gdb/TODO

   1 If you find inaccuracies in this list, please send mail to
   2 bug-gdb@prep.ai.mit.edu.
   3
   4 General To Do List
   5 ------------------
   6
   7 This to do list is probably not up to date, and opinions may vary
   8 about the importance or even desirability of some of the items.
   9
  10 START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that
  11 is its default value.  Clean this up.
  12
  13 It should be possible to use symbols from shared libraries before we know
  14 exactly where the libraries will be loaded.  E.g. "b perror" before running
  15 the program.  This could maybe be done as an extension of the "breakpoint
  16 re-evaluation" after new symbols are loaded.
  17
  18 Make single_step() insert and remove breakpoints in one operation.
  19
  20 Speed up single stepping by avoiding extraneous ptrace calls.
  21
  22 Speed up single stepping by not inserting and removing breakpoints
  23 each time the inferior starts and stops.
  24
  25 Breakpoints should not be inserted and deleted all the time.  Only the
  26 one(s) there should be removed when we have to step over one.  Support
  27 breakpoints that don't have to be removed to step over them.
  28
  29 Update gdbint.texinfo to include doc on the directory structure and
  30 the various tricks of building gdb.
  31
  32 Do a tutorial in gdb.texinfo on how to do simple things in gdb.
  33 E.g. how to set a breakpoint that just prints something and continues.
  34 How to break on aborts.  Etc.
  35
  36 Provide "voodoo" debugging of core files.  This creates a zombie
  37 process as a child of the debugger, and loads it up with the data,
  38 stack, and regs of the core file.  This allows you to call functions
  39 in the executable, to manipulate the data in the core file.
  40
  41 GDB reopens the source file on every line, as you "next" through it.
  42
  43 Referencing the vtbl member of a struct doesn't work.  It prints OK
  44 if you print the struct, but it gets 0 if you try to deref it.
  45
  46 Persistent command history: A feature where you could save off a list
  47 of the commands you did, so you can edit it into something that will bring
  48 the target to the same place every time you source it.
  49 This would also be useful for automated fast watchpointing; if you go
  50 past the place where it watchpoints, you just start it over again and
  51 do it more carefully.
  52
  53 Deal with the SunOS 4.0 and 4.1.1 ptrace bug that loses the registers if
  54 the stack is paged out.
  55
  56 Finish the C++ exception handling stub routines.  Lint points them out
  57 as unused statics functions.
  58
  59 Perhaps "i source" should take an argument like that of "list".
  60
  61 See if coredep.c's fetch_core_registers can be used on more machines.
  62 E.g. MIPS (mips-xdep.c).
  63
  64 unpack_double() does not handle IEEE float on the target unless the host
  65 is also IEEE.  Death on a vax.
  66
  67 Set up interface between GDB and INFO so that you can hop into interactive
  68 INFO and back out again.  When running under Emacs, should use Emacs
  69 info, else fork the info program.  Installation of GDB should install
  70 its texinfo files into the info tree automagically, including the readline
  71 texinfo files.
  72
  73 "help address" ought to find the "help set print address" entry.
  74
  75 Remove the VTBL internal guts from printouts of C++ structs, unless
  76 vtblprint is set.
  77
  78 Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if
  79 it matches the source line indicated.
  80
  81 The prompt at end of screen should accept space as well as CR.
  82
  83 Check STORE_RETURN_VALUE on all architectures.  Check near it in tm-sparc.h
  84 for other bogosities.
  85
  86 Check for storage leaks in GDB, I'm sure there are a lot!
  87
  88 vtblprint of a vtbl should demangle the names it's printing.
  89
  90 Backtrace should point out what the currently selected frame is, in
  91 its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar,
  92 ...)" rather than "#3 foo (bar, ...)".
  93
  94 "i program" should work for core files, and display more info, like what
  95 actually caused it to die.
  96
  97 "x/10i" should shorten the long name, if any, on subsequent lines.
  98
  99 Check through the code for FIXME comments and fix them.  dbxread.c,
 100 blockframe.c, and plenty more. (I count 634 as of 940621 - sts)
 101
 102 "next" over a function that longjumps, never stops until next time you happen
 103 to get to that spot by accident.  E.g. "n" over execute_command which has
 104 an error.
 105
 106 Watchpoints seem not entirely reliable, though they haven't failed me recently.
 107
 108 "set zeroprint off", don't bother printing members of structs which are entirely
 109 zero.  Useful for those big structs with few useful members.
 110
 111 GDB does four ioctl's for every command, probably switching terminal modes
 112 to/from inferior or for readline or something.
 113
 114 terminal_ours versus terminal_inferior: cache state.  Switch should be a noop
 115 if the state is the same, too.
 116
 117 ptype $i6 = void??!
 118
 119 Clean up invalid_float handling so gdb doesn't coredump when it tries to
 120 access a NaN.  While this might work on SPARC, other machines are not
 121 configured right.
 122
 123 "b value_at ; commands ; continue ; end" stops EVERY OTHER TIME!
 124 Then once you enter a command, it does the command, runs two more
 125 times, and then stops again!  Bizarre...  (This behaviour has been
 126 modified, but it is not yet 100% predictable when e.g. the commands
 127 call functions in the child, and while there, the child is interrupted
 128 with a signal, or hits a breakpoint.)
 129
 130 help completion, help history should work.
 131
 132 Check that we can handle stack trace through varargs AND alloca in same
 133 function, on 29K.
 134
 135 wait_for_inferior loops forever if wait() gives it an error.
 136
 137 "i frame" shows wrong "arglist at" location, doesn't show where the args
 138 should be found, only their actual values.
 139
 140 There should be a way for "set" commands to validate the new setting
 141 before it takes effect.
 142
 143 A mess of floating point opcodes are missing from sparc-opcode.h.
 144 Also, a little program should test the table for bits that are
 145 overspecified or underspecified.  E.g. if the must-be-ones bits
 146 and the must-be-zeroes bits leave some fields unexamined, and the format
 147 string leaves them unprinted, then point this out.  If multiple
 148 non-alias patterns match, point this out too.  Finally, there should
 149 be a sparc-optest.s file that tries each pattern out.  This file
 150 should end up coming back the same (modulo transformation comments)
 151 if fed to "gas" then the .o is fed to gdb for disassembly.
 152
 153 Eliminate all the core_file_command's in all the xdep files.
 154 Eliminate separate declarations of registers[] everywhere.
 155
 156 "ena d" is ambiguous, why?  "ena delete" seems to think it is a command!
 157
 158 Perhaps move the tdep, xdep, and nat files, into the config
 159 subdirectories.  If not, at least straighten out their names so that
 160 they all start with the machine name.
 161
 162 inferior_status should include stop_print_frame.  It won't need to be
 163 reset in wait_for_inferior after bpstat_stop_status call, then.
 164
 165 i line VAR produces "Line number not known for symbol ``var''.".  I
 166 thought we were stashing that info now!
 167
 168 We should be able to write to random files at hex offsets like adb.
 169
 170 Make "target xxx" command interruptible.
 171
 172 Handle add_file with separate text, data, and bss addresses.  Maybe
 173 handle separate addresses for each segment in the object file?
 174
 175 Handle free_named_symtab to cope with multiply-loaded object files
 176 in a dynamic linking environment.  Should remember the last copy loaded,
 177 but not get too snowed if it finds references to the older copy.
 178
 179 Generalize and Standardize the RPC interface to a target program,
 180 improve it beyond the "ptrace" interface, and see if it can become a
 181 standard for remote debugging.
 182
 183 Remove all references to:
 184         text_offset
 185         data_offset
 186         text_data_start
 187         text_end
 188         exec_data_offset
 189         ...
 190 now that we have BFD.  All remaining are in machine dependent files.
 191
 192 When quitting with a running program, if a core file was previously
 193 examined, you get "Couldn't read float regs from core file"...if
 194 indeed it can't.  generic_mourn_inferior...
 195
 196 Have remote targets give a warning on a signal argument to
 197 target_resume.  Or better yet, extend the protocols so that it works
 198 like it does on the Unix-like systems.
 199
 200 Sort help and info output.
 201
 202 Re-organize help categories into things that tend to fit on a screen
 203 and hang together.
 204
 205 renote-nindy.c handles interrupts poorly; it error()s out of badly
 206 chosen places, e.g. leaving current_frame zero, which causes core dumps
 207 on the next command.
 208
 209 Add in commands like ADB's for searching for patterns, etc.  We should
 210 be able to examine and patch raw unsymboled binaries as well in gdb as
 211 we can in adb.  (E.g. increase the timeout in /bin/login without source).
 212
 213 Those xdep files that call register_addr without defining it are
 214 probably simply broken.  When reconfiguring this part of gdb, I could
 215 only make guesses about how to redo some of those files, and I
 216 probably guessed wrong, or left them "for later" when I have a
 217 machine that can attempt to build them.
 218
 219 When doing "step" or "next", if a few lines of source are skipped between
 220 the previous line and the current one, print those lines, not just the
 221 last line of a multiline statement.
 222
 223 When searching for C++ superclasses in value_cast in valops.c, we must
 224 not search the "fields", only the "superclasses".  There might be a
 225 struct with a field name that matches the superclass name.  This can
 226 happen when the struct was defined before the superclass (before the
 227 name became a typedef).
 228
 229 Handling of "&" address-of operator needs some serious overhaul
 230 for ANSI C and consistency on arrays and functions.
 231   For "float point[15];":
 232 ptype &point[4]   ==> Attempt to take address of non-lvalue.
 233   For "char *malloc();":
 234 ptype malloc      ==> "char *()";  should be same as
 235 ptype &malloc     ==> "char *(*)()"
 236 call printf ("%x\n", malloc)   ==> weird value, should be same as
 237 call printf ("%x\n", &malloc)  ==> correct value
 238
 239 Fix dbxread.c symbol reading in the presence of interrupts.  It
 240 currently leaves a cleanup to blow away the entire symbol table when a
 241 QUIT occurs.  (What's wrong with that? -kingdon, 28 Oct 1993).
 242
 243 Mipsread.c reads include files depth-first, because the dependencies
 244 in the psymtabs are way too inclusive (it seems to me).  Figure out what
 245 really depends on what, to avoid recursing 20 or 30 times while reading
 246 real symtabs.
 247
 248 value_add() should be subtracting the lower bound of arrays, if known,
 249 and possibly checking against the upper bound for error reporting.
 250
 251 mipsread.c symbol table allocation and deallocation should be checked.
 252 My suspicion is that it's full of memory leaks.
 253
 254 SunOS should have a target_lookup_symbol() for common'd things allocated
 255 by the shared library linker ld.so.
 256
 257 When listing source lines, check for a preceding \n, to verify that
 258 the file hasn't changed out from under us.
 259
 260 When listing source lines, eat leading whitespace corresponding to the
 261 line-number prefix we print.  This avoids long lines wrapping.
 262
 263 mipsread.c needs to check for old symtabs and psymtabs for the same
 264 files, the way it happens for dbxread.c and coffread.c, for VxWorks
 265 incremental symbol table reloading.
 266
 267 Get all the remote systems (where the protocol allows it) to be able to
 268 stop the remote system when the GDB user types ^C (like remote.c
 269 does).  For ebmon, use ^Ak.
 270
 271 Possible feature: A version of the "disassemble" command which shows
 272 both source and assembly code ("set symbol-filename on" is a partial
 273 solution).
 274
 275 xcoffexec.c should be eliminated, contents going into either exec.c
 276 or rs6000-nat.c.
 277