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