floatformat.[ch].
 468
 469 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00466.html
 470
 471 --
 472
 473 The ``maintenance deprecate set endian big'' command doesn't notice
 474 that it is deprecating ``set endian'' and not ``set endian big'' (big
 475 is implemented using an enum).  Is anyone going to notice this?
 476
 477 --
 478
 479 When tab expanding something like ``set arch<tab>'' ignore the
 480 deprecated ``set archdebug'' and expand to ``set architecture''.
 481
 482 --
 483
 484 Eliminate ``arm_register_names[j] = (char *) regnames[j]'' and the
 485 like from arm-tdep.c.
 486
 487 --
 488
 489 Fix uses of ->function.cfunc = set_function().
 490
 491 The command.c code calls sfunc() when a set command.  Rather than
 492 change it suggest fixing the callback function so that it is more
 493 useful.  See:
 494
 495 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
 496
 497 See also ``Fix implementation of  ``target xxx''.'' below.
 498
 499 --
 500
 501 IRIX 3.x support is probably broken.
 502
 503 --
 504
 505 Delete sim/SIM_HAVE_BREAKPOINTS and gdb/SIM_HAS_BREAKPOINTS.
 506 http://sourceware.cygnus.com/ml/gdb-patches/2000-07/msg00042.html
 507
 508 Apart from the d30v, are there any sim/common simulators that make use
 509 of this?
 510
 511 A brief summary of what happended is that sim/common/sim-break.c was
 512 created as a good idea.  It turned out a better idea was to use
 513 SIM_SIGBREAK and have GDB pass back sim_resume (..., SIGBREAK).
 514
 515 --
 516
 517 Move remote_remove_hw_breakpoint, remote_insert_hw_breakpoint,
 518 remote_remove_watchpoint, remote_insert_watchpoint into target vector.
 519
 520 --
 521
 522                         New Features and Fixes
 523                         ======================
 524
 525 These are harder than cleanups but easier than work involving
 526 fundamental architectural change.
 527
 528 --
 529
 530 Add built-by, build-date, tm, xm, nm and anything else into gdb binary
 531 so that you can see how the GDB was created.
 532
 533 --
 534
 535 Add an "info bfd" command that displays supported object formats,
 536 similarly to objdump -i.
 537
 538 Is there a command already?
 539
 540 --
 541
 542 Fix ``I'm sorry, Dave, I can't do that.'' from symfile.c.
 543
 544 This requires internationalization.
 545
 546 --
 547
 548 Add support for:
 549
 550 (gdb) p fwprintf(stdout,L"%S\n", f)
 551 No symbol "L" in current context.
 552
 553 --
 554
 555 Cleanup configury support for optional sub-directories.
 556
 557 Check how GCC handles multiple front ends for an example of how things
 558 could work.  A tentative first step is to rationalize things so that
 559 all sub directories are handled in a fashion similar to gdb/mi.
 560
 561 See also automake above.
 562
 563 --
 564
 565 Add a transcript mechanism to GDB.
 566
 567 Such a mechanism might log all gdb input and output to a file in a
 568 form that would allow it to be replayed.  It could involve ``gdb
 569 --transcript=FILE'' or it could involve ``(gdb) transcript file''.
 570
 571 --
 572
 573 Can the xdep files be replaced by autoconf?
 574
 575 --
 576
 577 Document trace machinery
 578
 579 --
 580
 581 Document ui-out and ui-file.
 582
 583 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00121.html
 584
 585 --
 586
 587 Update texinfo.tex to latest?
 588
 589 --
 590
 591 Incorporate agentexpr.texi into gdb.texinfo
 592
 593 agentexpr.texi mostly describes the details of the byte code used for
 594 tracepoints, not the internals of the support for this in GDB.  So it
 595 looks like gdb.texinfo is a better place for this information.
 596
 597 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00566.html
 598
 599 --
 600
 601 Document overlay machinery.
 602
 603 --
 604
 605 ``(gdb) catch signal SIGNAL''
 606
 607 Overlaps with ``handle SIGNAL'' but the implied behavour is different.
 608 You can attach commands to a catch but not a handle.  A handle has a
 609 limited number of hardwired actions.
 610
 611 --
 612
 613 Get the TUI working on all platforms.
 614
 615 --
 616
 617 Add support for ``gdb --- PROGRAM ARGS ...''.
 618 Add support for ``gdb -cmd=...''
 619
 620 Along with many variations.  Check:
 621
 622 ????? for a full discussion.
 623
 624 for a discussion.
 625
 626 --
 627
 628 Implement ``(gdb) !ls''.
 629
 630 Which is very different from ``(gdb) ! ls''.  Implementing the latter
 631 is trivial.
 632
 633 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00034.html
 634
 635 --
 636
 637 Change the (char *list[]) to (const char (*)[]) so that dynamic lists can
 638 be passed.
 639
 640 --
 641
 642 When tab expanding something like ``set arch<tab>'' ignore the
 643 deprecated ``set archdebug'' and expand to ``set architecture''.
 644
 645 --
 646
 647 Replace the code that uses the host FPU with an emulator of the target
 648 FPU.
 649
 650 --
 651
 652                         Thread Support
 653                         ==============
 654
 655 --
 656
 657 Generic: lin-thread cannot handle thread exit (Mark Kettenis, Michael
 658 Snyder) http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00525.html
 659
 660 The thread_db assisted debugging code doesn't handle exiting threads
 661 properly, at least in combination with glibc 2.1.3 (the framework is
 662 there, just not the actual code).  There are at least two problems
 663 that prevent this from working.
 664
 665 As an additional reference point, the pre thread_db code did not work
 666 either.
 667
 668 --
 669
 670 GNU/Linux/x86 and random thread signals (and Solaris/SPARC but not
 671 Solaris/x86).
 672 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00336.html
 673
 674 Christopher Blizzard writes:
 675
 676 So, I've done some more digging into this and it looks like Jim
 677 Kingdon has reported this problem in the past:
 678
 679 http://sourceware.cygnus.com/ml/bug-gdb/1999-10/msg00058.html
 680
 681 I can reproduce this problem both with and without Tom's patch.  Has
 682 anyone seen this before?  Maybe have a solution for it hanging around?
 683 :)
 684
 685 There's a test case for this documented at:
 686
 687 when debugging threaded applications you get extra SIGTRAPs
 688 http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=9565
 689
 690 [There should be a GDB testcase - cagney]
 691
 692 --
 693
 694 GDB5 TOT on unixware 7
 695 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00119.html
 696
 697 Robert Lipe writes:
 698 > I just spun the top of tree of the GDB5 branch on UnixWare 7.  As a
 699 > practical matter, the current thread support is somewhat more annoying
 700 > than when GDB was thread-unaware.
 701
 702 --
 703
 704 Migrate qfThreadInfo packet -> qThreadInfo. (Andrew Cagney)
 705
 706 Add support for packet enable/disable commands with these thread
 707 packets.  General cleanup.
 708
 709 [PATCH] Document the ThreadInfo remote protocol queries
 710 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00832.html
 711
 712 [PATCH] "info threads" queries for remote.c
 713 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00831.html
 714
 715 --
 716
 717                         Language Support
 718                         ================
 719
 720 New languages come onto the scene all the time.
 721
 722 --
 723
 724 Re: Various C++ things
 725
 726 value_headof/value_from_vtable_info are worthless, and should be
 727 removed.  The one place in printcmd.c that uses it should use the RTTI
 728 functions.
 729
 730 RTTI for g++ should be using the typeinfo functions rather than the
 731 vtables.  The typeinfo functions are always at offset 4 from the
 732 beginning of the vtable, and are always right. The vtables will have
 733 weird names like E::VB sometimes.  The typeinfo function will always
 734 be "E type_info function", or somesuch.
 735
 736 value_virtual_fn_field needs to be fixed so there are no failures for
 737 virtual functions for C++ using g++.
 738
 739 Testsuite cases are the major priority right now for C++ support,
 740 since i have to make a lot of changes that could potentially break
 741 each other.
 742
 743 --
 744
 745 Add support for Modula3
 746
 747 Get DEC/Compaq to contribute their Modula-3 support.
 748
 749 --
 750
 751                 Remote Protocol Support
 752                 =======================
 753
 754 --
 755
 756 Remote protocol doco feedback.
 757
 758 Too much feedback to mention needs to be merged in (901660).  Search
 759 for the word ``remote''.
 760
 761
 762 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00023.html
 763 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00056.html
 764 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00382.html
 765
 766 --
 767
 768 GDB doesn't recover gracefully from remote protocol errors.
 769
 770 GDB wasn't checking for NAKs from the remote target.  Instead a NAK is
 771 ignored and a timeout is required before GDB retries.  A pre-cursor to
 772 fixing this this is making GDB's remote protocol packet more robust.
 773
 774 While downloading to a remote protocol target, gdb ignores packet
 775 errors in so far as it will continue to download with chunk N+1 even
 776 if chunk N was not correctly sent.  This causes gdb.base/remote.exp to
 777 take a painfully long time to run.  As a PS that test needs to be
 778 fixed so that it builds on 16 bit machines.
 779
 780 --
 781
 782 Add the cycle step command.
 783
 784 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00237.html
 785
 786 --
 787
 788 Resolve how to scale things to support very large packets.
 789
 790 --
 791
 792 Resolve how to handle a target that changes things like its endianess
 793 on the fly - should it be returned in the ``T'' packet?
 794
 795 Underlying problem is that the register file is target endian.  If the
 796 target endianess changes gdb doesn't know.
 797
 798 --
 799
 800 Rename read_register{,_pid}() to read_unsigned_register{,_pid}().
 801
 802 --
 803
 804                         Symbol Support
 805                         ==============
 806
 807 If / when GDB starts to support the debugging of multi-processor
 808 (rather than multi-thread) applications the symtab code will need to
 809 be updated a little so that several independant symbol tables are
 810 active at a given time.
 811
 812 The other interesting change is a clarification of the exact meaning
 813 of CORE_ADDR and that has had consequences for a few targets (that
 814 were abusing that data type).
 815
 816 --
 817
 818 Investiagate ways of reducing memory.
 819
 820 --
 821
 822 Investigate ways of improving load time.
 823
 824 --
 825
 826 Get the d10v to use POINTER_TO_ADDRESS and ADDRESS_TO_POINTER.
 827
 828 Consequence of recent symtab clarification.  No marks for figuring out
 829 who maintains the d10v.
 830
 831 --
 832
 833 Get the MIPS to correctly sign extend all address <-> pointer
 834 conversions.
 835
 836 Consequence of recent symtab clarification.  No marks for figuring out
 837 who maintains the MIPS.
 838
 839 --
 840
 841 GDB truncates 64 bit enums.
 842
 843 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00290.html
 844
 845 --
 846
 847                         Testsuite Support
 848                         =================
 849
 850 There are never to many testcases.
 851
 852 --
 853
 854 Better thread testsuite.
 855
 856 --
 857
 858 Better C++ testsuite.
 859
 860 --
 861
 862 Look at adding a GDB specific testsuite directory so that white box
 863 tests of key internals can be added (eg ui_file).
 864
 865 --
 866
 867 Separate out tests that involve the floating point (FP).
 868
 869 (Something for people brining up new targets).  FP and non-fp tests
 870 are combined.  I think there should be set of basic tests that
 871 exercise pure integer support and then a more expanded set that
 872 exercise FP and FP/integer interactions.
 873
 874 As an example, the MIPS, for n32 as problems with passing FP's and
 875 structs.  Since most inferior call tests include FP it is difficult to
 876 determine of the integer tests are ok.
 877
 878 --
 879
 880                 Architectural Changes: General
 881                 ==============================
 882
 883 These are harder than simple cleanups / fixes and, consequently
 884 involve more work.  Typically an Architectural Change will be broken
 885 down into a more digestible set of cleanups and fixes.
 886
 887 --
 888
 889 Cleanup software single step.
 890
 891 At present many targets implement software single step by directly
 892 blatting memory (see rs6000-tdep.c).  Those targets should register
 893 the applicable breakpoints using the breakpoint framework.  Perhaphs a
 894 new internal breakpoint class ``step'' is needed.
 895
 896 --
 897
 898 Replace READ_FP() with FRAME_HANDLE().
 899
 900 READ_FP() is a hangover from the days of the vax when the ABI really
 901 did have a frame pointer register.  Modern architectures typically
 902 construct a virtual frame-handle from the stack pointer and various
 903 other bits of string.
 904
 905 Unfortunatly GDB still treats this synthetic FP register as though it
 906 is real.  That in turn really confuses users (arm and ``print $fp'' VS
 907 ``info registers fp'').  The synthetic FP should be separated out of
 908 the true register set presented to the user.
 909
 910 --
 911
 912 Register Cache Cleanup (below from Andrew Cagney)
 913
 914 I would depict the current register architecture as something like:
 915
 916         High GDB  -->   Low GDB
 917             |              |
 918            \|/            \|/
 919             --- REG NR -----
 920                    |
 921     register + REGISTER_BYTE(reg_nr)
 922                    |
 923                   \|/
 924         -------------------------
 925         |  extern register[]    |
 926         -------------------------
 927
 928 where neither the high (valops.c et.al.) or low gdb (*-tdep.c) are
 929 really clear on what mechanisms they should be using to manipulate that
 930 buffer.   Further, much code assumes, dangerously, that registers are
 931 contigious.  Having got mips-tdep.c to support multiple ABIs, believe
 932 me, that is a bad assumption.  Finally, that register cache layout is
 933 determined by the current remote/local target and _not_ the less
 934 specific target ISA.  In fact, in many cases it is determined by the
 935 somewhat arbitrary layout of the [gG] packets!
 936
 937
 938 How I would like the register file to work is more like:
 939
 940
 941           High GDB
 942              |
 943             \|/
 944        pseudo reg-nr
 945              |
 946         map pseudo <->
 947         random cache
 948         bytes
 949              |
 950             \|/
 951         ------------
 952         | register |
 953         |  cache   |
 954         ------------
 955             /|\
 956              |
 957         map random cache
 958         bytes to target
 959         dependant i-face
 960             /|\
 961              |
 962        target dependant
 963         such as [gG] packet
 964         or ptrace buffer
 965
 966 The main objectives being:
 967
 968         o       a clear separation between the low
 969                 level target and the high level GDB
 970
 971         o       a mechanism that solves the general
 972                 problem of register aliases, overlaps
 973                 etc instead of treating them as optional
 974                 extras that can be wedged in as an after
 975                 thought (that is a reasonable description
 976                 of the current code).
 977
 978                 Identify then solve the hard case and the
 979                 rest just falls out.  GDB solved the easy
 980                 case and then tried to ignore the real
 981                 world :-)
 982
 983         o       a removal of the assumption that the
 984                 mapping between the register cache
 985                 and virtual registers is largely static.
 986                 If you flip the USR/SSR stack register
 987                 select bit in the status-register then
 988                 the corresponding stack registers should
 989                 reflect the change.
 990
 991         o       a mechanism that clearly separates the
 992                 gdb internal register cache from any
 993                 target (not architecture) dependant
 994                 specifics such as [gG] packets.
 995
 996 Of course, like anything, it sounds good in theory.  In reality, it
 997 would have to contend with many<->many relationships at both the
 998 virt<->cache and cache<->target level.  For instance:
 999
1000         virt<->cache
1001                 Modifying an mmx register may involve
1002                 scattering values across both FP and
1003                 mmpx specific parts of a buffer
1004
1005         cache<->target
1006                 When writing back a SP it may need to
1007                 both be written to both SP and USP.
1008
1009
1010 Hmm,
1011
1012 Rather than let this like the last time it was discussed, just slip, I'm
1013 first going to add this e-mail (+ references) to TODO.  I'd then like to
1014 sketch out a broad strategy I think could get us there.
1015
1016
1017 First thing I'd suggest is separating out the ``extern registers[]''
1018 code so that we can at least identify what is using it.  At present
1019 things are scattered across many files.  That way we can at least
1020 pretend that there is a cache instead of a global array :-)
1021
1022 I'd then suggest someone putting up a proposal for the pseudo-reg /
1023 high-level side interface so that code can be adopted to it.  For old
1024 code, initially a blanket rename of write_register_bytes() to
1025 deprecated_write_register_bytes() would help.
1026
1027 Following that would, finaly be the corresponding changes to the target.
1028
1029 --
1030
1031 Check that GDB can handle all BFD architectures (Andrew Cagney)
1032
1033 There should be a test that checks that BFD/GDB are in sync with
1034 regard to architecture changes.  Something like a test that first
1035 queries GDB for all supported architectures and then feeds each back
1036 to GDB..  Anyone interested in learning how to write tests?  :-)
1037
1038 --
1039
1040                 Architectural Change: Multi-arch et al.
1041                 =======================================
1042
1043 The long term objective is to remove all assumptions that there is a
1044 single target with a single address space with a single instruction
1045 set architecture and single application binary interface.
1046
1047 This is an ongoing effort.  The first milestone is to enable
1048 ``multi-arch'' where by all architectural decisions are made at
1049 runtime.
1050
1051 It should be noted that ``gdbarch'' is really ``gdbabi'' and
1052 ``gdbisa''.  Once things are multi-arched breaking that down correctly
1053 will become much easier.
1054
1055 --
1056
1057 GDBARCH cleanup (Andrew Cagney)
1058
1059 The non-generated parts of gdbarch.{sh,h,c} should be separated out
1060 into arch-utils.[hc].
1061
1062 Document that gdbarch_init_ftype could easily fail because it didn't
1063 identify an architecture.
1064
1065 --
1066
1067 Fix BELIEVE_PPC_PROMOTION.  Change it to BELIEVE_PPC_PROMOTION_P?
1068
1069 At present there is still #ifdef BELIEVE_PPC_PROMOTION code in the
1070 symtab file.
1071
1072 --
1073
1074 Fix target_signal_from_host() etc.
1075
1076 The name is wrong for starters.  ``target_signal'' should probably be
1077 ``gdb_signal''.  ``from_host'' should be ``from_target_signal''.
1078 After that it needs to be multi-arched and made independant of any
1079 host signal numbering.
1080
1081 --
1082
1083 Update ALPHA so that it uses ``struct frame_extra_info'' instead of
1084 EXTRA_FRAME_INFO.
1085
1086 This is a barrier to replacing mips_extra_func_info with something
1087 that works with multi-arch.
1088
1089 --
1090
1091 Multi-arch  mips_extra_func_info.
1092
1093 This first needs the alpha to be updated so that it uses ``struct
1094 frame_extra_info''.
1095
1096 --
1097
1098 Rationalize TARGET_SINGLE_FORMAT and TARGET_SINGLE_BIT et al.
1099
1100 Surely one of them is redundant.
1101
1102 --
1103
1104 Convert ALL architectures to MULTI-ARCH.
1105
1106 --
1107
1108 Select the initial multi-arch ISA / ABI based on --target or similar.
1109
1110 At present the default is based on what ever is first in the BFD
1111 archures table.  It should be determined based on the ``--target=...''
1112 name.
1113
1114 --
1115
1116 Make MIPS pure multi-arch.
1117
1118 It is only at the multi-arch enabled stage.
1119
1120 --
1121
1122 Truly multi-arch.
1123
1124 Enable the code to recognize --enable-targets=.... like BINUTILS does.
1125
1126 Can the tm.h and nm.h files be eliminated by multi-arch.
1127
1128 --
1129
1130         Architectural Change: MI, LIBGDB and scripting languages
1131         ========================================================
1132
1133 See also architectural changes related to the event loop.  LIBGDB
1134 can't be finished until there is a generic event loop being used by
1135 all targets.
1136
1137 The long term objective is it to be possible to integrate GDB into
1138 scripting languages.
1139
1140 --
1141
1142 Implement generic ``(gdb) commmand > file''
1143
1144 Once everything is going through ui_file it should be come fairly
1145 easy.
1146
1147 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00104.html
1148
1149 --
1150
1151 Replace gdb_stdtarg with gdb_targout (and possibly gdb_targerr).
1152
1153 gdb_stdtarg is easily confused with gdb_stdarg.
1154
1155 --
1156
1157 Extra ui_file methods - dump.
1158
1159 Very useful for whitebox testing.
1160
1161 --
1162
1163 Eliminate error_begin().
1164
1165 With ui_file, there is no need for the statefull error_begin ()
1166 function.
1167
1168 --
1169
1170 Send normal output to gdb_stdout.
1171 Send error messages to gdb_stderror.
1172 Send debug and log output log gdb_stdlog.
1173
1174 GDB still contains many cases where (f)printf or printf_filtered () is
1175 used when it should be sending the messages to gdb_stderror or
1176 gdb_stdlog.  The thought of #defining printf to something has crossed
1177 peoples minds ;-)
1178
1179 --
1180
1181 Re-do GDB's output pager.
1182
1183 GDB's output pager still relies on people correctly using *_filtered
1184 for gdb_stdout and *_unfiltered for gdb_stdlog / gdb_stderr.
1185 Hopefully, with all normal output going to gdb_stdout, the pager can
1186 just look at the ui_file that the output is on and then use that to
1187 decide what to do about paging.  Sounds good in theory.
1188
1189 --
1190
1191 Check/cleanup MI documentation.
1192
1193 The list of commands specified in the documentation needs to be
1194 checked against the mi-cmds.c table in a mechanical way (so that they
1195 two can be kept up-to-date).
1196
1197 --
1198
1199 Convert MI into libgdb
1200
1201 MI provides a text interface into what should be many of the libgdb
1202 functions.  The implementation of those functions should be separated
1203 into the MI interface and the functions proper.  Those functions being
1204 moved to gdb/lib say.
1205
1206 --
1207
1208 Create libgdb.h
1209
1210 The first part can already be found in defs.h.
1211
1212 --
1213
1214 MI's input does not use buffering.
1215
1216 At present the MI interface reads raw characters of from an unbuffered
1217 FD.  This is to avoid several nasty buffer/race conditions.  That code
1218 should be changed so that it registers its self with the event loop
1219 (on the input FD) and then push commands up to MI as they arrive.
1220
1221 The serial code already does this.
1222
1223 --
1224
1225 Make MI interface accessable from existing CLI.
1226
1227 --
1228
1229 Add a breakpoint-edit command to MI.
1230
1231 It would be similar to MI's breakpoint create but would apply to an
1232 existing breakpoint.  It saves the need to delete/create breakpoints
1233 when ever they are changed.
1234
1235 --
1236
1237 Add directory path to MI breakpoint.
1238
1239 That way the GUI's task of finding the file within which the
1240 breakpoint was set is simplified.
1241
1242 --
1243
1244 Add a mechanism to reject certain expression classes to MI
1245
1246 There are situtations where you don't want GDB's expression
1247 parser/evaluator to perform inferior function calls or variable
1248 assignments.  A way of restricting the expression parser so that such
1249 operations are not accepted would be very helpful.
1250
1251 --
1252
1253 Remove sideffects from libgdb breakpoint create function.
1254
1255 The user can use the CLI to create a breakpoint with partial
1256 information - no file (gdb would use the file from the last
1257 breakpoint).
1258
1259 The libgdb interface currently affects that environment which can lead
1260 to confusion when a user is setting breakpoints via both the MI and
1261 the CLI.
1262
1263 This is also a good example of how getting the CLI ``right'' will be
1264 hard.
1265
1266 --
1267
1268 Move gdb_lasterr to ui_out?
1269
1270 The way GDB throws errors and records them needs a re-think.  ui_out
1271 handles the correct output well.  It doesn't resolve what to do with
1272 output / error-messages when things go wrong.
1273
1274 --
1275
1276 do_setshow_command contains a 1024 byte buffer.
1277
1278 The function assumes that there will never be any more than 1024 bytes
1279 of enum.  It should use mem_file.
1280
1281 --
1282
1283 Should struct cmd_list_element . completer take the command as an
1284 argument?
1285
1286 --
1287
1288 Should the bulk of top.c:line_completion_function() be moved to
1289 command.[hc]?  complete_on_cmdlist() and complete_on_enums() could
1290 then be made private.
1291
1292 --
1293
1294 top.c (execute_command): Should a command being valid when the target
1295 is running be made an attribute (predicate) to the command rather than
1296 an explicit set of tests.
1297
1298 --
1299
1300 top.c (execute_command): Should the bulk of this function be moved
1301 into command.[hc] so that top.c doesn't grub around in the command
1302 internals?
1303
1304 --
1305
1306                 Architectural Change: Async
1307                 ===========================
1308
1309 While GDB uses an event loop when prompting the user for input.  That
1310 event loop is not exploited by targets when they allow the target
1311 program to continue.  Typically targets still block in (target_wait())
1312 until the program again halts.
1313
1314 The closest a target comes to supporting full asynchronous mode are
1315 the remote targets ``async'' and ``extended-async''.
1316
1317 --
1318
1319 Asynchronous expression evaluator
1320
1321 Inferior function calls hang GDB.
1322
1323 --
1324
1325 Fix implementation of  ``target xxx''.
1326
1327 At present when the user specifies ``target xxxx'', the CLI maps that
1328 directly onto a target open method.  It is then assumed that the
1329 target open method should do all sorts of complicated things as this
1330 is the only chance it has.  Check how the various remote targets
1331 duplicate the target operations.  Check also how the various targets
1332 behave differently for purely arbitrary reasons.
1333
1334 What should happen is that ``target xxxx'' should call a generic
1335 ``target'' function and that should then co-ordinate the opening of
1336 ``xxxx''.  This becomes especially important when you're trying to
1337 open an asynchronous target that may need to perform background tasks
1338 as part of the ``attach'' phase.
1339
1340 Unfortunatly, due to limitations in the old/creaking command.h
1341 interface, that isn't possible.  The function being called isn't told
1342 of the ``xxx'' or any other context information.
1343
1344 Consequently a precursor to fixing ``target xxxx'' is to clean up the
1345 CLI code so that it passes to the callback function (attatched to a
1346 command) useful information such as the actual command and a context
1347 for that command.  Other changes such as making ``struct command''
1348 opaque may also help.
1349
1350 See also:
1351 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
1352
1353 --
1354
1355 Make "target xxx" command interruptible.
1356
1357 As things become async this becomes possible.  A target would start
1358 the connect and then return control to the event loop.  A cntrl-c
1359 would notify the target that the operation is to be abandoned and the
1360 target code could respond.
1361
1362 --
1363
1364 Add a "suspend" subcommand of the "continue" command to suspend gdb
1365 while continuing  execution of the subprocess.  Useful when you are
1366 debugging servers and you want to dodge out and initiate a connection
1367 to a server running under gdb.
1368
1369 [hey async!!]
1370
1371 --
1372
1373                         TODO FAQ
1374                         ========
1375
1376 Frequently requested but not approved requests.
1377
1378 --
1379
1380 Eliminate unused argument warnings using ATTRIBUTE_UNUSED.
1381
1382 The benefits on this one are thought to be marginal - GDBs design
1383 means that unused parameters are very common.  GCC 3.0 will also
1384 include the option -Wno-unused-parameter which means that ``-Wall
1385 -Wno-unused-parameters -Werror'' can be specified.
1386
1387 --
1388
1389
1390
1391                         Legacy Wish List
1392                         ================
1393
1394 This list is not up to date, and opinions vary about the importance or
1395 even desirability of some of the items.  If you do fix something, it
1396 always pays to check the below.
1397
1398 --
1399
1400 @c This does not work (yet if ever).  FIXME.
1401 @c @item --parse=@var{lang} @dots{}
1402 @c Configure the @value{GDBN} expression parser to parse the listed languages.
1403 @c @samp{all} configures @value{GDBN} for all supported languages.  To get a
1404 @c list of all supported languages, omit the argument.  Without this
1405 @c option, @value{GDBN} is configured to parse all supported languages.
1406
1407 --
1408
1409 START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that
1410 is its default value.  Clean this up.
1411
1412 --
1413
1414 It should be possible to use symbols from shared libraries before we know
1415 exactly where the libraries will be loaded.  E.g. "b perror" before running
1416 the program.  This could maybe be done as an extension of the "breakpoint
1417 re-evaluation" after new symbols are loaded.
1418
1419 --
1420
1421 Make single_step() insert and remove breakpoints in one operation.
1422
1423 [If this is talking about having single_step() insert the breakpoints,
1424 run the target then pull the breakpoints then it is wrong.  The
1425 function has to return as control has to eventually be passed back to
1426 the main event loop.]
1427
1428 --
1429
1430 Speed up single stepping by avoiding extraneous ptrace calls.
1431
1432 --
1433
1434 Speed up single stepping by not inserting and removing breakpoints
1435 each time the inferior starts and stops.
1436
1437 Breakpoints should not be inserted and deleted all the time.  Only the
1438 one(s) there should be removed when we have to step over one.  Support
1439 breakpoints that don't have to be removed to step over them.
1440
1441 [this has resulted in numerous debates.  The issue isn't clear cut]
1442
1443 --
1444
1445 Provide "voodoo" debugging of core files.  This creates a zombie
1446 process as a child of the debugger, and loads it up with the data,
1447 stack, and regs of the core file.  This allows you to call functions
1448 in the executable, to manipulate the data in the core file.
1449
1450 [you wish]
1451
1452 --
1453
1454 GDB reopens the source file on every line, as you "next" through it.
1455
1456 [still true? I've a memory of this being fixed]
1457
1458 --
1459
1460 Perhaps "i source" should take an argument like that of "list".
1461
1462 --
1463
1464 Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if
1465 it matches the source line indicated.
1466
1467 --
1468
1469 The prompt at end of screen should accept space as well as CR.
1470
1471 --
1472
1473 Backtrace should point out what the currently selected frame is, in
1474 its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar,
1475 ...)" rather than "#3 foo (bar, ...)".
1476
1477 --
1478
1479 "i program" should work for core files, and display more info, like what
1480 actually caused it to die.
1481
1482 --
1483
1484 "x/10i" should shorten the long name, if any, on subsequent lines.
1485
1486 --
1487
1488 "next" over a function that longjumps, never stops until next time you happen
1489 to get to that spot by accident.  E.g. "n" over execute_command which has
1490 an error.
1491
1492 --
1493
1494 "set zeroprint off", don't bother printing members of structs which
1495 are entirely zero.  Useful for those big structs with few useful
1496 members.
1497
1498 --
1499
1500 GDB does four ioctl's for every command, probably switching terminal modes
1501 to/from inferior or for readline or something.
1502
1503 --
1504
1505 terminal_ours versus terminal_inferior: cache state.  Switch should be a noop
1506 if the state is the same, too.
1507
1508 --
1509
1510 "i frame" shows wrong "arglist at" location, doesn't show where the args
1511 should be found, only their actual values.
1512
1513 --
1514
1515 There should be a way for "set" commands to validate the new setting
1516 before it takes effect.
1517
1518 --
1519
1520 "ena d" is ambiguous, why?  "ena delete" seems to think it is a command!
1521
1522 --
1523
1524 i line VAR produces "Line number not known for symbol ``var''.".  I
1525 thought we were stashing that info now!
1526
1527 --
1528
1529 We should be able to write to random files at hex offsets like adb.
1530
1531 --
1532
1533 [elena - delete this]
1534
1535 Handle add_file with separate text, data, and bss addresses.  Maybe
1536 handle separate addresses for each segment in the object file?
1537
1538 --
1539
1540 [Jimb/Elena delete this one]
1541
1542 Handle free_named_symtab to cope with multiply-loaded object files
1543 in a dynamic linking environment.  Should remember the last copy loaded,
1544 but not get too snowed if it finds references to the older copy.
1545
1546 --
1547
1548 [elena delete this also]
1549
1550 Remove all references to:
1551         text_offset
1552         data_offset
1553         text_data_start
1554         text_end
1555         exec_data_offset
1556         ...
1557 now that we have BFD.  All remaining are in machine dependent files.
1558
1559 --
1560
1561 Re-organize help categories into things that tend to fit on a screen
1562 and hang together.
1563
1564 --
1565
1566 Add in commands like ADB's for searching for patterns, etc.  We should
1567 be able to examine and patch raw unsymboled binaries as well in gdb as
1568 we can in adb.  (E.g. increase the timeout in /bin/login without source).
1569
1570 [actually, add ADB interface :-]
1571
1572 --
1573
1574 When doing "step" or "next", if a few lines of source are skipped between
1575 the previous line and the current one, print those lines, not just the
1576 last line of a multiline statement.
1577
1578 --
1579
1580 Handling of "&" address-of operator needs some serious overhaul
1581 for ANSI C and consistency on arrays and functions.
1582   For "float point[15];":
1583 ptype &point[4]   ==> Attempt to take address of non-lvalue.
1584   For "char *malloc();":
1585 ptype malloc      ==> "char *()";  should be same as
1586 ptype &malloc     ==> "char *(*)()"
1587 call printf ("%x\n", malloc)   ==> weird value, should be same as
1588 call printf ("%x\n", &malloc)  ==> correct value
1589
1590 --
1591
1592 Fix dbxread.c symbol reading in the presence of interrupts.  It
1593 currently leaves a cleanup to blow away the entire symbol table when a
1594 QUIT occurs.  (What's wrong with that? -kingdon, 28 Oct 1993).
1595
1596 [I suspect that the grype was that, on a slow system, you might want
1597 to cntrl-c and get just half the symbols and then load the rest later
1598 - scary to be honest]
1599
1600 --
1601
1602 Mipsread.c reads include files depth-first, because the dependencies
1603 in the psymtabs are way too inclusive (it seems to me).  Figure out what
1604 really depends on what, to avoid recursing 20 or 30 times while reading
1605 real symtabs.
1606
1607 --
1608
1609 value_add() should be subtracting the lower bound of arrays, if known,
1610 and possibly checking against the upper bound for error reporting.
1611
1612 --
1613
1614 When listing source lines, check for a preceding \n, to verify that
1615 the file hasn't changed out from under us.
1616
1617 [fixed by some other means I think.  That hack wouldn't actually work
1618 reliably - the file might move such that another \n appears. ]
1619
1620 --
1621
1622 Get all the remote systems (where the protocol allows it) to be able to
1623 stop the remote system when the GDB user types ^C (like remote.c
1624 does).  For ebmon, use ^Ak.
1625
1626 --
1627
1628 Possible feature: A version of the "disassemble" command which shows
1629 both source and assembly code ("set symbol-filename on" is a partial
1630 solution).
1631
1632 [has this been done?  It was certainly done for MI and GDBtk]
1633
1634 --
1635
1636 investigate "x/s 0" (right now stops early) (I think maybe GDB is
1637 using a 0 address for bad purposes internally).
1638
1639 --
1640
1641 Make "info path" and path_command work again (but independent of the
1642 environment either of gdb or that we'll pass to the inferior).
1643
1644 --
1645
1646 Make GDB understand the GCC feature for putting octal constants in
1647 enums.  Make it so overflow on an enum constant does not error_type
1648 the whole type.  Allow arbitrarily large enums with type attributes.
1649 Put all this stuff in the testsuite.
1650
1651 --
1652
1653 Make TYPE_CODE_ERROR with a non-zero TYPE_LENGTH more useful (print
1654 the value in hex; process type attributes).  Add this to the
1655 testsuite.  This way future compilers can add new types and old
1656 versions of GDB can do something halfway reasonable.
1657
1658 --
1659
1660 Fix mdebugread.c:parse_type to do fundamental types right (see
1661 rs6000_builtin_type in stabsread.c for what "right" is--the point is
1662 that the debug format fixes the sizes of these things and it shouldn't
1663 depend on stuff like TARGET_PTR_BIT and so on.  For mdebug, there seem
1664 to be separate bt* codes for 64 bit and 32 bit things, and GDB should
1665 be aware of that).  Also use a switch statement for clarity and speed.
1666
1667 --
1668
1669 Investigate adding symbols in target_load--some targets do, some
1670 don't.
1671
1672 --
1673
1674 Put dirname in psymtabs and change lookup*symtab to use dirname (so
1675 /foo/bar.c works whether compiled by cc /foo/bar.c, or cd /foo; cc
1676 bar.c).
1677
1678 --
1679
1680 Merge xcoffread.c and coffread.c.  Use breakpoint_re_set instead of
1681 fixup_breakpoints.
1682
1683 --
1684
1685 Make a watchpoint which contains a function call an error (it is
1686 broken now, making it work is probably not worth the effort).
1687
1688 --
1689
1690 New test case based on weird.exp but in which type numbers are not
1691 renumbered (thus multiply defining a type).  This currently causes an
1692 infinite loop on "p v_comb".
1693
1694 --
1695
1696 [Hey! Hint Hint Delete Delete!!!]
1697
1698 Fix 386 floating point so that floating point registers are real
1699 registers (but code can deal at run-time if they are missing, like
1700 mips and 68k).  This would clean up "info float" and related stuff.
1701
1702 --
1703
1704 gcc -g -c enummask.c then gdb enummask.o, then "p v".  GDB complains
1705 about not being able to access memory location 0.
1706
1707 -------------------- enummask.c
1708 enum mask
1709 {
1710   ANIMAL = 0,
1711   VEGETABLE = 1,
1712   MINERAL = 2,
1713   BASIC_CATEGORY = 3,
1714
1715   WHITE = 0,
1716   BLUE = 4,
1717   GREEN = 8,
1718   BLACK = 0xc,
1719   COLOR = 0xc,
1720
1721   ALIVE = 0x10,
1722
1723   LARGE = 0x20
1724 } v;
1725
1726 --
1727
1728 If try to modify value in file with "set write off" should give
1729 appropriate error not "cannot access memory at address 0x65e0".
1730
1731 --
1732
1733 Allow core file without exec file on RS/6000.
1734
1735 --
1736
1737 Make sure "shell" with no arguments works right on DOS.
1738
1739 --
1740
1741 Make gdb.ini (as well as .gdbinit) be checked on all platforms, so
1742 the same directory can be NFS-mounted on unix or DOS, and work the
1743 same way.
1744
1745 --
1746
1747 [Is this another delete???]
1748
1749 Get SECT_OFF_TEXT stuff out of objfile_relocate (might be needed to
1750 get RS/6000 to work right, might not be immediately relevant).
1751
1752 --
1753
1754 Work out some kind of way to allow running the inferior to be done as
1755 a sub-execution of, eg. breakpoint command lists.  Currently running
1756 the inferior interupts any command list execution.  This would require
1757 some rewriting of wait_for_inferior & friends, and hence should
1758 probably be done in concert with the above.
1759
1760 --
1761
1762 Add function arguments to gdb user defined functions.
1763
1764 --
1765
1766 Add convenience variables that refer to exec file, symbol file,
1767 selected frame source file, selected frame function, selected frame
1768 line number, etc.
1769
1770 --
1771
1772 Modify the handling of symbols grouped through BINCL/EINCL stabs to
1773 allocate a partial symtab for each BINCL/EINCL grouping.  This will
1774 seriously decrease the size of inter-psymtab dependencies and hence
1775 lessen the amount that needs to be read in when a new source file is
1776 accessed.
1777
1778 --
1779
1780 Add a command for searching memory, a la adb.  It specifies size,
1781 mask, value, start address.  ADB searches until it finds it or hits
1782 an error (or is interrupted).
1783
1784 --
1785
1786 Remove the range and type checking code and documentation, if not
1787 going to implement.
1788
1789 # Local Variables:
1790 # mode: text
1791 # End: