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