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                 Known problems in GDB 5.0
   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 a follow-on
  12 release.
  13
  14 --
  15
  16 The BFD directory requires bug-fixed AUTOMAKE et.al.
  17
  18 AUTOMAKE 1.4 incorrectly set the TEXINPUTS environment variable.  It
  19 contained the full path to texinfo.tex when it should have only
  20 contained the directory.  The bug has been fixed in the current
  21 AUTOMAKE sources.  Automake snapshots can be found in:
  22         ftp://sourceware.cygnus.com/pub/gdb/snapshots
  23 and     ftp://sourceware.cygnus.com/pub/binutils
  24
  25 --
  26
  27 RFD: infrun.c: No bpstat_stop_status call after proceed over break?
  28 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00665.html
  29
  30 GDB misses watchpoint triggers after proceeding over a breakpoint on
  31 x86 targets.
  32
  33 --
  34
  35 x86 linux GDB and SIGALRM (???)
  36 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00803.html
  37
  38 I know there are problems with single stepping through signal
  39 handlers.  These problems were present in 4.18.  They were just masked
  40 because 4.18 failed to recognize signal handlers.  Fixing it is not
  41 easy, and will require changes to handle_inferior_event(), that I
  42 prefer not to make before the 5.0 release.
  43
  44 Mark
  45
  46 --
  47
  48 Revised UDP support (was: Re: [Fwd: [patch] UDP transport support])
  49 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00000.html
  50
  51 (Broken) support for GDB's remote protocol across UDP is to be
  52 included in the follow-on release.
  53
  54 --
  55
  56 Can't build IRIX -> arm GDB.
  57 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00356.html
  58
  59 David Whedon writes:
  60 > Now I'm building for an embedded arm target.  If there is a way of turning
  61 > remote-rdi off, I couldn't find it. It looks like it gets built by default
  62 > in gdb/configure.tgt(line 58) Anyway, the build dies in
  63 > gdb/rdi-share/unixcomm.c.  SERPORT1 et. al. never get defined because we
  64 > aren't one of the architectures supported.
  65
  66 --
  67
  68 Problem with weak functions
  69 http://sourceware.cygnus.com/ml/gdb/2000-05/msg00060.html
  70
  71 Dan Nicolaescu writes:
  72 > It seems that gdb-4.95.1  does not display correctly the function when
  73 > stoping in weak functions.
  74 >
  75 > It stops in a function that is defined as weak, not in the function
  76 > that is actualy run...
  77
  78 --
  79
  80 GDB 5.0 doesn't work on Linux/SPARC
  81
  82 --
  83
  84                 Code Cleanups: Next Release
  85                 ===========================
  86
  87 The following are small cleanups that will hopefully be completed by
  88 the follow on to 5.0.
  89
  90 --
  91
  92 Delete macro TARGET_BYTE_ORDER_SELECTABLE.
  93
  94 Patches in the database.
  95
  96 --
  97
  98 Purge PARAMS.
  99
 100 Eliminate all uses of PARAMS in GDB's source code.
 101
 102 --
 103
 104 Fix copyright notices.
 105
 106 Turns out that ``1998-2000'' isn't considered valid :-(
 107
 108 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00467.html
 109
 110 --
 111
 112 Find something better than DEFAULT_BFD_ARCH, DEFAULT_BFD_VEC to
 113 determine the default isa/byte-order.
 114
 115 --
 116
 117 Rely on BFD_BIG_ENDIAN and BFD_LITTLE_ENDIAN instead of host dependant
 118 BIG_ENDIAN and LITTLE_ENDIAN.
 119
 120 --
 121
 122                 Code Cleanups: General
 123                 ======================
 124
 125 The following are more general cleanups and fixes.  They are not tied
 126 to any specific release.
 127
 128 --
 129
 130 Eliminate more compiler warnings.
 131
 132 Of course there also needs to be the usual debate over which warnings
 133 are valid and how to best go about this.
 134
 135 One method: choose a single option; get agreement that it is
 136 reasonable; try it out to see if there isn't anything silly about it
 137 (-Wunused-parameters is an example of that) then incrementally hack
 138 away.
 139
 140 The other method is to enable all warnings and eliminate them from one
 141 file at a time.
 142
 143 --
 144
 145 Elimination of ``(catch_errors_ftype *) func''.
 146
 147 Like make_cleanup_func it isn't portable.
 148 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00791.html
 149 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00814.html
 150
 151
 152 --
 153
 154 Nuke USG define.
 155
 156 --
 157
 158 [PATCH/5] src/intl/Makefile.in:distclean additions
 159 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00363.html
 160
 161 Do not forget to merge the patch back into the trunk.
 162
 163 --
 164
 165 Rationalize the host-endian code (grep for HOST_BYTE_ORDER).
 166
 167 At present defs.h includes <endian.h> (which is linux specific) yet
 168 almost nothing depends on it.  Suggest "gdb_endian.h" which can also
 169 handle <machine/endian.h> and only include that where it is really
 170 needed.
 171
 172 --
 173
 174 Replace asprintf() calls with xasprintf() calls.
 175
 176 As with things like strdup() most calls to asprintf() don't check the
 177 return value.
 178
 179 --
 180
 181 Replace strsave() + mstrsave() with libiberty:xstrdup().
 182
 183 --
 184
 185 Replace savestring() with something from libiberty.
 186
 187 An xstrldup()? but that would have different semantics.
 188
 189 --
 190
 191 Rationalize use of floatformat_unknown in GDB sources.
 192
 193 Instead of defaulting to floatformat_unknown, should hosts/targets
 194 specify the value explicitly?
 195
 196 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
 197
 198 --
 199
 200 Add a ``name'' member to include/floatformat.h:struct floatformat.
 201 Print that name in gdbarch.c.
 202
 203 --
 204
 205 Sort out the harris mess in include/floatformat.h (it hardwires two
 206 different floating point formats).
 207
 208 --
 209
 210 See of the GDB local floatformat_do_doublest() and libiberty's
 211 floatformat_to_double (which was once GDB's ...) can be merged some
 212 how.
 213
 214 --
 215
 216 Eliminate mmalloc() from GDB.
 217
 218 Also eliminate it from defs.h.
 219
 220 --
 221
 222 Eliminate PTR.  ISO-C allows ``void *''.
 223
 224 --
 225
 226 Eliminate abort ().
 227
 228 GDB should never abort.  GDB should either throw ``error ()'' or
 229 ``internal_error ()''.  Better still GDB should naturally unwind with
 230 an error status.
 231
 232 --
 233
 234 GDB probably doesn't build on FreeBSD pre 2.2.x
 235 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00378.html
 236
 237 Fixes to get FreeBSD working on 2.2.x, 3.x and 4.x caused the code to
 238 suffer bit rot.
 239
 240 --
 241
 242 Updated readline
 243
 244 Readline 4.? is out.  A merge wouldn't hurt.  Patches are in:
 245
 246 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00436.html
 247
 248 --
 249
 250 Deprecate "fg".  Apparently ``fg'' is actually continue.
 251
 252 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00417.html
 253
 254 --
 255
 256 Deprecate current use of ``floatformat_unknown''.
 257
 258 Require all targets to explicitly provide their float format instead
 259 of defaulting to floatformat unknown.  Doing the latter leads to nasty
 260 bugs.
 261
 262 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
 263
 264 --
 265
 266 Rationalize floatformat_to_double() vs floatformat_to_doublest().
 267
 268 Looks like GDB migrated floatformat_to_double() to libiberty but then
 269 turned around and created a ..._to_doublest() the latter containing
 270 several bug fixes.
 271
 272 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00472.html
 273
 274 --
 275
 276 Move floatformat_ia64_ext to libiberty/include floatformat.[ch].
 277
 278 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00466.html
 279
 280 --
 281
 282 Always build ser-tcp.c.
 283
 284 The patch as submitted was just going to add ser-tcp.c to the Alpha's
 285 makefile.  A better patch is to instead add ser-tcp.c to SER_HARDWARE
 286 and make it a standard part of all debuggers.
 287
 288 If problems occure then configure.in can sort them out.
 289
 290 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00544.html
 291
 292 --
 293
 294 Follow through `make check' with --enable-shared.
 295
 296 When the srcware tree is configured with --enable-shared, the `expect'
 297 program won't run properly.  Jim Wilson found out gdb has a local hack
 298 to set LD_LIBRARY_PATH, but, AFAIK, no other project has been hacked
 299 similarly.
 300
 301 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00845.html
 302
 303 --
 304
 305 The ``maintenance deprecate set endian big'' command doesn't notice
 306 that it is deprecating ``set endian'' and not ``set endian big'' (big
 307 is implemented using an enum).  Is anyone going to notice this?
 308
 309 --
 310
 311 When tab expanding something like ``set arch<tab>'' ignore the
 312 deprecated ``set archdebug'' and expand to ``set architecture''.
 313
 314 --
 315
 316 Eliminate ``arm_register_names[j] = (char *) regnames[j]'' and the
 317 like from arm-tdep.c.
 318
 319 --
 320
 321 Fix uses of ->function.cfunc = set_function().
 322
 323 The command.c code calls sfunc() when a set command.  Rather than
 324 change it suggest fixing the callback function so that it is more
 325 useful.  See:
 326
 327 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
 328
 329 See also ``Fix implementation of  ``target xxx''.'' below.
 330
 331 --
 332
 333                         New Features and Fixes
 334                         ======================
 335
 336 These are harder than cleanups but easier than work involving
 337 fundamental architectural change.
 338
 339 --
 340
 341 Add built-by, build-date, tm, xm, nm and anything else into gdb binary
 342 so that you can see how the GDB was created.
 343
 344 Some of these (*m.h) would be added to the generated config.h.  That
 345 in turn would fix a long standing bug where by the build process many
 346 not notice a changed tm.h file.  Since everything depends on config.h,
 347 a change to *m.h forces a change to config.h and, consequently forces
 348 a rebuild.
 349
 350 --
 351
 352 Add an "info bfd" command that displays supported object formats,
 353 similarly to objdump -i.
 354
 355 Is there a command already?
 356
 357 --
 358
 359 Fix ``I'm sorry, Dave, I can't do that.'' from symfile.c.
 360
 361 This requires internationalization.
 362
 363 --
 364
 365 Convert GDB build process to AUTOMAKE.
 366
 367 See also sub-directory configure below.
 368
 369 The current convention is (kind of) to use $(<header>_h) in all
 370 dependency lists.  It isn't done in a consistent way.
 371
 372 --
 373
 374 Cleanup configury support for optional sub-directories.
 375
 376 Check how GCC handles multiple front ends for an example of how things
 377 could work.  A tentative first step is to rationalize things so that
 378 all sub directories are handled in a fashion similar to gdb/mi.
 379
 380 See also automake above.
 381
 382 --
 383
 384 Restructure gdb directory tree so that it avoids any 8.3 and 14
 385 filename problems.
 386
 387 --
 388
 389 Add a transcript mechanism to GDB.
 390
 391 Such a mechanism might log all gdb input and output to a file in a
 392 form that would allow it to be replayed.  It could involve ``gdb
 393 --transcript=FILE'' or it could involve ``(gdb) transcript file''.
 394
 395 --
 396
 397 Can the xdep files be replaced by autoconf?
 398
 399 --
 400
 401 Document trace machinery
 402
 403 --
 404
 405 Document ui-out and ui-file.
 406
 407 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00121.html
 408
 409 --
 410
 411 Update texinfo.tex to latest?
 412
 413
 414
 415 --
 416
 417 Incorporate agentexpr.texi into gdb.texinfo
 418
 419 agentexpr.texi mostly describes the details of the byte code used for
 420 tracepoints, not the internals of the support for this in GDB.  So it
 421 looks like gdb.texinfo is a better place for this information.
 422
 423 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00566.html
 424
 425 --
 426
 427 Document overlay machinery.
 428
 429 --
 430
 431 ``(gdb) catch signal SIGNAL''
 432
 433 Overlaps with ``handle SIGNAL'' but the implied behavour is different.
 434 You can attach commands to a catch but not a handle.  A handle has a
 435 limited number of hardwired actions.
 436
 437 --
 438
 439 Get the TUI working on all platforms.
 440
 441 --
 442
 443 Add support for ``gdb --- PROGRAM ARGS ...''.
 444 Add support for ``gdb -cmd=...''
 445
 446 Along with many variations.  Check:
 447
 448 ????? for a full discussion.
 449
 450 for a discussion.
 451
 452 --
 453
 454 Implement ``(gdb) !ls''.
 455
 456 Which is very different from ``(gdb) ! ls''.  Implementing the latter
 457 is trivial.
 458
 459 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00034.html
 460
 461 --
 462
 463 Change the (char *list[]) to (const char (*)[]) so that dynamic lists can
 464 be passed.
 465
 466 --
 467
 468 When tab expanding something like ``set arch<tab>'' ignore the
 469 deprecated ``set archdebug'' and expand to ``set architecture''.
 470
 471 --
 472
 473 Replace the code that uses the host FPU with an emulator of the target
 474 FPU.
 475
 476 --
 477
 478                         Thread Support
 479                         ==============
 480
 481 --
 482
 483 Generic: lin-thread cannot handle thread exit (Mark Kettenis, Michael
 484 Snyder) http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00525.html
 485
 486 The thread_db assisted debugging code doesn't handle exiting threads
 487 properly, at least in combination with glibc 2.1.3 (the framework is
 488 there, just not the actual code).  There are at least two problems
 489 that prevent this from working.
 490
 491 As an additional reference point, the pre thread_db code did not work
 492 either.
 493
 494 --
 495
 496 GNU/Linux/x86 and random thread signals (and Solaris/SPARC but not
 497 Solaris/x86).
 498 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00336.html
 499
 500 Christopher Blizzard writes:
 501
 502 So, I've done some more digging into this and it looks like Jim
 503 Kingdon has reported this problem in the past:
 504
 505 http://sourceware.cygnus.com/ml/bug-gdb/1999-10/msg00058.html
 506
 507 I can reproduce this problem both with and without Tom's patch.  Has
 508 anyone seen this before?  Maybe have a solution for it hanging around?
 509 :)
 510
 511 There's a test case for this documented at:
 512
 513 when debugging threaded applications you get extra SIGTRAPs
 514 http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=9565
 515
 516 [There should be a GDB testcase - cagney]
 517
 518 --
 519
 520 GDB5 TOT on unixware 7
 521 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00119.html
 522
 523 Robert Lipe writes:
 524 > I just spun the top of tree of the GDB5 branch on UnixWare 7.  As a
 525 > practical matter, the current thread support is somewhat more annoying
 526 > than when GDB was thread-unaware.
 527
 528 --
 529
 530 Migrate qfThreadInfo packet -> qThreadInfo. (Andrew Cagney)
 531
 532 Add support for packet enable/disable commands with these thread
 533 packets.  General cleanup.
 534
 535 [PATCH] Document the ThreadInfo remote protocol queries
 536 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00832.html
 537
 538 [PATCH] "info threads" queries for remote.c
 539 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00831.html
 540
 541 --
 542
 543                         Language Support
 544                         ================
 545
 546 New languages come onto the scene all the time.
 547
 548 --
 549
 550 Pascal (Pierre Muller, David Taylor)
 551
 552 Pierre Muller has contributed patches for adding Pascal Language
 553 support to GDB.
 554
 555 2 pascal language patches inserted in database
 556 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00521.html
 557
 558 Indent -gnu ?
 559 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00496.html
 560
 561 --
 562
 563 Java (Anthony Green, David Taylor)
 564
 565 Anthony Green has a number of Java patches that did not make it into
 566 the 5.0 release.
 567
 568 Patch: java tests
 569 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00512.html
 570
 571 Patch: java booleans
 572 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00515.html
 573
 574 Patch: handle N_MAIN stab
 575 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00527.html
 576
 577 --
 578
 579 [Comming...]
 580
 581 Modify gdb to work correctly with Pascal.
 582
 583 --
 584
 585 Re: Various C++ things
 586
 587 value_headof/value_from_vtable_info are worthless, and should be
 588 removed.  The one place in printcmd.c that uses it should use the RTTI
 589 functions.
 590
 591 RTTI for g++ should be using the typeinfo functions rather than the
 592 vtables.  The typeinfo functions are always at offset 4 from the
 593 beginning of the vtable, and are always right. The vtables will have
 594 weird names like E::VB sometimes.  The typeinfo function will always
 595 be "E type_info function", or somesuch.
 596
 597 value_virtual_fn_field needs to be fixed so there are no failures for
 598 virtual functions for C++ using g++.
 599
 600 Testsuite cases are the major priority right now for C++ support,
 601 since i have to make a lot of changes that could potentially break
 602 each other.
 603
 604 --
 605
 606 Add support for Modula3
 607
 608 Get DEC/Compaq to contribute their Modula-3 support.
 609
 610 --
 611
 612                 Remote Protocol Support
 613                 =======================
 614
 615 --
 616
 617 set/show remote X-packet ...
 618
 619 ``(gdb) help set remote X-packet'' doesn't list the applicable
 620 responses.  The help message needs to be expanded.
 621
 622 --
 623
 624 Remote protocol doco feedback.
 625
 626 Too much feedback to mention needs to be merged in (901660).  Search
 627 for the word ``remote''.
 628
 629
 630 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00023.html
 631 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00056.html
 632 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00382.html
 633
 634 --
 635
 636 GDB doesn't recover gracefully from remote protocol errors.
 637
 638 GDB wasn't checking for NAKs from the remote target.  Instead a NAK is
 639 ignored and a timeout is required before GDB retries.  A pre-cursor to
 640 fixing this this is making GDB's remote protocol packet more robust.
 641
 642 While downloading to a remote protocol target, gdb ignores packet
 643 errors in so far as it will continue to edownload with chunk N+1 even
 644 if chunk N was not correctly sent.  This causes gdb.base/remote.exp to
 645 take a painfully long time to run.  As a PS that test needs to be
 646 fixed so that it builds on 16 bit machines.
 647
 648 --
 649
 650 Add the cycle step command.
 651
 652 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00237.html
 653
 654 --
 655
 656 Resolve how to scale things to support very large packets.
 657
 658 --
 659
 660 Resolve how to handle a target that changes things like its endianess
 661 on the fly - should it be returned in the ``T'' packet?
 662
 663 Underlying problem is that the register file is target endian.  If the
 664 target endianess changes gdb doesn't know.
 665
 666 --
 667
 668                         Symbol Support
 669                         ==============
 670
 671 If / when GDB starts to support the debugging of multi-processor
 672 (rather than multi-thread) applications the symtab code will need to
 673 be updated a little so that several independant symbol tables are
 674 active at a given time.
 675
 676 The other interesting change is a clarification of the exact meaning
 677 of CORE_ADDR and that has had consequences for a few targets (that
 678 were abusing that data type).
 679
 680 --
 681
 682 Investiagate ways of reducing memory.
 683
 684 --
 685
 686 Investigate ways of improving load time.
 687
 688 --
 689
 690 Get the d10v to use POINTER_TO_ADDRESS and ADDRESS_TO_POINTER.
 691
 692 Consequence of recent symtab clarification.  No marks for figuring out
 693 who maintains the d10v.
 694
 695 --
 696
 697 Get the MIPS to correctly sign extend all address <-> pointer
 698 conversions.
 699
 700 Consequence of recent symtab clarification.  No marks for figuring out
 701 who maintains the MIPS.
 702
 703 --
 704
 705                         Testsuite Support
 706                         =================
 707
 708 There are never to many testcases.
 709
 710 --
 711
 712 Better thread testsuite.
 713
 714 --
 715
 716 Better C++ testsuite.
 717
 718 --
 719
 720 Look at adding a GDB specific testsuite directory so that white box
 721 tests of key internals can be added (eg ui_file).
 722
 723 --
 724
 725 Separate out tests that involve the floating point (FP).
 726
 727 (Something for people brining up new targets).  FP and non-fp tests
 728 are combined.  I think there should be set of basic tests that
 729 exercise pure integer support and then a more expanded set that
 730 exercise FP and FP/integer interactions.
 731
 732 As an example, the MIPS, for n32 as problems with passing FP's and
 733 structs.  Since most inferior call tests include FP it is difficult to
 734 determine of the integer tests are ok.
 735
 736 --
 737
 738                 Architectural Changes: General
 739                 ==============================
 740
 741 These are harder than simple cleanups / fixes and, consequently
 742 involve more work.  Typically an Architectural Change will be broken
 743 down into a more digestible set of cleanups and fixes.
 744
 745 --
 746
 747 Cleanup software single step.
 748
 749 At present many targets implement software single step by directly
 750 blatting memory (see rs6000-tdep.c).  Those targets should register
 751 the applicable breakpoints using the breakpoint framework.  Perhaphs a
 752 new internal breakpoint class ``step'' is needed.
 753
 754 --
 755
 756 Replace READ_FP() with FRAME_HANDLE().
 757
 758 READ_FP() is a hangover from the days of the vax when the ABI really
 759 did have a frame pointer register.  Modern architectures typically
 760 construct a virtual frame-handle from the stack pointer and various
 761 other bits of string.
 762
 763 Unfortunatly GDB still treats this synthetic FP register as though it
 764 is real.  That in turn really confuses users (arm and ``print $fp'' VS
 765 ``info registers fp'').  The synthetic FP should be separated out of
 766 the true register set presented to the user.
 767
 768 --
 769
 770 Register Cache Cleanup (below from Andrew Cagney)
 771
 772 I would depict the current register architecture as something like:
 773
 774         High GDB  -->   Low GDB
 775             |              |
 776            \|/            \|/
 777             --- REG NR -----
 778                    |
 779     register + REGISTER_BYTE(reg_nr)
 780                    |
 781                   \|/
 782         -------------------------
 783         |  extern register[]    |
 784         -------------------------
 785
 786 where neither the high (valops.c et.al.) or low gdb (*-tdep.c) are
 787 really clear on what mechanisms they should be using to manipulate that
 788 buffer.   Further, much code assumes, dangerously, that registers are
 789 contigious.  Having got mips-tdep.c to support multiple ABIs, believe
 790 me, that is a bad assumption.  Finally, that register cache layout is
 791 determined by the current remote/local target and _not_ the less
 792 specific target ISA.  In fact, in many cases it is determined by the
 793 somewhat arbitrary layout of the [gG] packets!
 794
 795
 796 How I would like the register file to work is more like:
 797
 798
 799           High GDB
 800              |
 801             \|/
 802        pseudo reg-nr
 803              |
 804         map pseudo <->
 805         random cache
 806         bytes
 807              |
 808             \|/
 809         ------------
 810         | register |
 811         |  cache   |
 812         ------------
 813             /|\
 814              |
 815         map random cache
 816         bytes to target
 817         dependant i-face
 818             /|\
 819              |
 820        target dependant
 821         such as [gG] packet
 822         or ptrace buffer
 823
 824 The main objectives being:
 825
 826         o       a clear separation between the low
 827                 level target and the high level GDB
 828
 829         o       a mechanism that solves the general
 830                 problem of register aliases, overlaps
 831                 etc instead of treating them as optional
 832                 extras that can be wedged in as an after
 833                 thought (that is a reasonable description
 834                 of the current code).
 835
 836                 Identify then solve the hard case and the
 837                 rest just falls out.  GDB solved the easy
 838                 case and then tried to ignore the real
 839                 world :-)
 840
 841         o       a removal of the assumption that the
 842                 mapping between the register cache
 843                 and virtual registers is largely static.
 844                 If you flip the USR/SSR stack register
 845                 select bit in the status-register then
 846                 the corresponding stack registers should
 847                 reflect the change.
 848
 849         o       a mechanism that clearly separates the
 850                 gdb internal register cache from any
 851                 target (not architecture) dependant
 852                 specifics such as [gG] packets.
 853
 854 Of course, like anything, it sounds good in theory.  In reality, it
 855 would have to contend with many<->many relationships at both the
 856 virt<->cache and cache<->target level.  For instance:
 857
 858         virt<->cache
 859                 Modifying an mmx register may involve
 860                 scattering values across both FP and
 861                 mmpx specific parts of a buffer
 862
 863         cache<->target
 864                 When writing back a SP it may need to
 865                 both be written to both SP and USP.
 866
 867
 868 Hmm,
 869
 870 Rather than let this like the last time it was discussed, just slip, I'm
 871 first going to add this e-mail (+ references) to TODO.  I'd then like to
 872 sketch out a broad strategy I think could get us there.
 873
 874
 875 First thing I'd suggest is separating out the ``extern registers[]''
 876 code so that we can at least identify what is using it.  At present
 877 things are scattered across many files.  That way we can at least
 878 pretend that there is a cache instead of a global array :-)
 879
 880 I'd then suggest someone putting up a proposal for the pseudo-reg /
 881 high-level side interface so that code can be adopted to it.  For old
 882 code, initially a blanket rename of write_register_bytes() to
 883 deprecated_write_register_bytes() would help.
 884
 885 Following that would, finaly be the corresponding changes to the target.
 886
 887 --
 888
 889 Check that GDB can handle all BFD architectures (Andrew Cagney)
 890
 891 There should be a test that checks that BFD/GDB are in sync with
 892 regard to architecture changes.  Something like a test that first
 893 queries GDB for all supported architectures and then feeds each back
 894 to GDB..  Anyone interested in learning how to write tests?  :-)
 895
 896 --
 897
 898                 Architectural Change: Multi-arch et al.
 899                 =======================================
 900
 901 The long term objective is to remove all assumptions that there is a
 902 single target with a single address space with a single instruction
 903 set architecture and single application binary interface.
 904
 905 This is an ongoing effort.  The first milestone is to enable
 906 ``multi-arch'' where by all architectural decisions are made at
 907 runtime.
 908
 909 It should be noted that ``gdbarch'' is really ``gdbabi'' and
 910 ``gdbisa''.  Once things are multi-arched breaking that down correctly
 911 will become much easier.
 912
 913 --
 914
 915 GDBARCH cleanup (Andrew Cagney)
 916
 917 The non-generated parts of gdbarch.{sh,h,c} should be separated out
 918 into arch-utils.[hc].
 919
 920 Document that gdbarch_init_ftype could easily fail because it didn't
 921 identify an architecture.
 922
 923 --
 924
 925 Fix BELIEVE_PPC_PROMOTION.  Change it to BELIEVE_PPC_PROMOTION_P?
 926
 927 At present there is still #ifdef BELIEVE_PPC_PROMOTION code in the
 928 symtab file.
 929
 930 --
 931
 932 Fix target_signal_from_host() etc.
 933
 934 The name is wrong for starters.  ``target_signal'' should probably be
 935 ``gdb_signal''.  ``from_host'' should be ``from_target_signal''.
 936 After that it needs to be multi-arched and made independant of any
 937 host signal numbering.
 938
 939 --
 940
 941 Update ALPHA so that it uses ``struct frame_extra_info'' instead of
 942 EXTRA_FRAME_INFO.
 943
 944 This is a barrier to replacing mips_extra_func_info with something
 945 that works with multi-arch.
 946
 947 --
 948
 949 Multi-arch  mips_extra_func_info.
 950
 951 This first needs the alpha to be updated so that it uses ``struct
 952 frame_extra_info''.
 953
 954 --
 955
 956 Rationalize TARGET_SINGLE_FORMAT and TARGET_SINGLE_BIT et al.
 957
 958 Surely one of them is redundant.
 959
 960 --
 961
 962 Convert ALL architectures to MULTI-ARCH.
 963
 964 --
 965
 966 Select the initial multi-arch ISA / ABI based on --target or similar.
 967
 968 At present the default is based on what ever is first in the BFD
 969 archures table.  It should be determined based on the ``--target=...''
 970 name.
 971
 972 --
 973
 974 Truly multi-arch.
 975
 976 Enable the code to recognize --enable-targets=.... like BINUTILS does.
 977
 978 Can the tm.h and nm.h files be eliminated by multi-arch.
 979
 980 --
 981
 982         Architectural Change: MI, LIBGDB and scripting languages
 983         ========================================================
 984
 985 See also architectural changes related to the event loop.  LIBGDB
 986 can't be finished until there is a generic event loop being used by
 987 all targets.
 988
 989 The long term objective is it to be possible to integrate GDB into
 990 scripting languages.
 991
 992 --
 993
 994 Implement generic ``(gdb) commmand > file''
 995
 996 Once everything is going through ui_file it should be come fairly
 997 easy.
 998
 999 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00104.html
1000
1001 --
1002
1003 Replace gdb_stdtarg with gdb_targout (and possibly gdb_targerr).
1004
1005 gdb_stdtarg is easily confused with gdb_stdarg.
1006
1007 --
1008
1009 Extra ui_file methods - dump.
1010
1011 Very useful for whitebox testing.
1012
1013 --
1014
1015 Eliminate error_begin().
1016
1017 With ui_file, there is no need for the statefull error_begin ()
1018 function.
1019
1020 --
1021
1022 Send normal output to gdb_stdout.
1023 Send error messages to gdb_stderror.
1024 Send debug and log output log gdb_stdlog.
1025
1026 GDB still contains many cases where (f)printf or printf_filtered () is
1027 used when it should be sending the messages to gdb_stderror or
1028 gdb_stdlog.  The thought of #defining printf to something has crossed
1029 peoples minds ;-)
1030
1031 --
1032
1033 Re-do GDB's output pager.
1034
1035 GDB's output pager still relies on people correctly using *_filtered
1036 for gdb_stdout and *_unfiltered for gdb_stdlog / gdb_stderr.
1037 Hopefully, with all normal output going to gdb_stdout, the pager can
1038 just look at the ui_file that the output is on and then use that to
1039 decide what to do about paging.  Sounds good in theory.
1040
1041 --
1042
1043 Check/cleanup MI documentation.
1044
1045 The list of commands specified in the documentation needs to be
1046 checked against the mi-cmds.c table in a mechanical way (so that they
1047 two can be kept up-to-date).
1048
1049 --
1050
1051 Convert MI into libgdb
1052
1053 MI provides a text interface into what should be many of the libgdb
1054 functions.  The implementation of those functions should be separated
1055 into the MI interface and the functions proper.  Those functions being
1056 moved to gdb/lib say.
1057
1058 --
1059
1060 Create libgdb.h
1061
1062 The first part can already be found in defs.h.
1063
1064 --
1065
1066 MI's input does not use buffering.
1067
1068 At present the MI interface reads raw characters of from an unbuffered
1069 FD.  This is to avoid several nasty buffer/race conditions.  That code
1070 should be changed so that it registers its self with the event loop
1071 (on the input FD) and then push commands up to MI as they arrive.
1072
1073 The serial code already does this.
1074
1075 --
1076
1077 Make MI interface accessable from existing CLI.
1078
1079 --
1080
1081 Add a breakpoint-edit command to MI.
1082
1083 It would be similar to MI's breakpoint create but would apply to an
1084 existing breakpoint.  It saves the need to delete/create breakpoints
1085 when ever they are changed.
1086
1087 --
1088
1089 Add directory path to MI breakpoint.
1090
1091 That way the GUI's task of finding the file within which the
1092 breakpoint was set is simplified.
1093
1094 --
1095
1096 Add a mechanism to reject certain expression classes to MI
1097
1098 There are situtations where you don't want GDB's expression
1099 parser/evaluator to perform inferior function calls or variable
1100 assignments.  A way of restricting the expression parser so that such
1101 operations are not accepted would be very helpful.
1102
1103 --
1104
1105 Remove sideffects from libgdb breakpoint create function.
1106
1107 The user can use the CLI to create a breakpoint with partial
1108 information - no file (gdb would use the file from the last
1109 breakpoint).
1110
1111 The libgdb interface currently affects that environment which can lead
1112 to confusion when a user is setting breakpoints via both the MI and
1113 the CLI.
1114
1115 This is also a good example of how getting the CLI ``right'' will be
1116 hard.
1117
1118 --
1119
1120 Move gdb_lasterr to ui_out?
1121
1122 The way GDB throws errors and records them needs a re-think.  ui_out
1123 handles the correct output well.  It doesn't resolve what to do with
1124 output / error-messages when things go wrong.
1125
1126 --
1127
1128                 Architectural Change: Async
1129                 ===========================
1130
1131 While GDB uses an event loop when prompting the user for input.  That
1132 event loop is not exploited by targets when they allow the target
1133 program to continue.  Typically targets still block in (target_wait())
1134 until the program again halts.
1135
1136 The closest a target comes to supporting full asynchronous mode are
1137 the remote targets ``async'' and ``extended-async''.
1138
1139 --
1140
1141 Asynchronous expression evaluator
1142
1143 Inferior function calls hang GDB.
1144
1145 --
1146
1147 Fix implementation of  ``target xxx''.
1148
1149 At present when the user specifies ``target xxxx'', the CLI maps that
1150 directly onto a target open method.  It is then assumed that the
1151 target open method should do all sorts of complicated things as this
1152 is the only chance it has.  Check how the various remote targets
1153 duplicate the target operations.  Check also how the various targets
1154 behave differently for purely arbitrary reasons.
1155
1156 What should happen is that ``target xxxx'' should call a generic
1157 ``target'' function and that should then co-ordinate the opening of
1158 ``xxxx''.  This becomes especially important when you're trying to
1159 open an asynchronous target that may need to perform background tasks
1160 as part of the ``attach'' phase.
1161
1162 Unfortunatly, due to limitations in the old/creaking command.h
1163 interface, that isn't possible.  The function being called isn't told
1164 of the ``xxx'' or any other context information.
1165
1166 Consequently a precursor to fixing ``target xxxx'' is to clean up the
1167 CLI code so that it passes to the callback function (attatched to a
1168 command) useful information such as the actual command and a context
1169 for that command.  Other changes such as making ``struct command''
1170 opaque may also help.
1171
1172 See also:
1173 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
1174
1175 --
1176
1177 Make "target xxx" command interruptible.
1178
1179 As things become async this becomes possible.  A target would start
1180 the connect and then return control to the event loop.  A cntrl-c
1181 would notify the target that the operation is to be abandoned and the
1182 target code could respond.
1183
1184 --
1185
1186 Add a "suspend" subcommand of the "continue" command to suspend gdb
1187 while continuing  execution of the subprocess.  Useful when you are
1188 debugging servers and you want to dodge out and initiate a connection
1189 to a server running under gdb.
1190
1191 [hey async!!]
1192
1193 --
1194
1195                         TODO FAQ
1196                         ========
1197
1198 Frequently requested but not approved requests.
1199
1200 --
1201
1202 Eliminate unused argument warnings using ATTRIBUTE_UNUSED.
1203
1204 The benefits on this one are thought to be marginal - GDBs design
1205 means that unused parameters are very common.  GCC 3.0 will also
1206 include the option -Wno-unused-parameter which means that ``-Wall
1207 -Wno-unused-parameters -Werror'' can be specified.
1208
1209 --
1210
1211
1212
1213                         Legacy Wish List
1214                         ================
1215
1216 This list is not up to date, and opinions vary about the importance or
1217 even desirability of some of the items.  If you do fix something, it
1218 always pays to check the below.
1219
1220 --
1221
1222 @c This does not work (yet if ever).  FIXME.
1223 @c @item --parse=@var{lang} @dots{}
1224 @c Configure the @value{GDBN} expression parser to parse the listed languages.
1225 @c @samp{all} configures @value{GDBN} for all supported languages.  To get a
1226 @c list of all supported languages, omit the argument.  Without this
1227 @c option, @value{GDBN} is configured to parse all supported languages.
1228
1229 --
1230
1231 START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that
1232 is its default value.  Clean this up.
1233
1234 --
1235
1236 It should be possible to use symbols from shared libraries before we know
1237 exactly where the libraries will be loaded.  E.g. "b perror" before running
1238 the program.  This could maybe be done as an extension of the "breakpoint
1239 re-evaluation" after new symbols are loaded.
1240
1241 --
1242
1243 Make single_step() insert and remove breakpoints in one operation.
1244
1245 [If this is talking about having single_step() insert the breakpoints,
1246 run the target then pull the breakpoints then it is wrong.  The
1247 function has to return as control has to eventually be passed back to
1248 the main event loop.]
1249
1250 --
1251
1252 Speed up single stepping by avoiding extraneous ptrace calls.
1253
1254 --
1255
1256 Speed up single stepping by not inserting and removing breakpoints
1257 each time the inferior starts and stops.
1258
1259 Breakpoints should not be inserted and deleted all the time.  Only the
1260 one(s) there should be removed when we have to step over one.  Support
1261 breakpoints that don't have to be removed to step over them.
1262
1263 [this has resulted in numerous debates.  The issue isn't clear cut]
1264
1265 --
1266
1267 Provide "voodoo" debugging of core files.  This creates a zombie
1268 process as a child of the debugger, and loads it up with the data,
1269 stack, and regs of the core file.  This allows you to call functions
1270 in the executable, to manipulate the data in the core file.
1271
1272 [you wish]
1273
1274 --
1275
1276 GDB reopens the source file on every line, as you "next" through it.
1277
1278 [still true? I've a memory of this being fixed]
1279
1280 --
1281
1282 Perhaps "i source" should take an argument like that of "list".
1283
1284 --
1285
1286 Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if
1287 it matches the source line indicated.
1288
1289 --
1290
1291 The prompt at end of screen should accept space as well as CR.
1292
1293 --
1294
1295 Backtrace should point out what the currently selected frame is, in
1296 its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar,
1297 ...)" rather than "#3 foo (bar, ...)".
1298
1299 --
1300
1301 "i program" should work for core files, and display more info, like what
1302 actually caused it to die.
1303
1304 --
1305
1306 "x/10i" should shorten the long name, if any, on subsequent lines.
1307
1308 --
1309
1310 "next" over a function that longjumps, never stops until next time you happen
1311 to get to that spot by accident.  E.g. "n" over execute_command which has
1312 an error.
1313
1314 --
1315
1316 "set zeroprint off", don't bother printing members of structs which
1317 are entirely zero.  Useful for those big structs with few useful
1318 members.
1319
1320 --
1321
1322 GDB does four ioctl's for every command, probably switching terminal modes
1323 to/from inferior or for readline or something.
1324
1325 --
1326
1327 terminal_ours versus terminal_inferior: cache state.  Switch should be a noop
1328 if the state is the same, too.
1329
1330 --
1331
1332 "i frame" shows wrong "arglist at" location, doesn't show where the args
1333 should be found, only their actual values.
1334
1335 --
1336
1337 There should be a way for "set" commands to validate the new setting
1338 before it takes effect.
1339
1340 --
1341
1342 "ena d" is ambiguous, why?  "ena delete" seems to think it is a command!
1343
1344 --
1345
1346 i line VAR produces "Line number not known for symbol ``var''.".  I
1347 thought we were stashing that info now!
1348
1349 --
1350
1351 We should be able to write to random files at hex offsets like adb.
1352
1353 --
1354
1355 [elena - delete this]
1356
1357 Handle add_file with separate text, data, and bss addresses.  Maybe
1358 handle separate addresses for each segment in the object file?
1359
1360 --
1361
1362 [Jimb/Elena delete this one]
1363
1364 Handle free_named_symtab to cope with multiply-loaded object files
1365 in a dynamic linking environment.  Should remember the last copy loaded,
1366 but not get too snowed if it finds references to the older copy.
1367
1368 --
1369
1370 [elena delete this also]
1371
1372 Remove all references to:
1373         text_offset
1374         data_offset
1375         text_data_start
1376         text_end
1377         exec_data_offset
1378         ...
1379 now that we have BFD.  All remaining are in machine dependent files.
1380
1381 --
1382
1383 Re-organize help categories into things that tend to fit on a screen
1384 and hang together.
1385
1386 --
1387
1388 Add in commands like ADB's for searching for patterns, etc.  We should
1389 be able to examine and patch raw unsymboled binaries as well in gdb as
1390 we can in adb.  (E.g. increase the timeout in /bin/login without source).
1391
1392 [actually, add ADB interface :-]
1393
1394 --
1395
1396 When doing "step" or "next", if a few lines of source are skipped between
1397 the previous line and the current one, print those lines, not just the
1398 last line of a multiline statement.
1399
1400 --
1401
1402 Handling of "&" address-of operator needs some serious overhaul
1403 for ANSI C and consistency on arrays and functions.
1404   For "float point[15];":
1405 ptype &point[4]   ==> Attempt to take address of non-lvalue.
1406   For "char *malloc();":
1407 ptype malloc      ==> "char *()";  should be same as
1408 ptype &malloc     ==> "char *(*)()"
1409 call printf ("%x\n", malloc)   ==> weird value, should be same as
1410 call printf ("%x\n", &malloc)  ==> correct value
1411
1412 --
1413
1414 Fix dbxread.c symbol reading in the presence of interrupts.  It
1415 currently leaves a cleanup to blow away the entire symbol table when a
1416 QUIT occurs.  (What's wrong with that? -kingdon, 28 Oct 1993).
1417
1418 [I suspect that the grype was that, on a slow system, you might want
1419 to cntrl-c and get just half the symbols and then load the rest later
1420 - scary to be honest]
1421
1422 --
1423
1424 Mipsread.c reads include files depth-first, because the dependencies
1425 in the psymtabs are way too inclusive (it seems to me).  Figure out what
1426 really depends on what, to avoid recursing 20 or 30 times while reading
1427 real symtabs.
1428
1429 --
1430
1431 value_add() should be subtracting the lower bound of arrays, if known,
1432 and possibly checking against the upper bound for error reporting.
1433
1434 --
1435
1436 When listing source lines, check for a preceding \n, to verify that
1437 the file hasn't changed out from under us.
1438
1439 [fixed by some other means I think.  That hack wouldn't actually work
1440 reliably - the file might move such that another \n appears. ]
1441
1442 --
1443
1444 Get all the remote systems (where the protocol allows it) to be able to
1445 stop the remote system when the GDB user types ^C (like remote.c
1446 does).  For ebmon, use ^Ak.
1447
1448 --
1449
1450 Possible feature: A version of the "disassemble" command which shows
1451 both source and assembly code ("set symbol-filename on" is a partial
1452 solution).
1453
1454 [has this been done?  It was certainly done for MI and GDBtk]
1455
1456 --
1457
1458 investigate "x/s 0" (right now stops early) (I think maybe GDB is
1459 using a 0 address for bad purposes internally).
1460
1461 --
1462
1463 Make "info path" and path_command work again (but independent of the
1464 environment either of gdb or that we'll pass to the inferior).
1465
1466 --
1467
1468 Make GDB understand the GCC feature for putting octal constants in
1469 enums.  Make it so overflow on an enum constant does not error_type
1470 the whole type.  Allow arbitrarily large enums with type attributes.
1471 Put all this stuff in the testsuite.
1472
1473 --
1474
1475 Make TYPE_CODE_ERROR with a non-zero TYPE_LENGTH more useful (print
1476 the value in hex; process type attributes).  Add this to the
1477 testsuite.  This way future compilers can add new types and old
1478 versions of GDB can do something halfway reasonable.
1479
1480 --
1481
1482 Fix mdebugread.c:parse_type to do fundamental types right (see
1483 rs6000_builtin_type in stabsread.c for what "right" is--the point is
1484 that the debug format fixes the sizes of these things and it shouldn't
1485 depend on stuff like TARGET_PTR_BIT and so on.  For mdebug, there seem
1486 to be separate bt* codes for 64 bit and 32 bit things, and GDB should
1487 be aware of that).  Also use a switch statement for clarity and speed.
1488
1489 --
1490
1491 Investigate adding symbols in target_load--some targets do, some
1492 don't.
1493
1494 --
1495
1496 Put dirname in psymtabs and change lookup*symtab to use dirname (so
1497 /foo/bar.c works whether compiled by cc /foo/bar.c, or cd /foo; cc
1498 bar.c).
1499
1500 --
1501
1502 Merge xcoffread.c and coffread.c.  Use breakpoint_re_set instead of
1503 fixup_breakpoints.
1504
1505 --
1506
1507 Make a watchpoint which contains a function call an error (it is
1508 broken now, making it work is probably not worth the effort).
1509
1510 --
1511
1512 New test case based on weird.exp but in which type numbers are not
1513 renumbered (thus multiply defining a type).  This currently causes an
1514 infinite loop on "p v_comb".
1515
1516 --
1517
1518 [Hey! Hint Hint Delete Delete!!!]
1519
1520 Fix 386 floating point so that floating point registers are real
1521 registers (but code can deal at run-time if they are missing, like
1522 mips and 68k).  This would clean up "info float" and related stuff.
1523
1524 --
1525
1526 gcc -g -c enummask.c then gdb enummask.o, then "p v".  GDB complains
1527 about not being able to access memory location 0.
1528
1529 -------------------- enummask.c
1530 enum mask
1531 {
1532   ANIMAL = 0,
1533   VEGETABLE = 1,
1534   MINERAL = 2,
1535   BASIC_CATEGORY = 3,
1536
1537   WHITE = 0,
1538   BLUE = 4,
1539   GREEN = 8,
1540   BLACK = 0xc,
1541   COLOR = 0xc,
1542
1543   ALIVE = 0x10,
1544
1545   LARGE = 0x20
1546 } v;
1547
1548 --
1549
1550 If try to modify value in file with "set write off" should give
1551 appropriate error not "cannot access memory at address 0x65e0".
1552
1553 --
1554
1555 Allow core file without exec file on RS/6000.
1556
1557 --
1558
1559 Make sure "shell" with no arguments works right on DOS.
1560
1561 --
1562
1563 Make gdb.ini (as well as .gdbinit) be checked on all platforms, so
1564 the same directory can be NFS-mounted on unix or DOS, and work the
1565 same way.
1566
1567 --
1568
1569 [Is this another delete???]
1570
1571 Get SECT_OFF_TEXT stuff out of objfile_relocate (might be needed to
1572 get RS/6000 to work right, might not be immediately relevant).
1573
1574 --
1575
1576 Work out some kind of way to allow running the inferior to be done as
1577 a sub-execution of, eg. breakpoint command lists.  Currently running
1578 the inferior interupts any command list execution.  This would require
1579 some rewriting of wait_for_inferior & friends, and hence should
1580 probably be done in concert with the above.
1581
1582 --
1583
1584 Add function arguments to gdb user defined functions.
1585
1586 --
1587
1588 Add convenience variables that refer to exec file, symbol file,
1589 selected frame source file, selected frame function, selected frame
1590 line number, etc.
1591
1592 --
1593
1594 Modify the handling of symbols grouped through BINCL/EINCL stabs to
1595 allocate a partial symtab for each BINCL/EINCL grouping.  This will
1596 seriously decrease the size of inter-psymtab dependencies and hence
1597 lessen the amount that needs to be read in when a new source file is
1598 accessed.
1599
1600 --
1601
1602 Add a command for searching memory, a la adb.  It specifies size,
1603 mask, value, start address.  ADB searches until it finds it or hits
1604 an error (or is interrupted).
1605
1606 --
1607
1608 Remove the range and type checking code and documentation, if not
1609 going to implement.
1610
1611 # Local Variables:
1612 # mode: text
1613 # End: