gdb/TODO

   1 If you find inaccuracies in this list, please send mail to
   2 gdb-patches@sourceware.cygnus.com.  If you would like to work on any
   3 of these, you should consider sending mail to the same address, to
   4 find out whether anyone else is working on it.
   5
   6
   7                         GDB 5.1 - Fixes
   8                         ===============
   9
  10 Below is a list of problems identified during the GDB 5.0 release
  11 cycle.  People hope to have these problems fixed in 5.1.
  12
  13 -- 2001-03-08
  14
  15 Update GDB's coding standard documentation.  Known topics:
  16
  17        o     alloca/malloc et.al.
  18
  19        o     typedef and structs
  20
  21        o        ISO-C
  22
  23 and most likely also:
  24
  25        o         include conventions
  26
  27 --
  28
  29 Wow, three bug reports for the same problem in one day!  We should
  30 probably make fixing this a real priority :-).
  31
  32 Anyway, thanks for reporting.
  33
  34 The following patch will fix the problems with setting breakpoints in
  35 dynamically loaded objects:
  36
  37    http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00230.html
  38
  39 This patch isn't checked in yet (ping Michael/JimB), but I hope this
  40 will be in the next GDB release.
  41
  42 There should really be a test in the testsuite for this problem, since
  43 it keeps coming up :-(.  Any volunteers?
  44
  45 Mark
  46
  47 --
  48
  49 x86 linux GDB and SIGALRM (???)
  50 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00803.html
  51
  52 This problem has been fixed, but a regression test still needs to be
  53 added to the testsuite:
  54 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00309.html
  55
  56 Mark
  57
  58 [The test has been submitted for approval - cagney]
  59
  60 --
  61
  62 RFD: infrun.c: No bpstat_stop_status call after proceed over break?
  63 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00665.html
  64
  65 GDB misses watchpoint triggers after proceeding over a breakpoint on
  66 x86 targets.
  67
  68 --
  69
  70 GDB 5.0 doesn't work on Linux/SPARC
  71
  72 There are two parts to this.
  73
  74       o   GDB 5.0 doesn't work on GNU/Linux/SPARC32
  75
  76       o   GDB 5.0 doesn't work on the new target
  77           GNU/Linux/SPARC64
  78
  79 GDB does build on both these targets.
  80
  81 The first problem is the one that should be fixed.
  82
  83 --
  84
  85                 GDB 5.1 - New features
  86                 ======================
  87
  88 The following new features should be included in 5.1.
  89
  90 --
  91
  92 Enable MI by default.  Old code can be deleted after 5.1 is out.
  93
  94 Issues:
  95
  96         o       syntax change where a list would
  97                 look like:
  98                      [ foo=a, foo=b, foo=c ]
  99                 instead of
 100                      { foo=a, foo=b, foo=c }
 101
 102         o       kill off the idea of a reverse
 103                 query.
 104
 105         o       review test cases
 106
 107         o       enable it
 108
 109 --
 110
 111 Pascal (Pierre Muller, David Taylor)
 112
 113 Pierre Muller has contributed patches for adding Pascal Language
 114 support to GDB.
 115
 116 2 pascal language patches inserted in database
 117 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00521.html
 118
 119 Indent -gnu ?
 120 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00496.html
 121
 122 [I think this has been merged, need to confirm - cagney]
 123
 124 --
 125
 126 Java (Anthony Green, David Taylor)
 127
 128 Anthony Green has a number of Java patches that did not make it into
 129 the 5.0 release.  The first two are in cvs now, but the third needs
 130 some fixing up before it can go in.
 131
 132 Patch: java tests
 133 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00512.html
 134
 135 Patch: java booleans
 136 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00515.html
 137
 138 Patch: handle N_MAIN stab
 139 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00527.html
 140
 141 -- 2001-03-08
 142
 143 Add CRIS target.
 144
 145 A predicate to this is the multi-arching of SOFTWARE_SINGLE_STEP().  A
 146 patch has been submitted.
 147
 148 --
 149
 150                 GDB 5.1 - Cleanups
 151                 ==================
 152
 153 The following code cleanups will hopefully be applied to GDB 5.1.
 154
 155 -- 2001-03-26
 156
 157 Resolve the build status of all broken targets as identified by the
 158 MAINTAINERS file.
 159
 160         o       arm-* vs NetBSD's lack of ``unix''
 161         o       arm-* vs IRIX (see below)
 162
 163         o       delete mpw?
 164
 165 -- 2001-03-15
 166
 167                         Obsolete some targets.
 168
 169 Possible selection criteria are:
 170
 171         o       uses a deprecated feature
 172
 173         o       doesn't build
 174
 175         o       doesn't have a maintainer
 176
 177 Steps:
 178
 179         o       Identify TUPPLE : CPU / HOST / TARGET
 180                 combinations.
 181
 182                   TUPPLE                CPU      HOST     TARGET     ZAPPED
 183                   ------                ---      ----     ------     ------
 184                   ns32k-*-mach3*        ns32k    ns32km3  ns32km3    yes
 185                   ns32k-umax-*          ns32k    umax     -          yes
 186                   ns32k-utek-sysv*      ns32k    merlin   merlin     yes
 187                   ns32k-utek-*          ns32k    -        umax       yes
 188                   tic80-*               tic80    -        tic80
 189                   m68*-isi-*            m68k     isi      isi
 190                   m68*-sony-*           m68k     news     news
 191                   m68030-sony-*         m68k     news1000 -
 192                   m68*-rom68k-*         m68k     -        (monitor)
 193                   m68*-*bug-*           m68k     -        (monitor)
 194                   m68*-monitor-*        m68k     -        (monitor)
 195                   m68*-est-*            m68k     -        (monitor)
 196                   a29k-*-*              a29k     ultra3   -
 197                   a29k-*-sym1*          a29k     -        ultra3
 198                   powerpcle-*-solaris*  powerpc  solaris  solaris
 199                   powerpcle-*-cygwin*   powerpc  cygwin   cygwin
 200                   powerpc-*-netware*    powerpc  -        ppc-nw
 201                   w65-*-*               w65      -        w65
 202                   i[3456]86-*-sunos*    i386     sun386   sun386
 203
 204                 (DONE)
 205
 206         o       post proposals to gdb@
 207                 (DONE)
 208
 209         o       post revised proposals to gdb-announce@
 210                 crossed with gdb@ reply-to to gdb@
 211                 (DONE)
 212
 213         o       clobber the tupple and files:
 214                         configure.host
 215                         configure.tgt
 216                         config/CPU/*HOST*
 217                         config/CPU/*TARGET*
 218                         *-tdep.c *-nat.c *-xdep.c
 219
 220         o       update NEWS
 221
 222 --
 223
 224 Fix copyright notices.
 225
 226 Turns out that ``1998-2000'' isn't considered valid :-(
 227
 228 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00467.html
 229
 230 --
 231
 232                 GDB 5.1 - Known Problems
 233                 ========================
 234
 235 --
 236
 237 z8k
 238
 239 The z8k has suffered bit rot and is known to not build.  The problem
 240 was occuring in the opcodes directory.
 241
 242 --
 243
 244 m88k
 245
 246 The m88k has suffered bit rot and is known to not build.
 247
 248 --
 249
 250 Solaris 8 x86 CURSES_H problem
 251 http://sources.redhat.com/ml/gdb/2000-07/msg00038.html
 252
 253 The original problem was worked around with:
 254
 255     2000-06-06  Michael Snyder  <msnyder@cygnus.com>
 256
 257         * configure.in: Enable autoconf to find curses.h on Solaris 2.8.
 258         * configure: Regenerate.
 259
 260 When building both GDB and SID using the same source tree the problem
 261 will still occure. sid/component/configure.in mis-configures
 262 <curses.h> and leaves wrong information in the config cache.
 263
 264 --
 265
 266                 GDB 5.2 - Fixes
 267                 ===============
 268
 269 --
 270
 271 Thread support.  Right now, as soon as a thread finishes and exits,
 272 you're hosed.  This problem is reported once a week or so.
 273
 274 --
 275
 276                 GDB 5.2 - New features
 277                 ======================
 278
 279 --
 280
 281 GCC 3.0 ABI support (but hopefully sooner...).
 282
 283 --
 284
 285 Objective C/C++ support (but hopefully sooner...).
 286
 287 --
 288
 289                 GDB 5.2 - Cleanups
 290                 ==================
 291
 292 The following cleanups have been identified as part of GDB 5.2.
 293
 294 --
 295
 296 Remove old code that does not use ui_out functions and all the related
 297 "ifdef"s.  This also allows the elimination of -DUI_OUT from
 298 Makefile.in and configure.in.
 299
 300 --
 301
 302 Compiler warnings.
 303
 304 Eliminate all warnings for at least one host/target for the flags:
 305 -Wimplicit -Wreturn-type -Wcomment -Wtrigraphs -Wformat -Wparentheses
 306 -Wpointer-arith -Wuninitialized
 307
 308 --
 309
 310 Deprecate, if not delete, the following:
 311
 312         register[]
 313         register_valid[]
 314         register_buffer()
 315         REGISTER_BYTE()
 316                 Replaced by, on the target side
 317                   supply_register()
 318                 and on core-gdb side:
 319                   {read,write}_register_gen()
 320                 Remote.c will need to use something
 321                 other than REGISTER_BYTE() and
 322                 REGISTER_RAW_SIZE() when unpacking
 323                 [gG] packets.
 324
 325         STORE_PSEUDO_REGISTER
 326         FETCH_PSEUDO_REGISTER
 327                 Now handed by the methods
 328                   gdbarch_{read,write}_register()
 329                 which sits between core GDB and
 330                 the register cache.
 331
 332         REGISTER_CONVERTIBLE
 333         REGISTER_CONVERT_TO_RAW
 334         REGISTER_CONVERT_TO_VIRTUAL
 335                 I think these three are redundant.
 336                 gdbarch_register_{read,write} can
 337                 do any conversion it likes.
 338
 339         REGISTER_VIRTUAL_SIZE
 340         MAX_REGISTER_VIRTUAL_SIZE
 341         REGISTER_VIRTUAL_TYPE
 342                 I think these can be replaced by
 343                 the pair:
 344                   FRAME_REGISTER_TYPE(frame, regnum)
 345                   REGISTER_TYPE(regnum)
 346
 347         DO_REGISTERS_INFO
 348                 Replace with
 349                  FRAME_REGISTER_INFO (frame, ...)
 350
 351         REGISTER_SIM_REGNO()
 352                 If nothing else rename this so that
 353                 how it relates to rawreg and the
 354                 regnum is clear.
 355
 356         REGISTER_BYTES
 357                 The size of the cache can be computed
 358                 on the fly.
 359
 360 --
 361
 362 Restructure gdb directory tree so that it avoids any 8.3 and 14
 363 filename problems.
 364
 365 --
 366
 367 Convert GDB build process to AUTOMAKE.
 368
 369 See also sub-directory configure below.
 370
 371 The current convention is (kind of) to use $(<header>_h) in all
 372 dependency lists.  It isn't done in a consistent way.
 373
 374 --
 375
 376                 GDB 5.2 - Known Problems
 377                 ========================
 378
 379 --
 380
 381                 Code Cleanups: General
 382                 ======================
 383
 384 The following are more general cleanups and fixes.  They are not tied
 385 to any specific release.
 386
 387 --
 388
 389 Rename read_register{,_pid}() to read_unsigned_register{,_pid}().
 390
 391 --
 392
 393 Can't build IRIX -> arm GDB.
 394 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00356.html
 395
 396 David Whedon writes:
 397 > Now I'm building for an embedded arm target.  If there is a way of turning
 398 > remote-rdi off, I couldn't find it. It looks like it gets built by default
 399 > in gdb/configure.tgt(line 58) Anyway, the build dies in
 400 > gdb/rdi-share/unixcomm.c.  SERPORT1 et. al. never get defined because we
 401 > aren't one of the architectures supported.
 402
 403 --
 404
 405 Problem with weak functions
 406 http://sourceware.cygnus.com/ml/gdb/2000-05/msg00060.html
 407
 408 Dan Nicolaescu writes:
 409 > It seems that gdb-4.95.1  does not display correctly the function when
 410 > stoping in weak functions.
 411 >
 412 > It stops in a function that is defined as weak, not in the function
 413 > that is actually run...
 414
 415 --
 416
 417 Follow through `make check' with --enable-shared.
 418
 419 When the srcware tree is configured with --enable-shared, the `expect'
 420 program won't run properly.  Jim Wilson found out gdb has a local hack
 421 to set LD_LIBRARY_PATH, but, AFAIK, no other project has been hacked
 422 similarly.
 423
 424 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00845.html
 425
 426 --
 427
 428 Delete macro TARGET_BYTE_ORDER_SELECTABLE.
 429
 430 Patches in the database.
 431
 432 --
 433
 434 printcmd.c (print_address_numeric):
 435
 436 NOTE: This assumes that the significant address information is kept in
 437 the least significant bits of ADDR - the upper bits were either zero
 438 or sign extended.  Should ADDRESS_TO_POINTER() or some
 439 ADDRESS_TO_PRINTABLE() be used to do the conversion?
 440
 441 --
 442
 443 The BFD directory requires bug-fixed AUTOMAKE et.al.
 444
 445 AUTOMAKE 1.4 incorrectly set the TEXINPUTS environment variable.  It
 446 contained the full path to texinfo.tex when it should have only
 447 contained the directory.  The bug has been fixed in the current
 448 AUTOMAKE sources.  Automake snapshots can be found in:
 449         ftp://sourceware.cygnus.com/pub/gdb/snapshots
 450 and     ftp://sourceware.cygnus.com/pub/binutils
 451
 452 --
 453
 454 Find something better than DEFAULT_BFD_ARCH, DEFAULT_BFD_VEC to
 455 determine the default isa/byte-order.
 456
 457 --
 458
 459 Rely on BFD_BIG_ENDIAN and BFD_LITTLE_ENDIAN instead of host dependent
 460 BIG_ENDIAN and LITTLE_ENDIAN.
 461
 462 --
 463
 464 Eliminate more compiler warnings.
 465
 466 Of course there also needs to be the usual debate over which warnings
 467 are valid and how to best go about this.
 468
 469 One method: choose a single option; get agreement that it is
 470 reasonable; try it out to see if there isn't anything silly about it
 471 (-Wunused-parameters is an example of that) then incrementally hack
 472 away.
 473
 474 The other method is to enable all warnings and eliminate them from one
 475 file at a time.
 476
 477 --
 478
 479 Elimination of ``(catch_errors_ftype *) func''.
 480
 481 Like make_cleanup_func it isn't portable.
 482 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00791.html
 483 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00814.html
 484
 485 --
 486
 487 Nuke #define CONST_PTR.
 488
 489 --
 490
 491 Nuke USG define.
 492
 493 --
 494
 495 [PATCH/5] src/intl/Makefile.in:distclean additions
 496 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00363.html
 497
 498 Do not forget to merge the patch back into the trunk.
 499
 500 --
 501
 502 Rationalize the host-endian code (grep for HOST_BYTE_ORDER).
 503
 504 At present defs.h includes <endian.h> (which is linux specific) yet
 505 almost nothing depends on it.  Suggest "gdb_endian.h" which can also
 506 handle <machine/endian.h> and only include that where it is really
 507 needed.
 508
 509 --
 510
 511 Replace savestring() with something from libiberty.
 512
 513 An xstrldup()? but that would have different semantics.
 514
 515 --
 516
 517 Rationalize use of floatformat_unknown in GDB sources.
 518
 519 Instead of defaulting to floatformat_unknown, should hosts/targets
 520 specify the value explicitly?
 521
 522 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
 523
 524 --
 525
 526 Add a ``name'' member to include/floatformat.h:struct floatformat.
 527 Print that name in gdbarch.c.
 528
 529 --
 530
 531 Sort out the harris mess in include/floatformat.h (it hardwires two
 532 different floating point formats).
 533
 534 --
 535
 536 See of the GDB local floatformat_do_doublest() and libiberty's
 537 floatformat_to_double (which was once GDB's ...) can be merged some
 538 how.
 539
 540 --
 541
 542 Eliminate mmalloc(), mstrsave() et.al. from GDB.
 543
 544 Also eliminate it from defs.h.
 545
 546 --
 547
 548 Eliminate PTR.  ISO-C allows ``void *''.
 549
 550 --
 551
 552 Eliminate abort ().
 553
 554 GDB should never abort.  GDB should either throw ``error ()'' or
 555 ``internal_error ()''.  Better still GDB should naturally unwind with
 556 an error status.
 557
 558 --
 559
 560 GDB probably doesn't build on FreeBSD pre 2.2.x
 561 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00378.html
 562
 563 Fixes to get FreeBSD working on 2.2.x, 3.x and 4.x caused the code to
 564 suffer bit rot.
 565
 566 --
 567
 568 Deprecate "fg".  Apparently ``fg'' is actually continue.
 569
 570 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00417.html
 571
 572 --
 573
 574 Deprecate current use of ``floatformat_unknown''.
 575
 576 Require all targets to explicitly provide their float format instead
 577 of defaulting to floatformat unknown.  Doing the latter leads to nasty
 578 bugs.
 579
 580 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00447.html
 581
 582 --
 583
 584 Rationalize floatformat_to_double() vs floatformat_to_doublest().
 585
 586 Looks like GDB migrated floatformat_to_double() to libiberty but then
 587 turned around and created a ..._to_doublest() the latter containing
 588 several bug fixes.
 589
 590 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00472.html
 591
 592 --
 593
 594 Move floatformat_ia64_ext to libiberty/include floatformat.[ch].
 595
 596 http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00466.html
 597
 598 --
 599
 600 The ``maintenance deprecate set endian big'' command doesn't notice
 601 that it is deprecating ``set endian'' and not ``set endian big'' (big
 602 is implemented using an enum).  Is anyone going to notice this?
 603
 604 --
 605
 606 When tab expanding something like ``set arch<tab>'' ignore the
 607 deprecated ``set archdebug'' and expand to ``set architecture''.
 608
 609 --
 610
 611 Eliminate ``arm_register_names[j] = (char *) regnames[j]'' and the
 612 like from arm-tdep.c.
 613
 614 --
 615
 616 Fix uses of ->function.cfunc = set_function().
 617
 618 The command.c code calls sfunc() when a set command.  Rather than
 619 change it suggest fixing the callback function so that it is more
 620 useful.  See:
 621
 622 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
 623
 624 See also ``Fix implementation of  ``target xxx''.'' below.
 625
 626 --
 627
 628 IRIX 3.x support is probably broken.
 629
 630 --
 631
 632 Delete sim/SIM_HAVE_BREAKPOINTS and gdb/SIM_HAS_BREAKPOINTS.
 633 http://sourceware.cygnus.com/ml/gdb-patches/2000-07/msg00042.html
 634
 635 Apart from the d30v, are there any sim/common simulators that make use
 636 of this?
 637
 638 A brief summary of what happened is that sim/common/sim-break.c was
 639 created as a good idea.  It turned out a better idea was to use
 640 SIM_SIGBREAK and have GDB pass back sim_resume (..., SIGBREAK).
 641
 642 --
 643
 644 Move remote_remove_hw_breakpoint, remote_insert_hw_breakpoint,
 645 remote_remove_watchpoint, remote_insert_watchpoint into target vector.
 646
 647 --
 648
 649 Eliminate ``extern'' from C files.
 650
 651 --
 652
 653 Replace ``STREQ()'' et.al. with ``strcmp() == 0'' et.al.
 654
 655 Extreme care is recommeded - perhaps only modify tests that are
 656 exercised by the testsuite (as determined using some type of code
 657 coverage analysis).
 658
 659 --
 660
 661 Replace the file gdb/CONTRIBUTE with a file that is generated from the
 662 gdb/doc/*.texinfo directory.
 663
 664 --
 665
 666                         New Features and Fixes
 667                         ======================
 668
 669 These are harder than cleanups but easier than work involving
 670 fundamental architectural change.
 671
 672 --
 673
 674 Hardware watchpoint problems on x86 OSes, including Linux:
 675
 676 1. Delete/disable hardware watchpoints should free hardware debug
 677 registers.
 678 2. Watch for different values on a viariable with one hardware debug
 679 register.
 680
 681 According to Eli Zaretskii <eliz@delorie.com>:
 682
 683 These are not GDB/ia32 issues per se: the above features are all
 684 implemented in the DJGPP port of GDB and work in v5.0.  Every
 685 x86-based target should be able to lift the relevant parts of
 686 go32-nat.c and use them almost verbatim.  You get debug register
 687 sharing through reference counts, and the ability to watch large
 688 regions (up to 16 bytes) using multiple registers.  (The required
 689 infrastructure in high-level GDB application code, mostly in
 690 breakpoint.c, is also working since v5.0.)
 691
 692 --
 693
 694 Add built-by, build-date, tm, xm, nm and anything else into gdb binary
 695 so that you can see how the GDB was created.
 696
 697 --
 698
 699 Add an "info bfd" command that displays supported object formats,
 700 similarly to objdump -i.
 701
 702 Is there a command already?
 703
 704 --
 705
 706 Fix ``I'm sorry, Dave, I can't do that.'' from symfile.c.
 707
 708 This requires internationalization.
 709
 710 --
 711
 712 Add support for:
 713
 714 (gdb) p fwprintf(stdout,L"%S\n", f)
 715 No symbol "L" in current context.
 716
 717 --
 718
 719 Cleanup configury support for optional sub-directories.
 720
 721 Check how GCC handles multiple front ends for an example of how things
 722 could work.  A tentative first step is to rationalize things so that
 723 all sub directories are handled in a fashion similar to gdb/mi.
 724
 725 See also automake above.
 726
 727 --
 728
 729 Add a transcript mechanism to GDB.
 730
 731 Such a mechanism might log all gdb input and output to a file in a
 732 form that would allow it to be replayed.  It could involve ``gdb
 733 --transcript=FILE'' or it could involve ``(gdb) transcript file''.
 734
 735 --
 736
 737 Can the xdep files be replaced by autoconf?
 738
 739 --
 740
 741 Document trace machinery
 742
 743 --
 744
 745 Document ui-out and ui-file.
 746
 747 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00121.html
 748
 749 --
 750
 751 Update texinfo.tex to latest?
 752
 753 --
 754
 755 Incorporate agentexpr.texi into gdb.texinfo
 756
 757 agentexpr.texi mostly describes the details of the byte code used for
 758 tracepoints, not the internals of the support for this in GDB.  So it
 759 looks like gdb.texinfo is a better place for this information.
 760
 761 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00566.html
 762
 763 --
 764
 765 Document overlay machinery.
 766
 767 --
 768
 769 ``(gdb) catch signal SIGNAL''
 770
 771 Overlaps with ``handle SIGNAL'' but the implied behavior is different.
 772 You can attach commands to a catch but not a handle.  A handle has a
 773 limited number of hardwired actions.
 774
 775 --
 776
 777 Fix TUI
 778
 779      o  readline/*.h bitrot
 780
 781         The TUI isn't up-to-date with
 782         respect to the readline currently
 783         bundled with GDB.  Importing a
 784         new readline is on the 5.1 wish
 785         list so this can only get worse.
 786
 787         Grep for things like term_cursor_move.
 788
 789         (To be honest, I don't see anyone
 790         importing a new readline before 5.1 is
 791         out)
 792
 793      o  tui.c:va_catch_errors() bitrot
 794
 795         This nasty piece of work used knowledge
 796         of the internals of GDBs error functions :-(
 797         Ever since those internals were cleaned
 798         up this code has been broken. :-(
 799
 800      o  tuiWin.c:c_makeVisibleWithNewHeight() broken
 801         tuiLayout.c:_extractDisplayStartAddr() broken
 802
 803         Both these function call find_line_pc()
 804         incorrectly (wrong args, wrong return value).
 805
 806         I suspect this bug has always been there!
 807         It had been hidden because those files
 808         didn't include the necessary header files
 809         from gdb proper :-(
 810
 811      o  tuiRegs() host dependant
 812
 813         Not suprisingly, this isn't a very portable
 814         section of code.  However, I'm sure people
 815         could live with no regs in the short to
 816         medium term.
 817
 818      o  defs.h: #include "tui.h" et.al.
 819
 820         I'm not sure where this came from.
 821         It was a really bad idea.
 822
 823         To get things to compile I did a nasty
 824         hack (Just declare what was needed and
 825         replace any expressions like xx->y.z()
 826         in GDB proper with function calls).  I
 827         could commit it slightly cleaned up if
 828         you like.
 829
 830         Medium Term. the #ifdef TUI and TuiDo()
 831         should be changed to hooks (like GDBTK).
 832         The gdb-events.[hc] is there for that
 833         purpose (1)
 834
 835      o  tui.c:_tuiReset() host dependant
 836
 837         tui.c contains a lump of termio[s]
 838         I suspect an equivalent block of
 839         code can be lifted from readline.
 840         An equivalent readline function may
 841         even be available.
 842
 843      o  curses.h vs ncurses.h.
 844
 845         Simple portability problem.
 846
 847      o  subsetCompare()
 848
 849         This function is a mystery - where is it?
 850
 851      o  tui-file.[hc] cleanup
 852
 853         This can be significantly simplified.
 854
 855      o  The code should be pacified. (-Werror -W...)
 856
 857         There are plenty of #includes,
 858         duplicate #includes, missing function decls
 859         and the like.
 860
 861         Some of the problems I found were through
 862         fixing a few of the warnings.
 863
 864      o  The code should be GNUtified.
 865
 866         It would be very nice to have this code
 867         look like the rest of GDB.  That way people
 868         would be more accepting of it as a true
 869         gdb component.
 870
 871         Until it is GNUtified it is going to stick
 872         out like a sore thumb to the programmer.
 873
 874      o  The code should be clearly copyrighted
 875
 876         (FSF, with due credit to HP)
 877
 878 --
 879
 880 Add support for ``gdb --- PROGRAM ARGS ...''.
 881 Add support for ``gdb -cmd=...''
 882
 883 Along with many variations.  Check:
 884
 885 ????? for a full discussion.
 886
 887 for a discussion.
 888
 889 --
 890
 891 Implement ``(gdb) !ls''.
 892
 893 Which is very different from ``(gdb) ! ls''.  Implementing the latter
 894 is trivial.
 895
 896 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00034.html
 897
 898 --
 899
 900 Change the (char *list[]) to (const char (*)[]) so that dynamic lists can
 901 be passed.
 902
 903 --
 904
 905 When tab expanding something like ``set arch<tab>'' ignore the
 906 deprecated ``set archdebug'' and expand to ``set architecture''.
 907
 908 --
 909
 910 Replace the code that uses the host FPU with an emulator of the target
 911 FPU.
 912
 913 --
 914
 915 The "ocd reset" command needs to flush the dcache, which requires breaking
 916 the abstraction layer between the target independent and target code.  One
 917 way to address this is provide a generic "reset" command and target vector.
 918
 919 http://sources.redhat.com/ml/gdb-patches/2000-10/msg00011.html
 920
 921 --
 922
 923                         Thread Support
 924                         ==============
 925
 926 --
 927
 928 Generic: lin-thread cannot handle thread exit (Mark Kettenis, Michael
 929 Snyder) http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00525.html
 930
 931 The thread_db assisted debugging code doesn't handle exiting threads
 932 properly, at least in combination with glibc 2.1.3 (the framework is
 933 there, just not the actual code).  There are at least two problems
 934 that prevent this from working.
 935
 936 As an additional reference point, the pre thread_db code did not work
 937 either.
 938
 939 --
 940
 941 GNU/Linux/x86 and random thread signals (and Solaris/SPARC but not
 942 Solaris/x86).
 943 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00336.html
 944
 945 Christopher Blizzard writes:
 946
 947 So, I've done some more digging into this and it looks like Jim
 948 Kingdon has reported this problem in the past:
 949
 950 http://sourceware.cygnus.com/ml/bug-gdb/1999-10/msg00058.html
 951
 952 I can reproduce this problem both with and without Tom's patch.  Has
 953 anyone seen this before?  Maybe have a solution for it hanging around?
 954 :)
 955
 956 There's a test case for this documented at:
 957
 958 when debugging threaded applications you get extra SIGTRAPs
 959 http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=9565
 960
 961 [There should be a GDB testcase - cagney]
 962
 963 --
 964
 965 GDB5 TOT on unixware 7
 966 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00119.html
 967
 968 Robert Lipe writes:
 969 > I just spun the top of tree of the GDB5 branch on UnixWare 7.  As a
 970 > practical matter, the current thread support is somewhat more annoying
 971 > than when GDB was thread-unaware.
 972
 973 --
 974
 975                         Language Support
 976                         ================
 977
 978 New languages come onto the scene all the time.
 979
 980 --
 981
 982 Re: Various C++ things
 983
 984 value_headof/value_from_vtable_info are worthless, and should be
 985 removed.  The one place in printcmd.c that uses it should use the RTTI
 986 functions.
 987
 988 RTTI for g++ should be using the typeinfo functions rather than the
 989 vtables.  The typeinfo functions are always at offset 4 from the
 990 beginning of the vtable, and are always right. The vtables will have
 991 weird names like E::VB sometimes.  The typeinfo function will always
 992 be "E type_info function", or somesuch.
 993
 994 value_virtual_fn_field needs to be fixed so there are no failures for
 995 virtual functions for C++ using g++.
 996
 997 Testsuite cases are the major priority right now for C++ support,
 998 since i have to make a lot of changes that could potentially break
 999 each other.
1000
1001 --
1002
1003 Add support for Modula3
1004
1005 Get DEC/Compaq to contribute their Modula-3 support.
1006
1007 --
1008
1009                 Remote Protocol Support
1010                 =======================
1011
1012 --
1013
1014 Revised UDP support (was: Re: [Fwd: [patch] UDP transport support])
1015 http://sourceware.cygnus.com/ml/gdb-patches/2000-04/msg00000.html
1016
1017 (Broken) support for GDB's remote protocol across UDP is to be
1018 included in the follow-on release.
1019
1020 It should be noted that UDP can only work when the [Gg] packet fits in
1021 a single UDP packet.
1022
1023 There is also much debate over the merit of this.
1024
1025 --
1026
1027 Migrate qfThreadInfo packet -> qThreadInfo. (Andrew Cagney)
1028
1029 Add support for packet enable/disable commands with these thread
1030 packets.  General cleanup.
1031
1032 [PATCH] Document the ThreadInfo remote protocol queries
1033 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00832.html
1034
1035 [PATCH] "info threads" queries for remote.c
1036 http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00831.html
1037
1038 --
1039
1040 Remote protocol doco feedback.
1041
1042 Too much feedback to mention needs to be merged in (901660).  Search
1043 for the word ``remote''.
1044
1045
1046 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00023.html
1047 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00056.html
1048 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00382.html
1049
1050 --
1051
1052 GDB doesn't recover gracefully from remote protocol errors.
1053
1054 GDB wasn't checking for NAKs from the remote target.  Instead a NAK is
1055 ignored and a timeout is required before GDB retries.  A pre-cursor to
1056 fixing this this is making GDB's remote protocol packet more robust.
1057
1058 While downloading to a remote protocol target, gdb ignores packet
1059 errors in so far as it will continue to download with chunk N+1 even
1060 if chunk N was not correctly sent.  This causes gdb.base/remote.exp to
1061 take a painfully long time to run.  As a PS that test needs to be
1062 fixed so that it builds on 16 bit machines.
1063
1064 --
1065
1066 Fix the ``!'' packet.
1067
1068 JT reported that the existing targets do, in fact return ``OK'' so it
1069 is possible to merge remote and extended-remote targets.
1070
1071 --
1072
1073 Drop ``<address>'' from the [SsCc] packets.
1074
1075 I don't think that GDB generates them so having it in the protocol is
1076 silly.
1077
1078 --
1079
1080 Fix doco on the ``q'' packet.
1081
1082 It has evolved into a generic RPC.  The notes should reflect this and,
1083 perhaps, the ``Q'' packet can be deprecated.
1084
1085 The doco should mention that ``OK'' is a valid packet response.
1086
1087 The doco should explain why ``OK'' needs to be a valid packet
1088 response.
1089
1090 --
1091
1092 Add the cycle step command.
1093
1094 http://sourceware.cygnus.com/ml/gdb/2000-q1/msg00237.html
1095
1096 --
1097
1098 Resolve how to scale things to support very large packets.
1099
1100 --
1101
1102 Resolve how to handle a target that changes things like its endianess
1103 on the fly - should it be returned in the ``T'' packet?
1104
1105 Underlying problem is that the register file is target endian.  If the
1106 target endianess changes gdb doesn't know.
1107
1108 --
1109
1110                         Symbol Support
1111                         ==============
1112
1113 If / when GDB starts to support the debugging of multi-processor
1114 (rather than multi-thread) applications the symtab code will need to
1115 be updated a little so that several independent symbol tables are
1116 active at a given time.
1117
1118 The other interesting change is a clarification of the exact meaning
1119 of CORE_ADDR and that has had consequences for a few targets (that
1120 were abusing that data type).
1121
1122 --
1123
1124 Investiagate ways of reducing memory.
1125
1126 --
1127
1128 Investigate ways of improving load time.
1129
1130 --
1131
1132 Get the d10v to use POINTER_TO_ADDRESS and ADDRESS_TO_POINTER.
1133
1134 Consequence of recent symtab clarification.  No marks for figuring out
1135 who maintains the d10v.
1136
1137 --
1138
1139 Get the MIPS to correctly sign extend all address <-> pointer
1140 conversions.
1141
1142 Consequence of recent symtab clarification.  No marks for figuring out
1143 who maintains the MIPS.
1144
1145 --
1146
1147 GDB truncates 64 bit enums.
1148
1149 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00290.html
1150
1151 --
1152
1153                         Testsuite Support
1154                         =================
1155
1156 There are never to many testcases.
1157
1158 --
1159
1160 Better thread testsuite.
1161
1162 --
1163
1164 Better C++ testsuite.
1165
1166 --
1167
1168 Look at adding a GDB specific testsuite directory so that white box
1169 tests of key internals can be added (eg ui_file).
1170
1171 --
1172
1173 Separate out tests that involve the floating point (FP).
1174
1175 (Something for people brining up new targets).  FP and non-fp tests
1176 are combined.  I think there should be set of basic tests that
1177 exercise pure integer support and then a more expanded set that
1178 exercise FP and FP/integer interactions.
1179
1180 As an example, the MIPS, for n32 as problems with passing FP's and
1181 structs.  Since most inferior call tests include FP it is difficult to
1182 determine of the integer tests are ok.
1183
1184 --
1185
1186                 Architectural Changes: General
1187                 ==============================
1188
1189 These are harder than simple cleanups / fixes and, consequently
1190 involve more work.  Typically an Architectural Change will be broken
1191 down into a more digestible set of cleanups and fixes.
1192
1193 --
1194
1195 Cleanup software single step.
1196
1197 At present many targets implement software single step by directly
1198 blatting memory (see rs6000-tdep.c).  Those targets should register
1199 the applicable breakpoints using the breakpoint framework.  Perhaphs a
1200 new internal breakpoint class ``step'' is needed.
1201
1202 --
1203
1204 Replace READ_FP() with FRAME_HANDLE().
1205
1206 READ_FP() is a hangover from the days of the vax when the ABI really
1207 did have a frame pointer register.  Modern architectures typically
1208 construct a virtual frame-handle from the stack pointer and various
1209 other bits of string.
1210
1211 Unfortunately GDB still treats this synthetic FP register as though it
1212 is real.  That in turn really confuses users (arm and ``print $fp'' VS
1213 ``info registers fp'').  The synthetic FP should be separated out of
1214 the true register set presented to the user.
1215
1216 --
1217
1218 Register Cache Cleanup (below from Andrew Cagney)
1219
1220 I would depict the current register architecture as something like:
1221
1222         High GDB  -->   Low GDB
1223             |              |
1224            \|/            \|/
1225             --- REG NR -----
1226                    |
1227     register + REGISTER_BYTE(reg_nr)
1228                    |
1229                   \|/
1230         -------------------------
1231         |  extern register[]    |
1232         -------------------------
1233
1234 where neither the high (valops.c et.al.) or low gdb (*-tdep.c) are
1235 really clear on what mechanisms they should be using to manipulate that
1236 buffer.   Further, much code assumes, dangerously, that registers are
1237 contigious.  Having got mips-tdep.c to support multiple ABIs, believe
1238 me, that is a bad assumption.  Finally, that register cache layout is
1239 determined by the current remote/local target and _not_ the less
1240 specific target ISA.  In fact, in many cases it is determined by the
1241 somewhat arbitrary layout of the [gG] packets!
1242
1243
1244 How I would like the register file to work is more like:
1245
1246
1247           High GDB
1248              |
1249             \|/
1250        pseudo reg-nr
1251              |
1252         map pseudo <->
1253         random cache
1254         bytes
1255              |
1256             \|/
1257         ------------
1258         | register |
1259         |  cache   |
1260         ------------
1261             /|\
1262              |
1263         map random cache
1264         bytes to target
1265         dependent i-face
1266             /|\
1267              |
1268        target dependent
1269         such as [gG] packet
1270         or ptrace buffer
1271
1272 The main objectives being:
1273
1274         o       a clear separation between the low
1275                 level target and the high level GDB
1276
1277         o       a mechanism that solves the general
1278                 problem of register aliases, overlaps
1279                 etc instead of treating them as optional
1280                 extras that can be wedged in as an after
1281                 thought (that is a reasonable description
1282                 of the current code).
1283
1284                 Identify then solve the hard case and the
1285                 rest just falls out.  GDB solved the easy
1286                 case and then tried to ignore the real
1287                 world :-)
1288
1289         o       a removal of the assumption that the
1290                 mapping between the register cache
1291                 and virtual registers is largely static.
1292                 If you flip the USR/SSR stack register
1293                 select bit in the status-register then
1294                 the corresponding stack registers should
1295                 reflect the change.
1296
1297         o       a mechanism that clearly separates the
1298                 gdb internal register cache from any
1299                 target (not architecture) dependent
1300                 specifics such as [gG] packets.
1301
1302 Of course, like anything, it sounds good in theory.  In reality, it
1303 would have to contend with many<->many relationships at both the
1304 virt<->cache and cache<->target level.  For instance:
1305
1306         virt<->cache
1307                 Modifying an mmx register may involve
1308                 scattering values across both FP and
1309                 mmpx specific parts of a buffer
1310
1311         cache<->target
1312                 When writing back a SP it may need to
1313                 both be written to both SP and USP.
1314
1315
1316 Hmm,
1317
1318 Rather than let this like the last time it was discussed, just slip, I'm
1319 first going to add this e-mail (+ references) to TODO.  I'd then like to
1320 sketch out a broad strategy I think could get us there.
1321
1322
1323 First thing I'd suggest is separating out the ``extern registers[]''
1324 code so that we can at least identify what is using it.  At present
1325 things are scattered across many files.  That way we can at least
1326 pretend that there is a cache instead of a global array :-)
1327
1328 I'd then suggest someone putting up a proposal for the pseudo-reg /
1329 high-level side interface so that code can be adopted to it.  For old
1330 code, initially a blanket rename of write_register_bytes() to
1331 deprecated_write_register_bytes() would help.
1332
1333 Following that would, finaly be the corresponding changes to the target.
1334
1335 --
1336
1337 Check that GDB can handle all BFD architectures (Andrew Cagney)
1338
1339 There should be a test that checks that BFD/GDB are in sync with
1340 regard to architecture changes.  Something like a test that first
1341 queries GDB for all supported architectures and then feeds each back
1342 to GDB..  Anyone interested in learning how to write tests?  :-)
1343
1344 --
1345
1346                 Architectural Change: Multi-arch et al.
1347                 =======================================
1348
1349 The long term objective is to remove all assumptions that there is a
1350 single target with a single address space with a single instruction
1351 set architecture and single application binary interface.
1352
1353 This is an ongoing effort.  The first milestone is to enable
1354 ``multi-arch'' where by all architectural decisions are made at
1355 runtime.
1356
1357 It should be noted that ``gdbarch'' is really ``gdbabi'' and
1358 ``gdbisa''.  Once things are multi-arched breaking that down correctly
1359 will become much easier.
1360
1361 --
1362
1363 GDBARCH cleanup (Andrew Cagney)
1364
1365 The non-generated parts of gdbarch.{sh,h,c} should be separated out
1366 into arch-utils.[hc].
1367
1368 Document that gdbarch_init_ftype could easily fail because it didn't
1369 identify an architecture.
1370
1371 --
1372
1373 Fix BELIEVE_PPC_PROMOTION.  Change it to BELIEVE_PPC_PROMOTION_P?
1374
1375 At present there is still #ifdef BELIEVE_PPC_PROMOTION code in the
1376 symtab file.
1377
1378 --
1379
1380 Fix target_signal_from_host() etc.
1381
1382 The name is wrong for starters.  ``target_signal'' should probably be
1383 ``gdb_signal''.  ``from_host'' should be ``from_target_signal''.
1384 After that it needs to be multi-arched and made independent of any
1385 host signal numbering.
1386
1387 --
1388
1389 Update ALPHA so that it uses ``struct frame_extra_info'' instead of
1390 EXTRA_FRAME_INFO.
1391
1392 This is a barrier to replacing mips_extra_func_info with something
1393 that works with multi-arch.
1394
1395 --
1396
1397 Multi-arch  mips_extra_func_info.
1398
1399 This first needs the alpha to be updated so that it uses ``struct
1400 frame_extra_info''.
1401
1402 --
1403
1404 Rationalize TARGET_SINGLE_FORMAT and TARGET_SINGLE_BIT et al.
1405
1406 Surely one of them is redundant.
1407
1408 --
1409
1410 Convert ALL architectures to MULTI-ARCH.
1411
1412 --
1413
1414 Select the initial multi-arch ISA / ABI based on --target or similar.
1415
1416 At present the default is based on what ever is first in the BFD
1417 archures table.  It should be determined based on the ``--target=...''
1418 name.
1419
1420 --
1421
1422 Make MIPS pure multi-arch.
1423
1424 It is only at the multi-arch enabled stage.
1425
1426 --
1427
1428 Truly multi-arch.
1429
1430 Enable the code to recognize --enable-targets=.... like BINUTILS does.
1431
1432 Can the tm.h and nm.h files be eliminated by multi-arch.
1433
1434 --
1435
1436         Architectural Change: MI, LIBGDB and scripting languages
1437         ========================================================
1438
1439 See also architectural changes related to the event loop.  LIBGDB
1440 can't be finished until there is a generic event loop being used by
1441 all targets.
1442
1443 The long term objective is it to be possible to integrate GDB into
1444 scripting languages.
1445
1446 --
1447
1448 Implement generic ``(gdb) commmand > file''
1449
1450 Once everything is going through ui_file it should be come fairly
1451 easy.
1452
1453 http://sourceware.cygnus.com/ml/gdb/2000-04/msg00104.html
1454
1455 --
1456
1457 Replace gdb_stdtarg with gdb_targout (and possibly gdb_targerr).
1458
1459 gdb_stdtarg is easily confused with gdb_stdarg.
1460
1461 --
1462
1463 Extra ui_file methods - dump.
1464
1465 Very useful for whitebox testing.
1466
1467 --
1468
1469 Eliminate error_begin().
1470
1471 With ui_file, there is no need for the statefull error_begin ()
1472 function.
1473
1474 --
1475
1476 Send normal output to gdb_stdout.
1477 Send error messages to gdb_stderror.
1478 Send debug and log output log gdb_stdlog.
1479
1480 GDB still contains many cases where (f)printf or printf_filtered () is
1481 used when it should be sending the messages to gdb_stderror or
1482 gdb_stdlog.  The thought of #defining printf to something has crossed
1483 peoples minds ;-)
1484
1485 --
1486
1487 Re-do GDB's output pager.
1488
1489 GDB's output pager still relies on people correctly using *_filtered
1490 for gdb_stdout and *_unfiltered for gdb_stdlog / gdb_stderr.
1491 Hopefully, with all normal output going to gdb_stdout, the pager can
1492 just look at the ui_file that the output is on and then use that to
1493 decide what to do about paging.  Sounds good in theory.
1494
1495 --
1496
1497 Check/cleanup MI documentation.
1498
1499 The list of commands specified in the documentation needs to be
1500 checked against the mi-cmds.c table in a mechanical way (so that they
1501 two can be kept up-to-date).
1502
1503 --
1504
1505 Convert MI into libgdb
1506
1507 MI provides a text interface into what should be many of the libgdb
1508 functions.  The implementation of those functions should be separated
1509 into the MI interface and the functions proper.  Those functions being
1510 moved to gdb/lib say.
1511
1512 --
1513
1514 Create libgdb.h
1515
1516 The first part can already be found in defs.h.
1517
1518 --
1519
1520 MI's input does not use buffering.
1521
1522 At present the MI interface reads raw characters of from an unbuffered
1523 FD.  This is to avoid several nasty buffer/race conditions.  That code
1524 should be changed so that it registers its self with the event loop
1525 (on the input FD) and then push commands up to MI as they arrive.
1526
1527 The serial code already does this.
1528
1529 --
1530
1531 Make MI interface accessible from existing CLI.
1532
1533 --
1534
1535 Add a breakpoint-edit command to MI.
1536
1537 It would be similar to MI's breakpoint create but would apply to an
1538 existing breakpoint.  It saves the need to delete/create breakpoints
1539 when ever they are changed.
1540
1541 --
1542
1543 Add directory path to MI breakpoint.
1544
1545 That way the GUI's task of finding the file within which the
1546 breakpoint was set is simplified.
1547
1548 --
1549
1550 Add a mechanism to reject certain expression classes to MI
1551
1552 There are situtations where you don't want GDB's expression
1553 parser/evaluator to perform inferior function calls or variable
1554 assignments.  A way of restricting the expression parser so that such
1555 operations are not accepted would be very helpful.
1556
1557 --
1558
1559 Remove sideffects from libgdb breakpoint create function.
1560
1561 The user can use the CLI to create a breakpoint with partial
1562 information - no file (gdb would use the file from the last
1563 breakpoint).
1564
1565 The libgdb interface currently affects that environment which can lead
1566 to confusion when a user is setting breakpoints via both the MI and
1567 the CLI.
1568
1569 This is also a good example of how getting the CLI ``right'' will be
1570 hard.
1571
1572 --
1573
1574 Move gdb_lasterr to ui_out?
1575
1576 The way GDB throws errors and records them needs a re-think.  ui_out
1577 handles the correct output well.  It doesn't resolve what to do with
1578 output / error-messages when things go wrong.
1579
1580 --
1581
1582 do_setshow_command contains a 1024 byte buffer.
1583
1584 The function assumes that there will never be any more than 1024 bytes
1585 of enum.  It should use mem_file.
1586
1587 --
1588
1589 Should struct cmd_list_element . completer take the command as an
1590 argument?
1591
1592 --
1593
1594 Should the bulk of top.c:line_completion_function() be moved to
1595 command.[hc]?  complete_on_cmdlist() and complete_on_enums() could
1596 then be made private.
1597
1598 --
1599
1600 top.c (execute_command): Should a command being valid when the target
1601 is running be made an attribute (predicate) to the command rather than
1602 an explicit set of tests.
1603
1604 --
1605
1606 top.c (execute_command): Should the bulk of this function be moved
1607 into command.[hc] so that top.c doesn't grub around in the command
1608 internals?
1609
1610 --
1611
1612                 Architectural Change: Async
1613                 ===========================
1614
1615 While GDB uses an event loop when prompting the user for input.  That
1616 event loop is not exploited by targets when they allow the target
1617 program to continue.  Typically targets still block in (target_wait())
1618 until the program again halts.
1619
1620 The closest a target comes to supporting full asynchronous mode are
1621 the remote targets ``async'' and ``extended-async''.
1622
1623 --
1624
1625 Asynchronous expression evaluator
1626
1627 Inferior function calls hang GDB.
1628
1629 --
1630
1631 Fix implementation of  ``target xxx''.
1632
1633 At present when the user specifies ``target xxxx'', the CLI maps that
1634 directly onto a target open method.  It is then assumed that the
1635 target open method should do all sorts of complicated things as this
1636 is the only chance it has.  Check how the various remote targets
1637 duplicate the target operations.  Check also how the various targets
1638 behave differently for purely arbitrary reasons.
1639
1640 What should happen is that ``target xxxx'' should call a generic
1641 ``target'' function and that should then co-ordinate the opening of
1642 ``xxxx''.  This becomes especially important when you're trying to
1643 open an asynchronous target that may need to perform background tasks
1644 as part of the ``attach'' phase.
1645
1646 Unfortunately, due to limitations in the old/creaking command.h
1647 interface, that isn't possible.  The function being called isn't told
1648 of the ``xxx'' or any other context information.
1649
1650 Consequently a precursor to fixing ``target xxxx'' is to clean up the
1651 CLI code so that it passes to the callback function (attatched to a
1652 command) useful information such as the actual command and a context
1653 for that command.  Other changes such as making ``struct command''
1654 opaque may also help.
1655
1656 See also:
1657 http://sourceware.cygnus.com/ml/gdb-patches/2000-06/msg00062.html
1658
1659 --
1660
1661 Make "target xxx" command interruptible.
1662
1663 As things become async this becomes possible.  A target would start
1664 the connect and then return control to the event loop.  A cntrl-c
1665 would notify the target that the operation is to be abandoned and the
1666 target code could respond.
1667
1668 --
1669
1670 Add a "suspend" subcommand of the "continue" command to suspend gdb
1671 while continuing  execution of the subprocess.  Useful when you are
1672 debugging servers and you want to dodge out and initiate a connection
1673 to a server running under gdb.
1674
1675 [hey async!!]
1676
1677 --
1678
1679                         TODO FAQ
1680                         ========
1681
1682 Frequently requested but not approved requests.
1683
1684 --
1685
1686 Eliminate unused argument warnings using ATTRIBUTE_UNUSED.
1687
1688 The benefits on this one are thought to be marginal - GDBs design
1689 means that unused parameters are very common.  GCC 3.0 will also
1690 include the option -Wno-unused-parameter which means that ``-Wall
1691 -Wno-unused-parameters -Werror'' can be specified.
1692
1693 --
1694
1695
1696
1697                         Legacy Wish List
1698                         ================
1699
1700 This list is not up to date, and opinions vary about the importance or
1701 even desirability of some of the items.  If you do fix something, it
1702 always pays to check the below.
1703
1704 --
1705
1706 @c This does not work (yet if ever).  FIXME.
1707 @c @item --parse=@var{lang} @dots{}
1708 @c Configure the @value{GDBN} expression parser to parse the listed languages.
1709 @c @samp{all} configures @value{GDBN} for all supported languages.  To get a
1710 @c list of all supported languages, omit the argument.  Without this
1711 @c option, @value{GDBN} is configured to parse all supported languages.
1712
1713 --
1714
1715 START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that
1716 is its default value.  Clean this up.
1717
1718 --
1719
1720 It should be possible to use symbols from shared libraries before we know
1721 exactly where the libraries will be loaded.  E.g. "b perror" before running
1722 the program.  This could maybe be done as an extension of the "breakpoint
1723 re-evaluation" after new symbols are loaded.
1724
1725 --
1726
1727 Make single_step() insert and remove breakpoints in one operation.
1728
1729 [If this is talking about having single_step() insert the breakpoints,
1730 run the target then pull the breakpoints then it is wrong.  The
1731 function has to return as control has to eventually be passed back to
1732 the main event loop.]
1733
1734 --
1735
1736 Speed up single stepping by avoiding extraneous ptrace calls.
1737
1738 --
1739
1740 Speed up single stepping by not inserting and removing breakpoints
1741 each time the inferior starts and stops.
1742
1743 Breakpoints should not be inserted and deleted all the time.  Only the
1744 one(s) there should be removed when we have to step over one.  Support
1745 breakpoints that don't have to be removed to step over them.
1746
1747 [this has resulted in numerous debates.  The issue isn't clear cut]
1748
1749 --
1750
1751 Provide "voodoo" debugging of core files.  This creates a zombie
1752 process as a child of the debugger, and loads it up with the data,
1753 stack, and regs of the core file.  This allows you to call functions
1754 in the executable, to manipulate the data in the core file.
1755
1756 [you wish]
1757
1758 --
1759
1760 GDB reopens the source file on every line, as you "next" through it.
1761
1762 [still true? I've a memory of this being fixed]
1763
1764 --
1765
1766 Perhaps "i source" should take an argument like that of "list".
1767
1768 --
1769
1770 Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if
1771 it matches the source line indicated.
1772
1773 --
1774
1775 The prompt at end of screen should accept space as well as CR.
1776
1777 --
1778
1779 Backtrace should point out what the currently selected frame is, in
1780 its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar,
1781 ...)" rather than "#3 foo (bar, ...)".
1782
1783 --
1784
1785 "i program" should work for core files, and display more info, like what
1786 actually caused it to die.
1787
1788 --
1789
1790 "x/10i" should shorten the long name, if any, on subsequent lines.
1791
1792 --
1793
1794 "next" over a function that longjumps, never stops until next time you happen
1795 to get to that spot by accident.  E.g. "n" over execute_command which has
1796 an error.
1797
1798 --
1799
1800 "set zeroprint off", don't bother printing members of structs which
1801 are entirely zero.  Useful for those big structs with few useful
1802 members.
1803
1804 --
1805
1806 GDB does four ioctl's for every command, probably switching terminal modes
1807 to/from inferior or for readline or something.
1808
1809 --
1810
1811 terminal_ours versus terminal_inferior: cache state.  Switch should be a noop
1812 if the state is the same, too.
1813
1814 --
1815
1816 "i frame" shows wrong "arglist at" location, doesn't show where the args
1817 should be found, only their actual values.
1818
1819 --
1820
1821 There should be a way for "set" commands to validate the new setting
1822 before it takes effect.
1823
1824 --
1825
1826 "ena d" is ambiguous, why?  "ena delete" seems to think it is a command!
1827
1828 --
1829
1830 i line VAR produces "Line number not known for symbol ``var''.".  I
1831 thought we were stashing that info now!
1832
1833 --
1834
1835 We should be able to write to random files at hex offsets like adb.
1836
1837 --
1838
1839 [elena - delete this]
1840
1841 Handle add_file with separate text, data, and bss addresses.  Maybe
1842 handle separate addresses for each segment in the object file?
1843
1844 --
1845
1846 [Jimb/Elena delete this one]
1847
1848 Handle free_named_symtab to cope with multiply-loaded object files
1849 in a dynamic linking environment.  Should remember the last copy loaded,
1850 but not get too snowed if it finds references to the older copy.
1851
1852 --
1853
1854 [elena delete this also]
1855
1856 Remove all references to:
1857         text_offset
1858         data_offset
1859         text_data_start
1860         text_end
1861         exec_data_offset
1862         ...
1863 now that we have BFD.  All remaining are in machine dependent files.
1864
1865 --
1866
1867 Re-organize help categories into things that tend to fit on a screen
1868 and hang together.
1869
1870 --
1871
1872 Add in commands like ADB's for searching for patterns, etc.  We should
1873 be able to examine and patch raw unsymboled binaries as well in gdb as
1874 we can in adb.  (E.g. increase the timeout in /bin/login without source).
1875
1876 [actually, add ADB interface :-]
1877
1878 --
1879
1880 When doing "step" or "next", if a few lines of source are skipped between
1881 the previous line and the current one, print those lines, not just the
1882 last line of a multiline statement.
1883
1884 --
1885
1886 Handling of "&" address-of operator needs some serious overhaul
1887 for ANSI C and consistency on arrays and functions.
1888   For "float point[15];":
1889 ptype &point[4]   ==> Attempt to take address of non-lvalue.
1890   For "char *malloc();":
1891 ptype malloc      ==> "char *()";  should be same as
1892 ptype &malloc     ==> "char *(*)()"
1893 call printf ("%x\n", malloc)   ==> weird value, should be same as
1894 call printf ("%x\n", &malloc)  ==> correct value
1895
1896 --
1897
1898 Fix dbxread.c symbol reading in the presence of interrupts.  It
1899 currently leaves a cleanup to blow away the entire symbol table when a
1900 QUIT occurs.  (What's wrong with that? -kingdon, 28 Oct 1993).
1901
1902 [I suspect that the grype was that, on a slow system, you might want
1903 to cntrl-c and get just half the symbols and then load the rest later
1904 - scary to be honest]
1905
1906 --
1907
1908 Mipsread.c reads include files depth-first, because the dependencies
1909 in the psymtabs are way too inclusive (it seems to me).  Figure out what
1910 really depends on what, to avoid recursing 20 or 30 times while reading
1911 real symtabs.
1912
1913 --
1914
1915 value_add() should be subtracting the lower bound of arrays, if known,
1916 and possibly checking against the upper bound for error reporting.
1917
1918 --
1919
1920 When listing source lines, check for a preceding \n, to verify that
1921 the file hasn't changed out from under us.
1922
1923 [fixed by some other means I think.  That hack wouldn't actually work
1924 reliably - the file might move such that another \n appears. ]
1925
1926 --
1927
1928 Get all the remote systems (where the protocol allows it) to be able to
1929 stop the remote system when the GDB user types ^C (like remote.c
1930 does).  For ebmon, use ^Ak.
1931
1932 --
1933
1934 Possible feature: A version of the "disassemble" command which shows
1935 both source and assembly code ("set symbol-filename on" is a partial
1936 solution).
1937
1938 [has this been done?  It was certainly done for MI and GDBtk]
1939
1940 --
1941
1942 investigate "x/s 0" (right now stops early) (I think maybe GDB is
1943 using a 0 address for bad purposes internally).
1944
1945 --
1946
1947 Make "info path" and path_command work again (but independent of the
1948 environment either of gdb or that we'll pass to the inferior).
1949
1950 --
1951
1952 Make GDB understand the GCC feature for putting octal constants in
1953 enums.  Make it so overflow on an enum constant does not error_type
1954 the whole type.  Allow arbitrarily large enums with type attributes.
1955 Put all this stuff in the testsuite.
1956
1957 --
1958
1959 Make TYPE_CODE_ERROR with a non-zero TYPE_LENGTH more useful (print
1960 the value in hex; process type attributes).  Add this to the
1961 testsuite.  This way future compilers can add new types and old
1962 versions of GDB can do something halfway reasonable.
1963
1964 --
1965
1966 Fix mdebugread.c:parse_type to do fundamental types right (see
1967 rs6000_builtin_type in stabsread.c for what "right" is--the point is
1968 that the debug format fixes the sizes of these things and it shouldn't
1969 depend on stuff like TARGET_PTR_BIT and so on.  For mdebug, there seem
1970 to be separate bt* codes for 64 bit and 32 bit things, and GDB should
1971 be aware of that).  Also use a switch statement for clarity and speed.
1972
1973 --
1974
1975 Investigate adding symbols in target_load--some targets do, some
1976 don't.
1977
1978 --
1979
1980 Put dirname in psymtabs and change lookup*symtab to use dirname (so
1981 /foo/bar.c works whether compiled by cc /foo/bar.c, or cd /foo; cc
1982 bar.c).
1983
1984 --
1985
1986 Merge xcoffread.c and coffread.c.  Use breakpoint_re_set instead of
1987 fixup_breakpoints.
1988
1989 --
1990
1991 Make a watchpoint which contains a function call an error (it is
1992 broken now, making it work is probably not worth the effort).
1993
1994 --
1995
1996 New test case based on weird.exp but in which type numbers are not
1997 renumbered (thus multiply defining a type).  This currently causes an
1998 infinite loop on "p v_comb".
1999
2000 --
2001
2002 [Hey! Hint Hint Delete Delete!!!]
2003
2004 Fix 386 floating point so that floating point registers are real
2005 registers (but code can deal at run-time if they are missing, like
2006 mips and 68k).  This would clean up "info float" and related stuff.
2007
2008 --
2009
2010 gcc -g -c enummask.c then gdb enummask.o, then "p v".  GDB complains
2011 about not being able to access memory location 0.
2012
2013 -------------------- enummask.c
2014 enum mask
2015 {
2016   ANIMAL = 0,
2017   VEGETABLE = 1,
2018   MINERAL = 2,
2019   BASIC_CATEGORY = 3,
2020
2021   WHITE = 0,
2022   BLUE = 4,
2023   GREEN = 8,
2024   BLACK = 0xc,
2025   COLOR = 0xc,
2026
2027   ALIVE = 0x10,
2028
2029   LARGE = 0x20
2030 } v;
2031
2032 --
2033
2034 If try to modify value in file with "set write off" should give
2035 appropriate error not "cannot access memory at address 0x65e0".
2036
2037 --
2038
2039 Allow core file without exec file on RS/6000.
2040
2041 --
2042
2043 Make sure "shell" with no arguments works right on DOS.
2044
2045 --
2046
2047 Make gdb.ini (as well as .gdbinit) be checked on all platforms, so
2048 the same directory can be NFS-mounted on unix or DOS, and work the
2049 same way.
2050
2051 --
2052
2053 [Is this another delete???]
2054
2055 Get SECT_OFF_TEXT stuff out of objfile_relocate (might be needed to
2056 get RS/6000 to work right, might not be immediately relevant).
2057
2058 --
2059
2060 Work out some kind of way to allow running the inferior to be done as
2061 a sub-execution of, eg. breakpoint command lists.  Currently running
2062 the inferior interupts any command list execution.  This would require
2063 some rewriting of wait_for_inferior & friends, and hence should
2064 probably be done in concert with the above.
2065
2066 --
2067
2068 Add function arguments to gdb user defined functions.
2069
2070 --
2071
2072 Add convenience variables that refer to exec file, symbol file,
2073 selected frame source file, selected frame function, selected frame
2074 line number, etc.
2075
2076 --
2077
2078 Modify the handling of symbols grouped through BINCL/EINCL stabs to
2079 allocate a partial symtab for each BINCL/EINCL grouping.  This will
2080 seriously decrease the size of inter-psymtab dependencies and hence
2081 lessen the amount that needs to be read in when a new source file is
2082 accessed.
2083
2084 --
2085
2086 Add a command for searching memory, a la adb.  It specifies size,
2087 mask, value, start address.  ADB searches until it finds it or hits
2088 an error (or is interrupted).
2089
2090 --
2091
2092 Remove the range and type checking code and documentation, if not
2093 going to implement.
2094
2095 # Local Variables:
2096 # mode: text
2097 # End: