gdb/TODO

   1 If you find inaccuracies in this list, please send mail to
   2 gdb-patches@sourceware.cygnus.com.  If you would like to work on any
   3 of these, you should consider sending mail to the same address, to
   4 find out whether anyone else is working on it.
   5
   6
   7                         GDB 5.1 - Fixes
   8                         ===============
   9
  10 Below is a list of problems identified during the GDB 5.0 release
  11 cycle.  People hope to have these problems fixed in 5.1.
  12
  13 --
  14
  15 Wow, three bug reports for the same problem in one day!  We should
  16 probably make fixing this a real priority :-).
  17
  18 Anyway, thanks for reporting.
  19
  20 The following patch will fix the problems with setting breakpoints in
  21 dynamically loaded objects:
  22
  23    http://sourceware.cygnus.com/ml/gdb-patches/2000-05/msg00230.html
  24
  25 This patch isn't checked in yet (ping Michael/JimB), but I hope this
  26 will be in the next GDB release.
  27
  28 There should really be a test in the testsuite for this problem, since
  29 it keeps coming up :-(.  Any volunteers?
  30
  31 Mark
  32
  33 --
  34
  35                 GDB 5.1 - New features
  36                 ======================
  37
  38 The following new features should be included in 5.1.
  39
  40 --
  41
  42                 GDB 5.1 - Cleanups
  43                 ==================
  44
  45 The following code cleanups will hopefully be applied to GDB 5.1.
  46
  47 --
  48
  49                 GDB 5.1 - Known Problems
  50                 ========================
  51
  52 --
  53
  54 z8k
  55
  56 The z8k has suffered bit rot and is known to not build.  The problem
  57 was occuring in the opcodes directory.
  58
  59 --
  60
  61 The BFD directory requires bug-fixed AUTOMAKE et.al.
  62
  63 AUTOMAKE 1.4 incorrectly set the TEXINPUTS environment variable.  It
  64 contained the full path to texinfo.tex when it should have only
  65 contained the directory.  The bug has been fixed in the current
  66 AUTOMAKE sources.  Automake snapshots can be found in:
  67         ftp://sourceware.cygnus.com/pub/gdb/snapshots
  68 and     ftp://sourceware.cygnus.com/pub/binutils
  69
  70 --
  71
  72 Solaris 8 x86 CURSES_H problem
  73 http://sources.redhat.com/ml/gdb/2000-07/msg00038.html
  74
  75 The original problem was worked around with:
  76
  77     2000-06-06  Michael Snyder  <msnyder@cygnus.com>
  78
  79         * configure.in: Enable autoconf to find curses.h on Solaris 2.8.
  80         * configure: Regenerate.
  81
  82 When building both GDB and SID using the same source tree the problem
  83 will still occure. sid/component/configure.in mis-configures
  84 <curses.h> and leaves wrong information in the config cache.
  85
  86 --
  87
  88                 GDB 5.2 - Fixes
  89                 ===============
  90
  91 --
  92
  93                 GDB 5.2 - New features
  94                 ======================
  95
  96 --
  97
  98 GCC 3.0 ABI support (but hopefully sooner...).
  99
 100 --
 101
 102 Objective C/C++ support (but hopefully sooner...).
 103
 104 --
 105
 106 Import of readline 4.2
 107
 108 --
 109
 110                 GDB 5.2 - Cleanups
 111                 ==================
 112
 113 The following cleanups have been identified as part of GDB 5.2.
 114
 115 --
 116
 117 Remove old code that does not use ui_out functions and all the related
 118 "ifdef"s.  This also allows the elimination of -DUI_OUT from
 119 Makefile.in and configure.in.
 120
 121 --
 122
 123 Compiler warnings.
 124
 125 Eliminate warnings for all targets on at least one host for one of the
 126 -W flags.  Flags up for debate include: -Wswitch -Wcomment -trigraphs
 127 -Wtrigraphs -Wunused-function -Wunused-label -Wunused-variable
 128 -Wunused-value -Wchar-subscripts -Wtraditional -Wshadow -Wcast-qual
 129 -Wcast-align -Wwrite-strings -Wconversion -Wstrict-prototypes
 130 -Wmissing-prototypes -Wmissing-declarations -Wredundant-decls
 131 -Woverloaded-virtual -Winline
 132
 133 --
 134
 135 Deprecate, if not delete, the following:
 136
 137         register[]
 138         register_valid[]
 139         register_buffer()
 140         REGISTER_BYTE()
 141                 Replaced by, on the target side
 142                   supply_register()
 143                 and on core-gdb side:
 144                   {read,write}_register_gen()
 145                 Remote.c will need to use something
 146                 other than REGISTER_BYTE() and
 147                 REGISTER_RAW_SIZE() when unpacking
 148                 [gG] packets.
 149
 150         STORE_PSEUDO_REGISTER
 151         FETCH_PSEUDO_REGISTER
 152                 Now handed by the methods
 153                   gdbarch_{read,write}_register()
 154                 which sits between core GDB and
 155                 the register cache.
 156
 157         REGISTER_CONVERTIBLE
 158         REGISTER_CONVERT_TO_RAW
 159         REGISTER_CONVERT_TO_VIRTUAL
 160                 I think these three are redundant.
 161                 gdbarch_register_{read,write} can
 162                 do any conversion it likes.
 163
 164         REGISTER_VIRTUAL_SIZE
 165         MAX_REGISTER_VIRTUAL_SIZE
 166         REGISTER_VIRTUAL_TYPE
 167                 I think these can be replaced by
 168                 the pair:
 169                   FRAME_REGISTER_TYPE(frame, regnum)
 170                   REGISTER_TYPE(regnum)
 171
 172         DO_REGISTERS_INFO
 173                 Replace with
 174                  FRAME_REGISTER_INFO (frame, ...)
 175
 176         REGISTER_SIM_REGNO()
 177                 If nothing else rename this so that
 178                 how it relates to rawreg and the
 179                 regnum is clear.
 180
 181         REGISTER_BYTES
 182                 The size of the cache can be computed
 183                 on the fly.
 184
 185         IS_TRAPPED_INTERNALVAR
 186                 The pseudo registers should eventually make
 187                 this redundant.
 188
 189 --
 190
 191 Obsolete the targets:
 192
 193 arm*-wince-pe
 194 mips*-*-pe
 195 sh*-*-pe
 196
 197 --
 198
 199 Obsolete the protocols:
 200
 201 RDB?
 202
 203 ``As of version 5.3, WindRiver has removed the RDB server (RDB
 204 protocol support is built into gdb).''  -- Till.
 205
 206 --
 207
 208 Restructure gdb directory tree so that it avoids any 8.3 and 14
 209 filename problems.
 210
 211 --
 212
 213 Convert GDB build process to AUTOMAKE.
 214
 215 See also sub-directory configure below.
 216
 217 The current convention is (kind of) to use $(<header>_h) in all
 218 dependency lists.  It isn't done in a consistent way.
 219
 220 --
 221
 222                 GDB 5.2 - Known Problems
 223                 ========================
 224
 225 --
 226
 227                 Code Cleanups: General
 228                 ======================
 229
 230 The following are more general cleanups and fixes.  They are not tied
 231 to any specific release.
 232
 233
 234                         New Features and Fixes
 235                         ======================
 236
 237 These are harder than cleanups but easier than work involving
 238 fundamental architectural change.
 239
 240 --
 241
 242                         Language Support
 243                         ================
 244
 245 New languages come onto the scene all the time.
 246
 247 --
 248
 249 Re: Various C++ things
 250
 251 value_headof/value_from_vtable_info are worthless, and should be
 252 removed.  The one place in printcmd.c that uses it should use the RTTI
 253 functions.
 254
 255 RTTI for g++ should be using the typeinfo functions rather than the
 256 vtables.  The typeinfo functions are always at offset 4 from the
 257 beginning of the vtable, and are always right. The vtables will have
 258 weird names like E::VB sometimes.  The typeinfo function will always
 259 be "E type_info function", or somesuch.
 260
 261 value_virtual_fn_field needs to be fixed so there are no failures for
 262 virtual functions for C++ using g++.
 263
 264 Testsuite cases are the major priority right now for C++ support,
 265 since i have to make a lot of changes that could potentially break
 266 each other.
 267
 268 --
 269
 270
 271                         Symbol Support
 272                         ==============
 273
 274 --
 275
 276 Investiagate ways of reducing memory.
 277
 278 --
 279
 280 Investigate ways of improving load time.
 281
 282 --
 283
 284                         Testsuite Support
 285                         =================
 286
 287 There are never to many testcases.
 288
 289 --
 290
 291 Better thread testsuite.
 292
 293 --
 294
 295 Better C++ testsuite.
 296
 297 --
 298
 299                 Architectural Changes: General
 300                 ==============================
 301
 302 These are harder than simple cleanups / fixes and, consequently
 303 involve more work.  Typically an Architectural Change will be broken
 304 down into a more digestible set of cleanups and fixes.
 305
 306 --
 307
 308                 Architectural Change: Multi-arch et al.
 309                 =======================================
 310
 311 The long term objective is to remove all assumptions that there is a
 312 single target with a single address space with a single instruction
 313 set architecture and single application binary interface.
 314
 315 This is an ongoing effort.  The first milestone is to enable
 316 ``multi-arch'' where by all architectural decisions are made at
 317 runtime.
 318
 319 It should be noted that ``gdbarch'' is really ``gdbabi'' and
 320 ``gdbisa''.  Once things are multi-arched breaking that down correctly
 321 will become much easier.
 322
 323 --
 324
 325         Architectural Change: MI, LIBGDB and scripting languages
 326         ========================================================
 327
 328 See also architectural changes related to the event loop.  LIBGDB
 329 can't be finished until there is a generic event loop being used by
 330 all targets.
 331
 332 The long term objective is it to be possible to integrate GDB into
 333 scripting languages.
 334
 335 --
 336
 337                 Architectural Change: Async
 338                 ===========================
 339
 340 While GDB uses an event loop when prompting the user for input.  That
 341 event loop is not exploited by targets when they allow the target
 342 program to continue.  Typically targets still block in (target_wait())
 343 until the program again halts.
 344
 345 The closest a target comes to supporting full asynchronous mode are
 346 the remote targets ``async'' and ``extended-async''.
 347
 348 --
 349
 350 # Local Variables:
 351 # mode: text
 352 # End: