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 -- 2001-03-08
  14
  15 Update GDB's coding standard documentation.  Known topics:
  16
  17        o     alloca/malloc et.al.
  18
  19        o     typedef and structs
  20
  21        o        ISO-C
  22
  23 and most likely also:
  24
  25        o         include conventions
  26
  27 --
  28
  29 Wow, three bug reports for the same problem in one day!  We should
  30 probably make fixing this a real priority :-).
  31
  32 Anyway, thanks for reporting.
  33
  34 The following patch will fix the problems with setting breakpoints in
  35 dynamically loaded objects:
  36
  37    http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00230.html
  38
  39 This patch isn't checked in yet (ping Michael/JimB), but I hope this
  40 will be in the next GDB release.
  41
  42 There should really be a test in the testsuite for this problem, since
  43 it keeps coming up :-(.  Any volunteers?
  44
  45 Mark
  46
  47 --
  48
  49 x86 linux GDB and SIGALRM (???)
  50 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00803.html
  51
  52 This problem has been fixed, but a regression test still needs to be
  53 added to the testsuite:
  54 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00309.html
  55
  56 Mark
  57
  58 [The test has been submitted for approval - cagney]
  59
  60 --
  61
  62 RFD: infrun.c: No bpstat_stop_status call after proceed over break?
  63 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00665.html
  64
  65 GDB misses watchpoint triggers after proceeding over a breakpoint on
  66 x86 targets.
  67
  68 --
  69
  70 GDB 5.0 doesn't work on Linux/SPARC
  71
  72 There are two parts to this.
  73
  74       o   GDB 5.0 doesn't work on GNU/Linux/SPARC32
  75
  76       o   GDB 5.0 doesn't work on the new target
  77           GNU/Linux/SPARC64
  78
  79 GDB does build on both these targets.
  80
  81 The first problem is the one that should be fixed.
  82
  83 --
  84
  85                 GDB 5.1 - New features
  86                 ======================
  87
  88 The following new features should be included in 5.1.
  89
  90 --
  91
  92 Enable MI by default.  Old code can be deleted after 5.1 is out.
  93
  94 Issues:
  95
  96         o       syntax change where a list would
  97                 look like:
  98                      [ foo=a, foo=b, foo=c ]
  99                 instead of
 100                      { foo=a, foo=b, foo=c }
 101
 102         o       kill off the idea of a reverse
 103                 query.
 104
 105         o       review test cases
 106
 107         o       enable it
 108
 109 --
 110
 111 Pascal (Pierre Muller, David Taylor)
 112
 113 Pierre Muller has contributed patches for adding Pascal Language
 114 support to GDB.
 115
 116 2 pascal language patches inserted in database
 117 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00521.html
 118
 119 Indent -gnu ?
 120 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00496.html
 121
 122 [I think this has been merged, need to confirm - cagney]
 123
 124 --
 125
 126 Java (Anthony Green, David Taylor)
 127
 128 Anthony Green has a number of Java patches that did not make it into
 129 the 5.0 release.  The first two are in cvs now, but the third needs
 130 some fixing up before it can go in.
 131
 132 Patch: java tests
 133 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00512.html
 134
 135 Patch: java booleans
 136 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00515.html
 137
 138 Patch: handle N_MAIN stab
 139 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00527.html
 140
 141 -- 2001-03-08
 142
 143 Add CRIS target.
 144
 145 A predicate to this is the multi-arching of SOFTWARE_SINGLE_STEP().  A
 146 patch has been submitted.
 147
 148 --
 149
 150                 GDB 5.1 - Cleanups
 151                 ==================
 152
 153 The following code cleanups will hopefully be applied to GDB 5.1.
 154
 155 -- 2001-03-08
 156
 157 Resolve the status of all broken targets as identified by the MAINTAINERS
 158 file.
 159
 160 --
 161
 162 Change documentation to GFDL license.
 163
 164 ``It is time to make an effort to start using the GFDL more
 165 thoroughly.  Would all GNU maintainers please change the license to
 166 the GFDL, for all manuals and other major documentation files?
 167
 168 The GFDL and some instructions for using it can be found in
 169 http://www.gnu.org/copyleft/''
 170
 171         RMS
 172
 173 --
 174
 175 Fix copyright notices.
 176
 177 Turns out that ``1998-2000'' isn't considered valid :-(
 178
 179 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00467.html
 180
 181 --
 182
 183                 GDB 5.1 - Known Problems
 184                 ========================
 185
 186 --
 187
 188 z8k
 189
 190 The z8k has suffered bit rot and is known to not build.  The problem
 191 was occuring in the opcodes directory.
 192
 193 --
 194
 195 Solaris 8 x86 CURSES_H problem
 196 http://sources.redhat.com/ml/gdb/2000-07/msg00038.html
 197
 198 The original problem was worked around with:
 199
 200     2000-06-06  Michael Snyder  <msnyder@cygnus.com>
 201
 202         * configure.in: Enable autoconf to find curses.h on Solaris 2.8.
 203         * configure: Regenerate.
 204
 205 When building both GDB and SID using the same source tree the problem
 206 will still occure. sid/component/configure.in mis-configures
 207 <curses.h> and leaves wrong information in the config cache.
 208
 209 --
 210
 211                 GDB 5.2 - Fixes
 212                 ===============
 213
 214 --
 215
 216 Thread support.  Right now, as soon as a thread finishes and exits,
 217 you're hosed.  This problem is reported once a week or so.
 218
 219 --
 220
 221                 GDB 5.2 - New features
 222                 ======================
 223
 224 --
 225
 226 GCC 3.0 ABI support (but hopefully sooner...).
 227
 228 --
 229
 230 Objective C/C++ support (but hopefully sooner...).
 231
 232 --
 233
 234                 GDB 5.2 - Cleanups
 235                 ==================
 236
 237 The following cleanups have been identified as part of GDB 5.2.
 238
 239 --
 240
 241 Remove old code that does not use ui_out functions and all the related
 242 "ifdef"s.  This also allows the elimination of -DUI_OUT from
 243 Makefile.in and configure.in.
 244
 245 --
 246
 247 Compiler warnings.
 248
 249 Eliminate all warnings for at least one host/target for the flags:
 250 -Wimplicit -Wreturn-type -Wcomment -Wtrigraphs -Wformat -Wparentheses
 251 -Wpointer-arith -Wuninitialized
 252
 253 --
 254
 255 Restructure gdb directory tree so that it avoids any 8.3 and 14
 256 filename problems.
 257
 258 --
 259
 260 Convert GDB build process to AUTOMAKE.
 261
 262 See also sub-directory configure below.
 263
 264 The current convention is (kind of) to use $(<header>_h) in all
 265 dependency lists.  It isn't done in a consistent way.
 266
 267 --
 268
 269                 GDB 5.2 - Known Problems
 270                 ========================
 271
 272 --
 273
 274                 Code Cleanups: General
 275                 ======================
 276
 277 The following are more general cleanups and fixes.  They are not tied
 278 to any specific release.
 279
 280 --
 281
 282 Revised UDP support (was: Re: [Fwd: [patch] UDP transport support])
 283 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00000.html
 284
 285 (Broken) support for GDB's remote protocol across UDP is to be
 286 included in the follow-on release.
 287
 288 It should be noted that UDP can only work when the [Gg] packet fits in
 289 a single UDP packet.
 290
 291 There is also much debate over the merit of this.
 292
 293 --
 294
 295 Can't build IRIX -> arm GDB.
 296 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00356.html
 297
 298 David Whedon writes:
 299 > Now I'm building for an embedded arm target.  If there is a way of turning
 300 > remote-rdi off, I couldn't find it. It looks like it gets built by default
 301 > in gdb/configure.tgt(line 58) Anyway, the build dies in
 302 > gdb/rdi-share/unixcomm.c.  SERPORT1 et. al. never get defined because we
 303 > aren't one of the architectures supported.
 304
 305 --
 306
 307 Problem with weak functions
 308 http://sourceware.cygnus.com/ml/gdb/2000-05/msg00060.html
 309
 310 Dan Nicolaescu writes:
 311 > It seems that gdb-4.95.1  does not display correctly the function when
 312 > stoping in weak functions.
 313 >
 314 > It stops in a function that is defined as weak, not in the function
 315 > that is actually run...
 316
 317 --
 318
 319 Follow through `make check' with --enable-shared.
 320
 321 When the srcware tree is configured with --enable-shared, the `expect'
 322 program won't run properly.  Jim Wilson found out gdb has a local hack
 323 to set LD_LIBRARY_PATH, but, AFAIK, no other project has been hacked
 324 similarly.
 325
 326 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00845.html
 327
 328 --
 329
 330 Delete macro TARGET_BYTE_ORDER_SELECTABLE.
 331
 332 Patches in the database.
 333
 334 --
 335
 336 printcmd.c (print_address_numeric):
 337
 338 NOTE: This assumes that the significant address information is kept in
 339 the least significant bits of ADDR - the upper bits were either zero
 340 or sign extended.  Should ADDRESS_TO_POINTER() or some
 341 ADDRESS_TO_PRINTABLE() be used to do the conversion?
 342
 343 --
 344
 345 The BFD directory requires bug-fixed AUTOMAKE et.al.
 346
 347 AUTOMAKE 1.4 incorrectly set the TEXINPUTS environment variable.  It
 348 contained the full path to texinfo.tex when it should have only
 349 contained the directory.  The bug has been fixed in the current
 350 AUTOMAKE sources.  Automake snapshots can be found in:
 351         ftp://sourceware.cygnus.com/pub/gdb/snapshots
 352 and     ftp://sourceware.cygnus.com/pub/binutils
 353
 354 --
 355
 356 Find something better than DEFAULT_BFD_ARCH, DEFAULT_BFD_VEC to
 357 determine the default isa/byte-order.
 358
 359 --
 360
 361 Rely on BFD_BIG_ENDIAN and BFD_LITTLE_ENDIAN instead of host dependent
 362 BIG_ENDIAN and LITTLE_ENDIAN.
 363
 364 --
 365
 366 Eliminate more compiler warnings.
 367
 368 Of course there also needs to be the usual debate over which warnings
 369 are valid and how to best go about this.
 370
 371 One method: choose a single option; get agreement that it is
 372 reasonable; try it out to see if there isn't anything silly about it
 373 (-Wunused-parameters is an example of that) then incrementally hack
 374 away.
 375
 376 The other method is to enable all warnings and eliminate them from one
 377 file at a time.
 378
 379 --
 380
 381 Elimination of ``(catch_errors_ftype *) func''.
 382
 383 Like make_cleanup_func it isn't portable.
 384 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00791.html
 385 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00814.html
 386
 387 --
 388
 389 Nuke #define CONST_PTR.
 390
 391 --
 392
 393 Nuke USG define.
 394
 395 --
 396
 397 [PATCH/5] src/intl/Makefile.in:distclean additions
 398 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00363.html
 399
 400 Do not forget to merge the patch back into the trunk.
 401
 402 --
 403
 404 Rationalize the host-endian code (grep for HOST_BYTE_ORDER).
 405
 406 At present defs.h includes <endian.h> (which is linux specific) yet
 407 almost nothing depends on it.  Suggest "gdb_endian.h" which can also
 408 handle <machine/endian.h> and only include that where it is really
 409 needed.
 410
 411 --
 412
 413 Replace savestring() with something from libiberty.
 414
 415 An xstrldup()? but that would have different semantics.
 416
 417 --
 418
 419 Rationalize use of floatformat_unknown in GDB sources.
 420
 421 Instead of defaulting to floatformat_unknown, should hosts/targets
 422 specify the value explicitly?
 423
 424 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
 425
 426 --
 427
 428 Add a ``name'' member to include/floatformat.h:struct floatformat.
 429 Print that name in gdbarch.c.
 430
 431 --
 432
 433 Sort out the harris mess in include/floatformat.h (it hardwires two
 434 different floating point formats).
 435
 436 --
 437
 438 See of the GDB local floatformat_do_doublest() and libiberty's
 439 floatformat_to_double (which was once GDB's ...) can be merged some
 440 how.
 441
 442 --
 443
 444 Eliminate mmalloc(), mstrsave() et.al. from GDB.
 445
 446 Also eliminate it from defs.h.
 447
 448 --
 449
 450 Eliminate PTR.  ISO-C allows ``void *''.
 451
 452 --
 453
 454 Eliminate abort ().
 455
 456 GDB should never abort.  GDB should either throw ``error ()'' or
 457 ``internal_error ()''.  Better still GDB should naturally unwind with
 458 an error status.
 459
 460 --
 461
 462 GDB probably doesn't build on FreeBSD pre 2.2.x
 463 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00378.html
 464
 465 Fixes to get FreeBSD working on 2.2.x, 3.x and 4.x caused the code to
 466 suffer bit rot.
 467
 468 --
 469
 470 Deprecate "fg".  Apparently ``fg'' is actually continue.
 471
 472 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00417.html
 473
 474 --
 475
 476 Deprecate current use of ``floatformat_unknown''.
 477
 478 Require all targets to explicitly provide their float format instead
 479 of defaulting to floatformat unknown.  Doing the latter leads to nasty
 480 bugs.
 481
 482 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
 483
 484 --
 485
 486 Rationalize floatformat_to_double() vs floatformat_to_doublest().
 487
 488 Looks like GDB migrated floatformat_to_double() to libiberty but then
 489 turned around and created a ..._to_doublest() the latter containing
 490 several bug fixes.
 491
 492 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00472.html
 493
 494 --
 495
 496 Move floatformat_ia64_ext to libiberty/include floatformat.[ch].
 497
 498 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00466.html
 499
 500 --
 501
 502 The ``maintenance deprecate set endian big'' command doesn't notice
 503 that it is deprecating ``set endian'' and not ``set endian big'' (big
 504 is implemented using an enum).  Is anyone going to notice this?
 505
 506 --
 507
 508 When tab expanding something like ``set arch<tab>'' ignore the
 509 deprecated ``set archdebug'' and expand to ``set architecture''.
 510
 511 --
 512
 513 Eliminate ``arm_register_names[j] = (char *) regnames[j]'' and the
 514 like from arm-tdep.c.
 515
 516 --
 517
 518 Fix uses of ->function.cfunc = set_function().
 519
 520 The command.c code calls sfunc() when a set command.  Rather than
 521 change it suggest fixing the callback function so that it is more
 522 useful.  See:
 523
 524 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
 525
 526 See also ``Fix implementation of  ``target xxx''.'' below.
 527
 528 --
 529
 530 IRIX 3.x support is probably broken.
 531
 532 --
 533
 534 Delete sim/SIM_HAVE_BREAKPOINTS and gdb/SIM_HAS_BREAKPOINTS.
 535 http://sourceware.cygnus.com/ml/gdb-patches/2000-07/msg00042.html
 536
 537 Apart from the d30v, are there any sim/common simulators that make use
 538 of this?
 539
 540 A brief summary of what happened is that sim/common/sim-break.c was
 541 created as a good idea.  It turned out a better idea was to use
 542 SIM_SIGBREAK and have GDB pass back sim_resume (..., SIGBREAK).
 543
 544 --
 545
 546 Move remote_remove_hw_breakpoint, remote_insert_hw_breakpoint,
 547 remote_remove_watchpoint, remote_insert_watchpoint into target vector.
 548
 549 --
 550
 551 Eliminate ``extern'' from C files.
 552
 553 --
 554
 555 Replace ``STREQ()'' et.al. with ``strcmp() == 0'' et.al.
 556
 557 Extreme care is recommeded - perhaps only modify tests that are
 558 exercised by the testsuite (as determined using some type of code
 559 coverage analysis).
 560
 561 --
 562
 563 Replace the file gdb/CONTRIBUTE with a file that is generated from the
 564 gdb/doc/*.texinfo directory.
 565
 566 --
 567
 568                         New Features and Fixes
 569                         ======================
 570
 571 These are harder than cleanups but easier than work involving
 572 fundamental architectural change.
 573
 574 --
 575
 576 Hardware watchpoint problems on x86 OSes, including Linux:
 577
 578 1. Delete/disable hardware watchpoints should free hardware debug
 579 registers.
 580 2. Watch for different values on a viariable with one hardware debug
 581 register.
 582
 583 According to Eli Zaretskii <eliz@delorie.com>:
 584
 585 These are not GDB/ia32 issues per se: the above features are all
 586 implemented in the DJGPP port of GDB and work in v5.0.  Every
 587 x86-based target should be able to lift the relevant parts of
 588 go32-nat.c and use them almost verbatim.  You get debug register
 589 sharing through reference counts, and the ability to watch large
 590 regions (up to 16 bytes) using multiple registers.  (The required
 591 infrastructure in high-level GDB application code, mostly in
 592 breakpoint.c, is also working since v5.0.)
 593
 594 --
 595
 596 Add built-by, build-date, tm, xm, nm and anything else into gdb binary
 597 so that you can see how the GDB was created.
 598
 599 --
 600
 601 Add an "info bfd" command that displays supported object formats,
 602 similarly to objdump -i.
 603
 604 Is there a command already?
 605
 606 --
 607
 608 Fix ``I'm sorry, Dave, I can't do that.'' from symfile.c.
 609
 610 This requires internationalization.
 611
 612 --
 613
 614 Add support for:
 615
 616 (gdb) p fwprintf(stdout,L"%S\n", f)
 617 No symbol "L" in current context.
 618
 619 --
 620
 621 Cleanup configury support for optional sub-directories.
 622
 623 Check how GCC handles multiple front ends for an example of how things
 624 could work.  A tentative first step is to rationalize things so that
 625 all sub directories are handled in a fashion similar to gdb/mi.
 626
 627 See also automake above.
 628
 629 --
 630
 631 Add a transcript mechanism to GDB.
 632
 633 Such a mechanism might log all gdb input and output to a file in a
 634 form that would allow it to be replayed.  It could involve ``gdb
 635 --transcript=FILE'' or it could involve ``(gdb) transcript file''.
 636
 637 --
 638
 639 Can the xdep files be replaced by autoconf?
 640
 641 --
 642
 643 Document trace machinery
 644
 645 --
 646
 647 Document ui-out and ui-file.
 648
 649 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00121.html
 650
 651 --
 652
 653 Update texinfo.tex to latest?
 654
 655 --
 656
 657 Incorporate agentexpr.texi into gdb.texinfo
 658
 659 agentexpr.texi mostly describes the details of the byte code used for
 660 tracepoints, not the internals of the support for this in GDB.  So it
 661 looks like gdb.texinfo is a better place for this information.
 662
 663 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00566.html
 664
 665 --
 666
 667 Document overlay machinery.
 668
 669 --
 670
 671 ``(gdb) catch signal SIGNAL''
 672
 673 Overlaps with ``handle SIGNAL'' but the implied behavior is different.
 674 You can attach commands to a catch but not a handle.  A handle has a
 675 limited number of hardwired actions.
 676
 677 --
 678
 679 Get the TUI working on all platforms.
 680
 681 --
 682
 683 Add support for ``gdb --- PROGRAM ARGS ...''.
 684 Add support for ``gdb -cmd=...''
 685
 686 Along with many variations.  Check:
 687
 688 ????? for a full discussion.
 689
 690 for a discussion.
 691
 692 --
 693
 694 Implement ``(gdb) !ls''.
 695
 696 Which is very different from ``(gdb) ! ls''.  Implementing the latter
 697 is trivial.
 698
 699 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00034.html
 700
 701 --
 702
 703 Change the (char *list[]) to (const char (*)[]) so that dynamic lists can
 704 be passed.
 705
 706 --
 707
 708 When tab expanding something like ``set arch<tab>'' ignore the
 709 deprecated ``set archdebug'' and expand to ``set architecture''.
 710
 711 --
 712
 713 Replace the code that uses the host FPU with an emulator of the target
 714 FPU.
 715
 716 --
 717
 718 The "ocd reset" command needs to flush the dcache, which requires breaking
 719 the abstraction layer between the target independent and target code.  One
 720 way to address this is provide a generic "reset" command and target vector.
 721
 722 http://sources.redhat.com/ml/gdb-patches/2000-10/msg00011.html
 723
 724 --
 725
 726                         Thread Support
 727                         ==============
 728
 729 --
 730
 731 Generic: lin-thread cannot handle thread exit (Mark Kettenis, Michael
 732 Snyder) http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00525.html
 733
 734 The thread_db assisted debugging code doesn't handle exiting threads
 735 properly, at least in combination with glibc 2.1.3 (the framework is
 736 there, just not the actual code).  There are at least two problems
 737 that prevent this from working.
 738
 739 As an additional reference point, the pre thread_db code did not work
 740 either.
 741
 742 --
 743
 744 GNU/Linux/x86 and random thread signals (and Solaris/SPARC but not
 745 Solaris/x86).
 746 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00336.html
 747
 748 Christopher Blizzard writes:
 749
 750 So, I've done some more digging into this and it looks like Jim
 751 Kingdon has reported this problem in the past:
 752
 753 http://sourceware.cygnus.com/ml/bug-gdb/1999-10/msg00058.html
 754
 755 I can reproduce this problem both with and without Tom's patch.  Has
 756 anyone seen this before?  Maybe have a solution for it hanging around?
 757 :)
 758
 759 There's a test case for this documented at:
 760
 761 when debugging threaded applications you get extra SIGTRAPs
 762 http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=9565
 763
 764 [There should be a GDB testcase - cagney]
 765
 766 --
 767
 768 GDB5 TOT on unixware 7
 769 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00119.html
 770
 771 Robert Lipe writes:
 772 > I just spun the top of tree of the GDB5 branch on UnixWare 7.  As a
 773 > practical matter, the current thread support is somewhat more annoying
 774 > than when GDB was thread-unaware.
 775
 776 --
 777
 778 Migrate qfThreadInfo packet -> qThreadInfo. (Andrew Cagney)
 779
 780 Add support for packet enable/disable commands with these thread
 781 packets.  General cleanup.
 782
 783 [PATCH] Document the ThreadInfo remote protocol queries
 784 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00832.html
 785
 786 [PATCH] "info threads" queries for remote.c
 787 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00831.html
 788
 789 --
 790
 791                         Language Support
 792                         ================
 793
 794 New languages come onto the scene all the time.
 795
 796 --
 797
 798 Re: Various C++ things
 799
 800 value_headof/value_from_vtable_info are worthless, and should be
 801 removed.  The one place in printcmd.c that uses it should use the RTTI
 802 functions.
 803
 804 RTTI for g++ should be using the typeinfo functions rather than the
 805 vtables.  The typeinfo functions are always at offset 4 from the
 806 beginning of the vtable, and are always right. The vtables will have
 807 weird names like E::VB sometimes.  The typeinfo function will always
 808 be "E type_info function", or somesuch.
 809
 810 value_virtual_fn_field needs to be fixed so there are no failures for
 811 virtual functions for C++ using g++.
 812
 813 Testsuite cases are the major priority right now for C++ support,
 814 since i have to make a lot of changes that could potentially break
 815 each other.
 816
 817 --
 818
 819 Add support for Modula3
 820
 821 Get DEC/Compaq to contribute their Modula-3 support.
 822
 823 --
 824
 825                 Remote Protocol Support
 826                 =======================
 827
 828 --
 829
 830 Remote protocol doco feedback.
 831
 832 Too much feedback to mention needs to be merged in (901660).  Search
 833 for the word ``remote''.
 834
 835
 836 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00023.html
 837 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00056.html
 838 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00382.html
 839
 840 --
 841
 842 GDB doesn't recover gracefully from remote protocol errors.
 843
 844 GDB wasn't checking for NAKs from the remote target.  Instead a NAK is
 845 ignored and a timeout is required before GDB retries.  A pre-cursor to
 846 fixing this this is making GDB's remote protocol packet more robust.
 847
 848 While downloading to a remote protocol target, gdb ignores packet
 849 errors in so far as it will continue to download with chunk N+1 even
 850 if chunk N was not correctly sent.  This causes gdb.base/remote.exp to
 851 take a painfully long time to run.  As a PS that test needs to be
 852 fixed so that it builds on 16 bit machines.
 853
 854 --
 855
 856 Add the cycle step command.
 857
 858 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00237.html
 859
 860 --
 861
 862 Resolve how to scale things to support very large packets.
 863
 864 --
 865
 866 Resolve how to handle a target that changes things like its endianess
 867 on the fly - should it be returned in the ``T'' packet?
 868
 869 Underlying problem is that the register file is target endian.  If the
 870 target endianess changes gdb doesn't know.
 871
 872 --
 873
 874 Rename read_register{,_pid}() to read_unsigned_register{,_pid}().
 875
 876 --
 877
 878                         Symbol Support
 879                         ==============
 880
 881 If / when GDB starts to support the debugging of multi-processor
 882 (rather than multi-thread) applications the symtab code will need to
 883 be updated a little so that several independent symbol tables are
 884 active at a given time.
 885
 886 The other interesting change is a clarification of the exact meaning
 887 of CORE_ADDR and that has had consequences for a few targets (that
 888 were abusing that data type).
 889
 890 --
 891
 892 Investiagate ways of reducing memory.
 893
 894 --
 895
 896 Investigate ways of improving load time.
 897
 898 --
 899
 900 Get the d10v to use POINTER_TO_ADDRESS and ADDRESS_TO_POINTER.
 901
 902 Consequence of recent symtab clarification.  No marks for figuring out
 903 who maintains the d10v.
 904
 905 --
 906
 907 Get the MIPS to correctly sign extend all address <-> pointer
 908 conversions.
 909
 910 Consequence of recent symtab clarification.  No marks for figuring out
 911 who maintains the MIPS.
 912
 913 --
 914
 915 GDB truncates 64 bit enums.
 916
 917 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00290.html
 918
 919 --
 920
 921                         Testsuite Support
 922                         =================
 923
 924 There are never to many testcases.
 925
 926 --
 927
 928 Better thread testsuite.
 929
 930 --
 931
 932 Better C++ testsuite.
 933
 934 --
 935
 936 Look at adding a GDB specific testsuite directory so that white box
 937 tests of key internals can be added (eg ui_file).
 938
 939 --
 940
 941 Separate out tests that involve the floating point (FP).
 942
 943 (Something for people brining up new targets).  FP and non-fp tests
 944 are combined.  I think there should be set of basic tests that
 945 exercise pure integer support and then a more expanded set that
 946 exercise FP and FP/integer interactions.
 947
 948 As an example, the MIPS, for n32 as problems with passing FP's and
 949 structs.  Since most inferior call tests include FP it is difficult to
 950 determine of the integer tests are ok.
 951
 952 --
 953
 954                 Architectural Changes: General
 955                 ==============================
 956
 957 These are harder than simple cleanups / fixes and, consequently
 958 involve more work.  Typically an Architectural Change will be broken
 959 down into a more digestible set of cleanups and fixes.
 960
 961 --
 962
 963 Cleanup software single step.
 964
 965 At present many targets implement software single step by directly
 966 blatting memory (see rs6000-tdep.c).  Those targets should register
 967 the applicable breakpoints using the breakpoint framework.  Perhaphs a
 968 new internal breakpoint class ``step'' is needed.
 969
 970 --
 971
 972 Replace READ_FP() with FRAME_HANDLE().
 973
 974 READ_FP() is a hangover from the days of the vax when the ABI really
 975 did have a frame pointer register.  Modern architectures typically
 976 construct a virtual frame-handle from the stack pointer and various
 977 other bits of string.
 978
 979 Unfortunately GDB still treats this synthetic FP register as though it
 980 is real.  That in turn really confuses users (arm and ``print $fp'' VS
 981 ``info registers fp'').  The synthetic FP should be separated out of
 982 the true register set presented to the user.
 983
 984 --
 985
 986 Register Cache Cleanup (below from Andrew Cagney)
 987
 988 I would depict the current register architecture as something like:
 989
 990         High GDB  -->   Low GDB
 991             |              |
 992            \|/            \|/
 993             --- REG NR -----
 994                    |
 995     register + REGISTER_BYTE(reg_nr)
 996                    |
 997                   \|/
 998         -------------------------
 999         |  extern register[]    |
1000         -------------------------
1001
1002 where neither the high (valops.c et.al.) or low gdb (*-tdep.c) are
1003 really clear on what mechanisms they should be using to manipulate that
1004 buffer.   Further, much code assumes, dangerously, that registers are
1005 contigious.  Having got mips-tdep.c to support multiple ABIs, believe
1006 me, that is a bad assumption.  Finally, that register cache layout is
1007 determined by the current remote/local target and _not_ the less
1008 specific target ISA.  In fact, in many cases it is determined by the
1009 somewhat arbitrary layout of the [gG] packets!
1010
1011
1012 How I would like the register file to work is more like:
1013
1014
1015           High GDB
1016              |
1017             \|/
1018        pseudo reg-nr
1019              |
1020         map pseudo <->
1021         random cache
1022         bytes
1023              |
1024             \|/
1025         ------------
1026         | register |
1027         |  cache   |
1028         ------------
1029             /|\
1030              |
1031         map random cache
1032         bytes to target
1033         dependent i-face
1034             /|\
1035              |
1036        target dependent
1037         such as [gG] packet
1038         or ptrace buffer
1039
1040 The main objectives being:
1041
1042         o       a clear separation between the low
1043                 level target and the high level GDB
1044
1045         o       a mechanism that solves the general
1046                 problem of register aliases, overlaps
1047                 etc instead of treating them as optional
1048                 extras that can be wedged in as an after
1049                 thought (that is a reasonable description
1050                 of the current code).
1051
1052                 Identify then solve the hard case and the
1053                 rest just falls out.  GDB solved the easy
1054                 case and then tried to ignore the real
1055                 world :-)
1056
1057         o       a removal of the assumption that the
1058                 mapping between the register cache
1059                 and virtual registers is largely static.
1060                 If you flip the USR/SSR stack register
1061                 select bit in the status-register then
1062                 the corresponding stack registers should
1063                 reflect the change.
1064
1065         o       a mechanism that clearly separates the
1066                 gdb internal register cache from any
1067                 target (not architecture) dependent
1068                 specifics such as [gG] packets.
1069
1070 Of course, like anything, it sounds good in theory.  In reality, it
1071 would have to contend with many<->many relationships at both the
1072 virt<->cache and cache<->target level.  For instance:
1073
1074         virt<->cache
1075                 Modifying an mmx register may involve
1076                 scattering values across both FP and
1077                 mmpx specific parts of a buffer
1078
1079         cache<->target
1080                 When writing back a SP it may need to
1081                 both be written to both SP and USP.
1082
1083
1084 Hmm,
1085
1086 Rather than let this like the last time it was discussed, just slip, I'm
1087 first going to add this e-mail (+ references) to TODO.  I'd then like to
1088 sketch out a broad strategy I think could get us there.
1089
1090
1091 First thing I'd suggest is separating out the ``extern registers[]''
1092 code so that we can at least identify what is using it.  At present
1093 things are scattered across many files.  That way we can at least
1094 pretend that there is a cache instead of a global array :-)
1095
1096 I'd then suggest someone putting up a proposal for the pseudo-reg /
1097 high-level side interface so that code can be adopted to it.  For old
1098 code, initially a blanket rename of write_register_bytes() to
1099 deprecated_write_register_bytes() would help.
1100
1101 Following that would, finaly be the corresponding changes to the target.
1102
1103 --
1104
1105 Check that GDB can handle all BFD architectures (Andrew Cagney)
1106
1107 There should be a test that checks that BFD/GDB are in sync with
1108 regard to architecture changes.  Something like a test that first
1109 queries GDB for all supported architectures and then feeds each back
1110 to GDB..  Anyone interested in learning how to write tests?  :-)
1111
1112 --
1113
1114                 Architectural Change: Multi-arch et al.
1115                 =======================================
1116
1117 The long term objective is to remove all assumptions that there is a
1118 single target with a single address space with a single instruction
1119 set architecture and single application binary interface.
1120
1121 This is an ongoing effort.  The first milestone is to enable
1122 ``multi-arch'' where by all architectural decisions are made at
1123 runtime.
1124
1125 It should be noted that ``gdbarch'' is really ``gdbabi'' and
1126 ``gdbisa''.  Once things are multi-arched breaking that down correctly
1127 will become much easier.
1128
1129 --
1130
1131 GDBARCH cleanup (Andrew Cagney)
1132
1133 The non-generated parts of gdbarch.{sh,h,c} should be separated out
1134 into arch-utils.[hc].
1135
1136 Document that gdbarch_init_ftype could easily fail because it didn't
1137 identify an architecture.
1138
1139 --
1140
1141 Fix BELIEVE_PPC_PROMOTION.  Change it to BELIEVE_PPC_PROMOTION_P?
1142
1143 At present there is still #ifdef BELIEVE_PPC_PROMOTION code in the
1144 symtab file.
1145
1146 --
1147
1148 Fix target_signal_from_host() etc.
1149
1150 The name is wrong for starters.  ``target_signal'' should probably be
1151 ``gdb_signal''.  ``from_host'' should be ``from_target_signal''.
1152 After that it needs to be multi-arched and made independent of any
1153 host signal numbering.
1154
1155 --
1156
1157 Update ALPHA so that it uses ``struct frame_extra_info'' instead of
1158 EXTRA_FRAME_INFO.
1159
1160 This is a barrier to replacing mips_extra_func_info with something
1161 that works with multi-arch.
1162
1163 --
1164
1165 Multi-arch  mips_extra_func_info.
1166
1167 This first needs the alpha to be updated so that it uses ``struct
1168 frame_extra_info''.
1169
1170 --
1171
1172 Rationalize TARGET_SINGLE_FORMAT and TARGET_SINGLE_BIT et al.
1173
1174 Surely one of them is redundant.
1175
1176 --
1177
1178 Convert ALL architectures to MULTI-ARCH.
1179
1180 --
1181
1182 Select the initial multi-arch ISA / ABI based on --target or similar.
1183
1184 At present the default is based on what ever is first in the BFD
1185 archures table.  It should be determined based on the ``--target=...''
1186 name.
1187
1188 --
1189
1190 Make MIPS pure multi-arch.
1191
1192 It is only at the multi-arch enabled stage.
1193
1194 --
1195
1196 Truly multi-arch.
1197
1198 Enable the code to recognize --enable-targets=.... like BINUTILS does.
1199
1200 Can the tm.h and nm.h files be eliminated by multi-arch.
1201
1202 --
1203
1204         Architectural Change: MI, LIBGDB and scripting languages
1205         ========================================================
1206
1207 See also architectural changes related to the event loop.  LIBGDB
1208 can't be finished until there is a generic event loop being used by
1209 all targets.
1210
1211 The long term objective is it to be possible to integrate GDB into
1212 scripting languages.
1213
1214 --
1215
1216 Implement generic ``(gdb) commmand > file''
1217
1218 Once everything is going through ui_file it should be come fairly
1219 easy.
1220
1221 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00104.html
1222
1223 --
1224
1225 Replace gdb_stdtarg with gdb_targout (and possibly gdb_targerr).
1226
1227 gdb_stdtarg is easily confused with gdb_stdarg.
1228
1229 --
1230
1231 Extra ui_file methods - dump.
1232
1233 Very useful for whitebox testing.
1234
1235 --
1236
1237 Eliminate error_begin().
1238
1239 With ui_file, there is no need for the statefull error_begin ()
1240 function.
1241
1242 --
1243
1244 Send normal output to gdb_stdout.
1245 Send error messages to gdb_stderror.
1246 Send debug and log output log gdb_stdlog.
1247
1248 GDB still contains many cases where (f)printf or printf_filtered () is
1249 used when it should be sending the messages to gdb_stderror or
1250 gdb_stdlog.  The thought of #defining printf to something has crossed
1251 peoples minds ;-)
1252
1253 --
1254
1255 Re-do GDB's output pager.
1256
1257 GDB's output pager still relies on people correctly using *_filtered
1258 for gdb_stdout and *_unfiltered for gdb_stdlog / gdb_stderr.
1259 Hopefully, with all normal output going to gdb_stdout, the pager can
1260 just look at the ui_file that the output is on and then use that to
1261 decide what to do about paging.  Sounds good in theory.
1262
1263 --
1264
1265 Check/cleanup MI documentation.
1266
1267 The list of commands specified in the documentation needs to be
1268 checked against the mi-cmds.c table in a mechanical way (so that they
1269 two can be kept up-to-date).
1270
1271 --
1272
1273 Convert MI into libgdb
1274
1275 MI provides a text interface into what should be many of the libgdb
1276 functions.  The implementation of those functions should be separated
1277 into the MI interface and the functions proper.  Those functions being
1278 moved to gdb/lib say.
1279
1280 --
1281
1282 Create libgdb.h
1283
1284 The first part can already be found in defs.h.
1285
1286 --
1287
1288 MI's input does not use buffering.
1289
1290 At present the MI interface reads raw characters of from an unbuffered
1291 FD.  This is to avoid several nasty buffer/race conditions.  That code
1292 should be changed so that it registers its self with the event loop
1293 (on the input FD) and then push commands up to MI as they arrive.
1294
1295 The serial code already does this.
1296
1297 --
1298
1299 Make MI interface accessible from existing CLI.
1300
1301 --
1302
1303 Add a breakpoint-edit command to MI.
1304
1305 It would be similar to MI's breakpoint create but would apply to an
1306 existing breakpoint.  It saves the need to delete/create breakpoints
1307 when ever they are changed.
1308
1309 --
1310
1311 Add directory path to MI breakpoint.
1312
1313 That way the GUI's task of finding the file within which the
1314 breakpoint was set is simplified.
1315
1316 --
1317
1318 Add a mechanism to reject certain expression classes to MI
1319
1320 There are situtations where you don't want GDB's expression
1321 parser/evaluator to perform inferior function calls or variable
1322 assignments.  A way of restricting the expression parser so that such
1323 operations are not accepted would be very helpful.
1324
1325 --
1326
1327 Remove sideffects from libgdb breakpoint create function.
1328
1329 The user can use the CLI to create a breakpoint with partial
1330 information - no file (gdb would use the file from the last
1331 breakpoint).
1332
1333 The libgdb interface currently affects that environment which can lead
1334 to confusion when a user is setting breakpoints via both the MI and
1335 the CLI.
1336
1337 This is also a good example of how getting the CLI ``right'' will be
1338 hard.
1339
1340 --
1341
1342 Move gdb_lasterr to ui_out?
1343
1344 The way GDB throws errors and records them needs a re-think.  ui_out
1345 handles the correct output well.  It doesn't resolve what to do with
1346 output / error-messages when things go wrong.
1347
1348 --
1349
1350 do_setshow_command contains a 1024 byte buffer.
1351
1352 The function assumes that there will never be any more than 1024 bytes
1353 of enum.  It should use mem_file.
1354
1355 --
1356
1357 Should struct cmd_list_element . completer take the command as an
1358 argument?
1359
1360 --
1361
1362 Should the bulk of top.c:line_completion_function() be moved to
1363 command.[hc]?  complete_on_cmdlist() and complete_on_enums() could
1364 then be made private.
1365
1366 --
1367
1368 top.c (execute_command): Should a command being valid when the target
1369 is running be made an attribute (predicate) to the command rather than
1370 an explicit set of tests.
1371
1372 --
1373
1374 top.c (execute_command): Should the bulk of this function be moved
1375 into command.[hc] so that top.c doesn't grub around in the command
1376 internals?
1377
1378 --
1379
1380                 Architectural Change: Async
1381                 ===========================
1382
1383 While GDB uses an event loop when prompting the user for input.  That
1384 event loop is not exploited by targets when they allow the target
1385 program to continue.  Typically targets still block in (target_wait())
1386 until the program again halts.
1387
1388 The closest a target comes to supporting full asynchronous mode are
1389 the remote targets ``async'' and ``extended-async''.
1390
1391 --
1392
1393 Asynchronous expression evaluator
1394
1395 Inferior function calls hang GDB.
1396
1397 --
1398
1399 Fix implementation of  ``target xxx''.
1400
1401 At present when the user specifies ``target xxxx'', the CLI maps that
1402 directly onto a target open method.  It is then assumed that the
1403 target open method should do all sorts of complicated things as this
1404 is the only chance it has.  Check how the various remote targets
1405 duplicate the target operations.  Check also how the various targets
1406 behave differently for purely arbitrary reasons.
1407
1408 What should happen is that ``target xxxx'' should call a generic
1409 ``target'' function and that should then co-ordinate the opening of
1410 ``xxxx''.  This becomes especially important when you're trying to
1411 open an asynchronous target that may need to perform background tasks
1412 as part of the ``attach'' phase.
1413
1414 Unfortunately, due to limitations in the old/creaking command.h
1415 interface, that isn't possible.  The function being called isn't told
1416 of the ``xxx'' or any other context information.
1417
1418 Consequently a precursor to fixing ``target xxxx'' is to clean up the
1419 CLI code so that it passes to the callback function (attatched to a
1420 command) useful information such as the actual command and a context
1421 for that command.  Other changes such as making ``struct command''
1422 opaque may also help.
1423
1424 See also:
1425 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
1426
1427 --
1428
1429 Make "target xxx" command interruptible.
1430
1431 As things become async this becomes possible.  A target would start
1432 the connect and then return control to the event loop.  A cntrl-c
1433 would notify the target that the operation is to be abandoned and the
1434 target code could respond.
1435
1436 --
1437
1438 Add a "suspend" subcommand of the "continue" command to suspend gdb
1439 while continuing  execution of the subprocess.  Useful when you are
1440 debugging servers and you want to dodge out and initiate a connection
1441 to a server running under gdb.
1442
1443 [hey async!!]
1444
1445 --
1446
1447                         TODO FAQ
1448                         ========
1449
1450 Frequently requested but not approved requests.
1451
1452 --
1453
1454 Eliminate unused argument warnings using ATTRIBUTE_UNUSED.
1455
1456 The benefits on this one are thought to be marginal - GDBs design
1457 means that unused parameters are very common.  GCC 3.0 will also
1458 include the option -Wno-unused-parameter which means that ``-Wall
1459 -Wno-unused-parameters -Werror'' can be specified.
1460
1461 --
1462
1463
1464
1465                         Legacy Wish List
1466                         ================
1467
1468 This list is not up to date, and opinions vary about the importance or
1469 even desirability of some of the items.  If you do fix something, it
1470 always pays to check the below.
1471
1472 --
1473
1474 @c This does not work (yet if ever).  FIXME.
1475 @c @item --parse=@var{lang} @dots{}
1476 @c Configure the @value{GDBN} expression parser to parse the listed languages.
1477 @c @samp{all} configures @value{GDBN} for all supported languages.  To get a
1478 @c list of all supported languages, omit the argument.  Without this
1479 @c option, @value{GDBN} is configured to parse all supported languages.
1480
1481 --
1482
1483 START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that
1484 is its default value.  Clean this up.
1485
1486 --
1487
1488 It should be possible to use symbols from shared libraries before we know
1489 exactly where the libraries will be loaded.  E.g. "b perror" before running
1490 the program.  This could maybe be done as an extension of the "breakpoint
1491 re-evaluation" after new symbols are loaded.
1492
1493 --
1494
1495 Make single_step() insert and remove breakpoints in one operation.
1496
1497 [If this is talking about having single_step() insert the breakpoints,
1498 run the target then pull the breakpoints then it is wrong.  The
1499 function has to return as control has to eventually be passed back to
1500 the main event loop.]
1501
1502 --
1503
1504 Speed up single stepping by avoiding extraneous ptrace calls.
1505
1506 --
1507
1508 Speed up single stepping by not inserting and removing breakpoints
1509 each time the inferior starts and stops.
1510
1511 Breakpoints should not be inserted and deleted all the time.  Only the
1512 one(s) there should be removed when we have to step over one.  Support
1513 breakpoints that don't have to be removed to step over them.
1514
1515 [this has resulted in numerous debates.  The issue isn't clear cut]
1516
1517 --
1518
1519 Provide "voodoo" debugging of core files.  This creates a zombie
1520 process as a child of the debugger, and loads it up with the data,
1521 stack, and regs of the core file.  This allows you to call functions
1522 in the executable, to manipulate the data in the core file.
1523
1524 [you wish]
1525
1526 --
1527
1528 GDB reopens the source file on every line, as you "next" through it.
1529
1530 [still true? I've a memory of this being fixed]
1531
1532 --
1533
1534 Perhaps "i source" should take an argument like that of "list".
1535
1536 --
1537
1538 Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if
1539 it matches the source line indicated.
1540
1541 --
1542
1543 The prompt at end of screen should accept space as well as CR.
1544
1545 --
1546
1547 Backtrace should point out what the currently selected frame is, in
1548 its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar,
1549 ...)" rather than "#3 foo (bar, ...)".
1550
1551 --
1552
1553 "i program" should work for core files, and display more info, like what
1554 actually caused it to die.
1555
1556 --
1557
1558 "x/10i" should shorten the long name, if any, on subsequent lines.
1559
1560 --
1561
1562 "next" over a function that longjumps, never stops until next time you happen
1563 to get to that spot by accident.  E.g. "n" over execute_command which has
1564 an error.
1565
1566 --
1567
1568 "set zeroprint off", don't bother printing members of structs which
1569 are entirely zero.  Useful for those big structs with few useful
1570 members.
1571
1572 --
1573
1574 GDB does four ioctl's for every command, probably switching terminal modes
1575 to/from inferior or for readline or something.
1576
1577 --
1578
1579 terminal_ours versus terminal_inferior: cache state.  Switch should be a noop
1580 if the state is the same, too.
1581
1582 --
1583
1584 "i frame" shows wrong "arglist at" location, doesn't show where the args
1585 should be found, only their actual values.
1586
1587 --
1588
1589 There should be a way for "set" commands to validate the new setting
1590 before it takes effect.
1591
1592 --
1593
1594 "ena d" is ambiguous, why?  "ena delete" seems to think it is a command!
1595
1596 --
1597
1598 i line VAR produces "Line number not known for symbol ``var''.".  I
1599 thought we were stashing that info now!
1600
1601 --
1602
1603 We should be able to write to random files at hex offsets like adb.
1604
1605 --
1606
1607 [elena - delete this]
1608
1609 Handle add_file with separate text, data, and bss addresses.  Maybe
1610 handle separate addresses for each segment in the object file?
1611
1612 --
1613
1614 [Jimb/Elena delete this one]
1615
1616 Handle free_named_symtab to cope with multiply-loaded object files
1617 in a dynamic linking environment.  Should remember the last copy loaded,
1618 but not get too snowed if it finds references to the older copy.
1619
1620 --
1621
1622 [elena delete this also]
1623
1624 Remove all references to:
1625         text_offset
1626         data_offset
1627         text_data_start
1628         text_end
1629         exec_data_offset
1630         ...
1631 now that we have BFD.  All remaining are in machine dependent files.
1632
1633 --
1634
1635 Re-organize help categories into things that tend to fit on a screen
1636 and hang together.
1637
1638 --
1639
1640 Add in commands like ADB's for searching for patterns, etc.  We should
1641 be able to examine and patch raw unsymboled binaries as well in gdb as
1642 we can in adb.  (E.g. increase the timeout in /bin/login without source).
1643
1644 [actually, add ADB interface :-]
1645
1646 --
1647
1648 When doing "step" or "next", if a few lines of source are skipped between
1649 the previous line and the current one, print those lines, not just the
1650 last line of a multiline statement.
1651
1652 --
1653
1654 Handling of "&" address-of operator needs some serious overhaul
1655 for ANSI C and consistency on arrays and functions.
1656   For "float point[15];":
1657 ptype &point[4]   ==> Attempt to take address of non-lvalue.
1658   For "char *malloc();":
1659 ptype malloc      ==> "char *()";  should be same as
1660 ptype &malloc     ==> "char *(*)()"
1661 call printf ("%x\n", malloc)   ==> weird value, should be same as
1662 call printf ("%x\n", &malloc)  ==> correct value
1663
1664 --
1665
1666 Fix dbxread.c symbol reading in the presence of interrupts.  It
1667 currently leaves a cleanup to blow away the entire symbol table when a
1668 QUIT occurs.  (What's wrong with that? -kingdon, 28 Oct 1993).
1669
1670 [I suspect that the grype was that, on a slow system, you might want
1671 to cntrl-c and get just half the symbols and then load the rest later
1672 - scary to be honest]
1673
1674 --
1675
1676 Mipsread.c reads include files depth-first, because the dependencies
1677 in the psymtabs are way too inclusive (it seems to me).  Figure out what
1678 really depends on what, to avoid recursing 20 or 30 times while reading
1679 real symtabs.
1680
1681 --
1682
1683 value_add() should be subtracting the lower bound of arrays, if known,
1684 and possibly checking against the upper bound for error reporting.
1685
1686 --
1687
1688 When listing source lines, check for a preceding \n, to verify that
1689 the file hasn't changed out from under us.
1690
1691 [fixed by some other means I think.  That hack wouldn't actually work
1692 reliably - the file might move such that another \n appears. ]
1693
1694 --
1695
1696 Get all the remote systems (where the protocol allows it) to be able to
1697 stop the remote system when the GDB user types ^C (like remote.c
1698 does).  For ebmon, use ^Ak.
1699
1700 --
1701
1702 Possible feature: A version of the "disassemble" command which shows
1703 both source and assembly code ("set symbol-filename on" is a partial
1704 solution).
1705
1706 [has this been done?  It was certainly done for MI and GDBtk]
1707
1708 --
1709
1710 investigate "x/s 0" (right now stops early) (I think maybe GDB is
1711 using a 0 address for bad purposes internally).
1712
1713 --
1714
1715 Make "info path" and path_command work again (but independent of the
1716 environment either of gdb or that we'll pass to the inferior).
1717
1718 --
1719
1720 Make GDB understand the GCC feature for putting octal constants in
1721 enums.  Make it so overflow on an enum constant does not error_type
1722 the whole type.  Allow arbitrarily large enums with type attributes.
1723 Put all this stuff in the testsuite.
1724
1725 --
1726
1727 Make TYPE_CODE_ERROR with a non-zero TYPE_LENGTH more useful (print
1728 the value in hex; process type attributes).  Add this to the
1729 testsuite.  This way future compilers can add new types and old
1730 versions of GDB can do something halfway reasonable.
1731
1732 --
1733
1734 Fix mdebugread.c:parse_type to do fundamental types right (see
1735 rs6000_builtin_type in stabsread.c for what "right" is--the point is
1736 that the debug format fixes the sizes of these things and it shouldn't
1737 depend on stuff like TARGET_PTR_BIT and so on.  For mdebug, there seem
1738 to be separate bt* codes for 64 bit and 32 bit things, and GDB should
1739 be aware of that).  Also use a switch statement for clarity and speed.
1740
1741 --
1742
1743 Investigate adding symbols in target_load--some targets do, some
1744 don't.
1745
1746 --
1747
1748 Put dirname in psymtabs and change lookup*symtab to use dirname (so
1749 /foo/bar.c works whether compiled by cc /foo/bar.c, or cd /foo; cc
1750 bar.c).
1751
1752 --
1753
1754 Merge xcoffread.c and coffread.c.  Use breakpoint_re_set instead of
1755 fixup_breakpoints.
1756
1757 --
1758
1759 Make a watchpoint which contains a function call an error (it is
1760 broken now, making it work is probably not worth the effort).
1761
1762 --
1763
1764 New test case based on weird.exp but in which type numbers are not
1765 renumbered (thus multiply defining a type).  This currently causes an
1766 infinite loop on "p v_comb".
1767
1768 --
1769
1770 [Hey! Hint Hint Delete Delete!!!]
1771
1772 Fix 386 floating point so that floating point registers are real
1773 registers (but code can deal at run-time if they are missing, like
1774 mips and 68k).  This would clean up "info float" and related stuff.
1775
1776 --
1777
1778 gcc -g -c enummask.c then gdb enummask.o, then "p v".  GDB complains
1779 about not being able to access memory location 0.
1780
1781 -------------------- enummask.c
1782 enum mask
1783 {
1784   ANIMAL = 0,
1785   VEGETABLE = 1,
1786   MINERAL = 2,
1787   BASIC_CATEGORY = 3,
1788
1789   WHITE = 0,
1790   BLUE = 4,
1791   GREEN = 8,
1792   BLACK = 0xc,
1793   COLOR = 0xc,
1794
1795   ALIVE = 0x10,
1796
1797   LARGE = 0x20
1798 } v;
1799
1800 --
1801
1802 If try to modify value in file with "set write off" should give
1803 appropriate error not "cannot access memory at address 0x65e0".
1804
1805 --
1806
1807 Allow core file without exec file on RS/6000.
1808
1809 --
1810
1811 Make sure "shell" with no arguments works right on DOS.
1812
1813 --
1814
1815 Make gdb.ini (as well as .gdbinit) be checked on all platforms, so
1816 the same directory can be NFS-mounted on unix or DOS, and work the
1817 same way.
1818
1819 --
1820
1821 [Is this another delete???]
1822
1823 Get SECT_OFF_TEXT stuff out of objfile_relocate (might be needed to
1824 get RS/6000 to work right, might not be immediately relevant).
1825
1826 --
1827
1828 Work out some kind of way to allow running the inferior to be done as
1829 a sub-execution of, eg. breakpoint command lists.  Currently running
1830 the inferior interupts any command list execution.  This would require
1831 some rewriting of wait_for_inferior & friends, and hence should
1832 probably be done in concert with the above.
1833
1834 --
1835
1836 Add function arguments to gdb user defined functions.
1837
1838 --
1839
1840 Add convenience variables that refer to exec file, symbol file,
1841 selected frame source file, selected frame function, selected frame
1842 line number, etc.
1843
1844 --
1845
1846 Modify the handling of symbols grouped through BINCL/EINCL stabs to
1847 allocate a partial symtab for each BINCL/EINCL grouping.  This will
1848 seriously decrease the size of inter-psymtab dependencies and hence
1849 lessen the amount that needs to be read in when a new source file is
1850 accessed.
1851
1852 --
1853
1854 Add a command for searching memory, a la adb.  It specifies size,
1855 mask, value, start address.  ADB searches until it finds it or hits
1856 an error (or is interrupted).
1857
1858 --
1859
1860 Remove the range and type checking code and documentation, if not
1861 going to implement.
1862
1863 # Local Variables:
1864 # mode: text
1865 # End: