Commit | Line | Data |
---|---|---|
4a44dc1a | 1 | If you find inaccuracies in this list, please send mail to |
718bccf1 SS |
2 | bug-gdb@prep.ai.mit.edu. If you would like to work on any of these, |
3 | you should consider sending mail to the same address, to find out | |
4 | whether anyone else is working on it. | |
dd3b648e | 5 | |
4a44dc1a JK |
6 | General To Do List |
7 | ------------------ | |
8 | ||
718bccf1 SS |
9 | This list is probably not up to date, and opinions vary about the |
10 | importance or even desirability of some of the items. | |
dd3b648e | 11 | |
2b9fbee4 SS |
12 | START_INFERIOR_TRAPS_EXPECTED need never be defined to 2, since that |
13 | is its default value. Clean this up. | |
14 | ||
dd3b648e RP |
15 | It should be possible to use symbols from shared libraries before we know |
16 | exactly where the libraries will be loaded. E.g. "b perror" before running | |
17 | the program. This could maybe be done as an extension of the "breakpoint | |
18 | re-evaluation" after new symbols are loaded. | |
19 | ||
20 | Make single_step() insert and remove breakpoints in one operation. | |
21 | ||
22 | Speed up single stepping by avoiding extraneous ptrace calls. | |
23 | ||
24 | Speed up single stepping by not inserting and removing breakpoints | |
25 | each time the inferior starts and stops. | |
26 | ||
adf2bb58 JG |
27 | Breakpoints should not be inserted and deleted all the time. Only the |
28 | one(s) there should be removed when we have to step over one. Support | |
29 | breakpoints that don't have to be removed to step over them. | |
30 | ||
b1dcd6ac | 31 | Update gdbint.texinfo to include doc on the directory structure and |
dd3b648e RP |
32 | the various tricks of building gdb. |
33 | ||
34 | Do a tutorial in gdb.texinfo on how to do simple things in gdb. | |
35 | E.g. how to set a breakpoint that just prints something and continues. | |
36 | How to break on aborts. Etc. | |
37 | ||
dd3b648e RP |
38 | Provide "voodoo" debugging of core files. This creates a zombie |
39 | process as a child of the debugger, and loads it up with the data, | |
40 | stack, and regs of the core file. This allows you to call functions | |
41 | in the executable, to manipulate the data in the core file. | |
42 | ||
43 | GDB reopens the source file on every line, as you "next" through it. | |
44 | ||
45 | Referencing the vtbl member of a struct doesn't work. It prints OK | |
46 | if you print the struct, but it gets 0 if you try to deref it. | |
47 | ||
48 | Persistent command history: A feature where you could save off a list | |
49 | of the commands you did, so you can edit it into something that will bring | |
adf2bb58 | 50 | the target to the same place every time you source it. |
dd3b648e RP |
51 | This would also be useful for automated fast watchpointing; if you go |
52 | past the place where it watchpoints, you just start it over again and | |
53 | do it more carefully. | |
54 | ||
adf2bb58 JG |
55 | Deal with the SunOS 4.0 and 4.1.1 ptrace bug that loses the registers if |
56 | the stack is paged out. | |
dd3b648e RP |
57 | |
58 | Finish the C++ exception handling stub routines. Lint points them out | |
59 | as unused statics functions. | |
60 | ||
adf2bb58 | 61 | Perhaps "i source" should take an argument like that of "list". |
dd3b648e RP |
62 | |
63 | See if coredep.c's fetch_core_registers can be used on more machines. | |
64 | E.g. MIPS (mips-xdep.c). | |
65 | ||
dd3b648e RP |
66 | unpack_double() does not handle IEEE float on the target unless the host |
67 | is also IEEE. Death on a vax. | |
68 | ||
dd3b648e RP |
69 | Set up interface between GDB and INFO so that you can hop into interactive |
70 | INFO and back out again. When running under Emacs, should use Emacs | |
71 | info, else fork the info program. Installation of GDB should install | |
72 | its texinfo files into the info tree automagically, including the readline | |
adf2bb58 | 73 | texinfo files. |
dd3b648e | 74 | |
adf2bb58 | 75 | "help address" ought to find the "help set print address" entry. |
dd3b648e RP |
76 | |
77 | Remove the VTBL internal guts from printouts of C++ structs, unless | |
78 | vtblprint is set. | |
79 | ||
b1dcd6ac | 80 | Remove "at 0xnnnn" from the "b foo" response, if `print address off' and if |
dd3b648e RP |
81 | it matches the source line indicated. |
82 | ||
83 | The prompt at end of screen should accept space as well as CR. | |
84 | ||
dd3b648e RP |
85 | Check STORE_RETURN_VALUE on all architectures. Check near it in tm-sparc.h |
86 | for other bogosities. | |
87 | ||
88 | Check for storage leaks in GDB, I'm sure there are a lot! | |
89 | ||
90 | vtblprint of a vtbl should demangle the names it's printing. | |
91 | ||
4a44dc1a JK |
92 | Backtrace should point out what the currently selected frame is, in |
93 | its display, perhaps showing "@3 foo (bar, ...)" or ">3 foo (bar, | |
94 | ...)" rather than "#3 foo (bar, ...)". | |
dd3b648e RP |
95 | |
96 | "i program" should work for core files, and display more info, like what | |
97 | actually caused it to die. | |
98 | ||
dd3b648e RP |
99 | "x/10i" should shorten the long name, if any, on subsequent lines. |
100 | ||
101 | Check through the code for FIXME comments and fix them. dbxread.c, | |
054308ad | 102 | blockframe.c, and plenty more. (I count 634 as of 940621 - sts) |
dd3b648e RP |
103 | |
104 | "next" over a function that longjumps, never stops until next time you happen | |
105 | to get to that spot by accident. E.g. "n" over execute_command which has | |
106 | an error. | |
107 | ||
adf2bb58 | 108 | Watchpoints seem not entirely reliable, though they haven't failed me recently. |
dd3b648e RP |
109 | |
110 | "set zeroprint off", don't bother printing members of structs which are entirely | |
111 | zero. Useful for those big structs with few useful members. | |
112 | ||
113 | GDB does four ioctl's for every command, probably switching terminal modes | |
114 | to/from inferior or for readline or something. | |
115 | ||
116 | terminal_ours versus terminal_inferior: cache state. Switch should be a noop | |
117 | if the state is the same, too. | |
118 | ||
119 | ptype $i6 = void??! | |
120 | ||
121 | Clean up invalid_float handling so gdb doesn't coredump when it tries to | |
122 | access a NaN. While this might work on SPARC, other machines are not | |
123 | configured right. | |
124 | ||
125 | "b value_at ; commands ; continue ; end" stops EVERY OTHER TIME! | |
126 | Then once you enter a command, it does the command, runs two more | |
127 | times, and then stops again! Bizarre... (This behaviour has been | |
128 | modified, but it is not yet 100% predictable when e.g. the commands | |
129 | call functions in the child, and while there, the child is interrupted | |
130 | with a signal, or hits a breakpoint.) | |
131 | ||
dd3b648e RP |
132 | help completion, help history should work. |
133 | ||
dd3b648e RP |
134 | Check that we can handle stack trace through varargs AND alloca in same |
135 | function, on 29K. | |
136 | ||
137 | wait_for_inferior loops forever if wait() gives it an error. | |
138 | ||
dd3b648e RP |
139 | "i frame" shows wrong "arglist at" location, doesn't show where the args |
140 | should be found, only their actual values. | |
141 | ||
dd3b648e RP |
142 | There should be a way for "set" commands to validate the new setting |
143 | before it takes effect. | |
144 | ||
dd3b648e RP |
145 | A mess of floating point opcodes are missing from sparc-opcode.h. |
146 | Also, a little program should test the table for bits that are | |
147 | overspecified or underspecified. E.g. if the must-be-ones bits | |
148 | and the must-be-zeroes bits leave some fields unexamined, and the format | |
149 | string leaves them unprinted, then point this out. If multiple | |
150 | non-alias patterns match, point this out too. Finally, there should | |
151 | be a sparc-optest.s file that tries each pattern out. This file | |
152 | should end up coming back the same (modulo transformation comments) | |
153 | if fed to "gas" then the .o is fed to gdb for disassembly. | |
154 | ||
dd3b648e RP |
155 | Eliminate all the core_file_command's in all the xdep files. |
156 | Eliminate separate declarations of registers[] everywhere. | |
157 | ||
158 | "ena d" is ambiguous, why? "ena delete" seems to think it is a command! | |
159 | ||
77eb2e8f JK |
160 | Perhaps move the tdep, xdep, and nat files, into the config |
161 | subdirectories. If not, at least straighten out their names so that | |
dd3b648e RP |
162 | they all start with the machine name. |
163 | ||
164 | inferior_status should include stop_print_frame. It won't need to be | |
165 | reset in wait_for_inferior after bpstat_stop_status call, then. | |
166 | ||
167 | i line VAR produces "Line number not known for symbol ``var''.". I | |
168 | thought we were stashing that info now! | |
169 | ||
dd3b648e RP |
170 | We should be able to write to random files at hex offsets like adb. |
171 | ||
dd3b648e RP |
172 | Make "target xxx" command interruptible. |
173 | ||
174 | Handle add_file with separate text, data, and bss addresses. Maybe | |
175 | handle separate addresses for each segment in the object file? | |
176 | ||
177 | Handle free_named_symtab to cope with multiply-loaded object files | |
178 | in a dynamic linking environment. Should remember the last copy loaded, | |
179 | but not get too snowed if it finds references to the older copy. | |
180 | ||
dd3b648e | 181 | Generalize and Standardize the RPC interface to a target program, |
adf2bb58 JG |
182 | improve it beyond the "ptrace" interface, and see if it can become a |
183 | standard for remote debugging. | |
dd3b648e RP |
184 | |
185 | Remove all references to: | |
186 | text_offset | |
187 | data_offset | |
188 | text_data_start | |
189 | text_end | |
190 | exec_data_offset | |
191 | ... | |
192 | now that we have BFD. All remaining are in machine dependent files. | |
193 | ||
194 | When quitting with a running program, if a core file was previously | |
195 | examined, you get "Couldn't read float regs from core file"...if | |
196 | indeed it can't. generic_mourn_inferior... | |
197 | ||
4a44dc1a JK |
198 | Have remote targets give a warning on a signal argument to |
199 | target_resume. Or better yet, extend the protocols so that it works | |
200 | like it does on the Unix-like systems. | |
dd3b648e | 201 | |
dd3b648e RP |
202 | Sort help and info output. |
203 | ||
204 | Re-organize help categories into things that tend to fit on a screen | |
205 | and hang together. | |
206 | ||
dd3b648e RP |
207 | renote-nindy.c handles interrupts poorly; it error()s out of badly |
208 | chosen places, e.g. leaving current_frame zero, which causes core dumps | |
209 | on the next command. | |
210 | ||
211 | Add in commands like ADB's for searching for patterns, etc. We should | |
212 | be able to examine and patch raw unsymboled binaries as well in gdb as | |
213 | we can in adb. (E.g. increase the timeout in /bin/login without source). | |
214 | ||
215 | Those xdep files that call register_addr without defining it are | |
216 | probably simply broken. When reconfiguring this part of gdb, I could | |
217 | only make guesses about how to redo some of those files, and I | |
218 | probably guessed wrong, or left them "for later" when I have a | |
219 | machine that can attempt to build them. | |
220 | ||
dd3b648e RP |
221 | When doing "step" or "next", if a few lines of source are skipped between |
222 | the previous line and the current one, print those lines, not just the | |
223 | last line of a multiline statement. | |
224 | ||
225 | When searching for C++ superclasses in value_cast in valops.c, we must | |
226 | not search the "fields", only the "superclasses". There might be a | |
227 | struct with a field name that matches the superclass name. This can | |
228 | happen when the struct was defined before the superclass (before the | |
229 | name became a typedef). | |
230 | ||
36b9d39c JG |
231 | Handling of "&" address-of operator needs some serious overhaul |
232 | for ANSI C and consistency on arrays and functions. | |
233 | For "float point[15];": | |
dd3b648e | 234 | ptype &point[4] ==> Attempt to take address of non-lvalue. |
36b9d39c JG |
235 | For "char *malloc();": |
236 | ptype malloc ==> "char *()"; should be same as | |
237 | ptype &malloc ==> "char *(*)()" | |
9da0e790 | 238 | call printf ("%x\n", malloc) ==> weird value, should be same as |
36b9d39c | 239 | call printf ("%x\n", &malloc) ==> correct value |
dd3b648e | 240 | |
4a44dc1a JK |
241 | Fix dbxread.c symbol reading in the presence of interrupts. It |
242 | currently leaves a cleanup to blow away the entire symbol table when a | |
243 | QUIT occurs. (What's wrong with that? -kingdon, 28 Oct 1993). | |
ef98d5ac JG |
244 | |
245 | Mipsread.c reads include files depth-first, because the dependencies | |
246 | in the psymtabs are way too inclusive (it seems to me). Figure out what | |
247 | really depends on what, to avoid recursing 20 or 30 times while reading | |
248 | real symtabs. | |
249 | ||
250 | value_add() should be subtracting the lower bound of arrays, if known, | |
251 | and possibly checking against the upper bound for error reporting. | |
252 | ||
253 | mipsread.c symbol table allocation and deallocation should be checked. | |
254 | My suspicion is that it's full of memory leaks. | |
255 | ||
256 | SunOS should have a target_lookup_symbol() for common'd things allocated | |
257 | by the shared library linker ld.so. | |
258 | ||
259 | When listing source lines, check for a preceding \n, to verify that | |
260 | the file hasn't changed out from under us. | |
261 | ||
262 | When listing source lines, eat leading whitespace corresponding to the | |
263 | line-number prefix we print. This avoids long lines wrapping. | |
264 | ||
265 | mipsread.c needs to check for old symtabs and psymtabs for the same | |
266 | files, the way it happens for dbxread.c and coffread.c, for VxWorks | |
267 | incremental symbol table reloading. | |
268 | ||
4a44dc1a JK |
269 | Get all the remote systems (where the protocol allows it) to be able to |
270 | stop the remote system when the GDB user types ^C (like remote.c | |
271 | does). For ebmon, use ^Ak. | |
ce31a796 JK |
272 | |
273 | Possible feature: A version of the "disassemble" command which shows | |
2af4851f JK |
274 | both source and assembly code ("set symbol-filename on" is a partial |
275 | solution). | |
054308ad | 276 | |
40e4117a SS |
277 | investigate "x/s 0" (right now stops early) (I think maybe GDB is |
278 | using a 0 address for bad purposes internally). | |
279 | ||
91dc42c5 | 280 | Make "info path" and path_command work again (but independent of the |
40e4117a SS |
281 | environment either of gdb or that we'll pass to the inferior). |
282 | ||
283 | Make GDB understand the GCC feature for putting octal constants in | |
284 | enums. Make it so overflow on an enum constant does not error_type | |
285 | the whole type. Allow arbitrarily large enums with type attributes. | |
286 | Put all this stuff in the testsuite. | |
287 | ||
288 | Make TYPE_CODE_ERROR with a non-zero TYPE_LENGTH more useful (print | |
289 | the value in hex; process type attributes). Add this to the | |
290 | testsuite. This way future compilers can add new types and old | |
291 | versions of GDB can do something halfway reasonable. | |
292 | ||
293 | Clean up formatting of "info registers" on MIPS and 88k. See if it | |
294 | is possible to do this generically across all target architectures. | |
295 | ||
296 | GDB gets bfd/core.c and gdb/core.c confused (this should be easy to | |
297 | repeat even with something more recent than GDB 4.9). | |
298 | ||
299 | Check that unmatched RBRAC doesn't abort(). | |
300 | ||
301 | Fix mdebugread.c:parse_type to do fundamental types right (see | |
302 | rs6000_builtin_type in stabsread.c for what "right" is--the point is | |
303 | that the debug format fixes the sizes of these things and it shouldn't | |
304 | depend on stuff like TARGET_PTR_BIT and so on. For mdebug, there seem | |
305 | to be separate bt* codes for 64 bit and 32 bit things, and GDB should | |
306 | be aware of that). Also use a switch statement for clarity and speed. | |
307 | ||
308 | Investigate adding symbols in target_load--some targets do, some | |
309 | don't. | |
310 | ||
311 | Put dirname in psymtabs and change lookup*symtab to use dirname (so | |
312 | /foo/bar.c works whether compiled by cc /foo/bar.c, or cd /foo; cc | |
313 | bar.c). | |
314 | ||
806f810b | 315 | Merge xcoffread.c and coffread.c. Use breakpoint_re_set instead of |
40e4117a SS |
316 | fixup_breakpoints. |
317 | ||
318 | Fix byte order and int size sins in tm-a29k.h | |
319 | (EXTRACT_RETURN_VALUE). Perhaps should reproduce bug and verify fix | |
320 | (or perhaps should just fix it...). | |
321 | ||
322 | Make a watchpoint on a constant expression an error (or warning | |
323 | perhaps) | |
324 | ||
325 | Make a watchpoint which contains a function call an error (it is | |
326 | broken now, making it work is probably not worth the effort). | |
327 | ||
aac9c614 JK |
328 | Re-do calls to signal() in remote.c, and inflow.c (set_sigint_trap and |
329 | so on) to be independent of the debugging target, using target_stop to | |
330 | stop the inferior. Probably the part which is now handled by | |
331 | interrupt_query in remote.c can be done without any new features in | |
332 | the debugging target. | |
40e4117a SS |
333 | |
334 | New test case based on weird.exp but in which type numbers are not | |
335 | renumbered (thus multiply defining a type). This currently causes an | |
336 | infinite loop on "p v_comb". | |
337 | ||
338 | Nuke baseclass_addr. | |
339 | ||
340 | Nuke USG define. | |
341 | ||
342 | "source file more recent" loses on re-read | |
343 | ||
344 | Fix 386 floating point so that floating point registers are real | |
345 | registers (but code can deal at run-time if they are missing, like | |
346 | mips and 68k). This would clean up "info float" and related stuff. | |
347 | ||
348 | Look at Solaris bug in interrupt.exp. Can get out of syscall with | |
349 | PRSABORT (syscall will return EINTR) but merely doing that leads to a | |
350 | "can't read memory" error. | |
351 | ||
352 | gcc -g -c enummask.c then gdb enummask.o, then "p v". GDB complains | |
353 | about not being able to access memory location 0. | |
354 | ||
355 | -------------------- enummask.c | |
356 | enum mask | |
357 | { | |
358 | ANIMAL = 0, | |
359 | VEGETABLE = 1, | |
360 | MINERAL = 2, | |
361 | BASIC_CATEGORY = 3, | |
362 | ||
363 | WHITE = 0, | |
364 | BLUE = 4, | |
365 | GREEN = 8, | |
366 | BLACK = 0xc, | |
367 | COLOR = 0xc, | |
368 | ||
369 | ALIVE = 0x10, | |
370 | ||
371 | LARGE = 0x20 | |
372 | } v; | |
373 | ||
374 | If try to modify value in file with "set write off" should give | |
375 | appropriate error not "cannot access memory at address 0x65e0". | |
376 | ||
a36a812c JK |
377 | Why do we allow a target to omit standard register names (NO_STD_REGS |
378 | in tm-z8k.h)? I thought the standard register names were supposed to | |
379 | be just that, standard. | |
40e4117a SS |
380 | |
381 | Make DEBUG_EXPRESSIONS a maintenance command, dependent on | |
382 | MAINTENANCE_COMMANDS. | |
383 | ||
384 | Allow core file without exec file on RS/6000. | |
385 | ||
386 | Make sure "shell" with no arguments works right on DOS. | |
387 | ||
388 | Make gdb.ini (as well as .gdbinit) be checked on all platforms, so | |
389 | the same directory can be NFS-mounted on unix or DOS, and work the | |
390 | same way. | |
391 | ||
392 | cd ~/tmp/<M-?> causes infinite loop (where ~/tmp is a directory). | |
393 | ||
394 | Get SECT_OFF_TEXT stuff out of objfile_relocate (might be needed to | |
395 | get RS/6000 to work right, might not be immediately relevant). | |
396 | ||
397 | Clean up add_toc_to_loadinfo | |
398 | ||
399 | Think about attached processes and sharing terminal. | |
400 | ||
401 | John sez in reference to ignoring errors from tcsegpgrp if attach_flag: | |
402 | set_tty_state should not have any trouble with attached processes. | |
403 | Instead, the tty handling should leave the pgrp of the tty alone when | |
404 | attaching to processes (perhaps pass terminal_init_inferior a flag | |
405 | saying whether we're attaching). | |
406 | ||
407 | PAGE_SIZE redefined warnings on AIX. Probably should be using | |
408 | BFD_PAGE_SIZE throughout BFD. | |
409 | ||
718bccf1 SS |
410 | Rewrite proceed, wait_for_inferior, and normal_stop to clean them up. |
411 | Suggestions: | |
412 | ||
413 | 1) Make each test in wait_for_inferior a seperate subroutine | |
414 | call. | |
415 | 2) Combine wait_for_inferior and normal_stop to clean up | |
416 | communication via global variables. | |
417 | 3) See if you can find some way to clean up the global | |
418 | variables that are used; possibly group them by data flow | |
419 | and information content? | |
420 | ||
421 | Work out some kind of way to allow running the inferior to be done as | |
422 | a sub-execution of, eg. breakpoint command lists. Currently running | |
423 | the inferior interupts any command list execution. This would require | |
424 | some rewriting of wait_for_inferior & friends, and hence should | |
425 | probably be done in concert with the above. | |
426 | ||
427 | Add function arguments to gdb user defined functions. | |
428 | ||
429 | Add convenience variables that refer to exec file, symbol file, | |
430 | selected frame source file, selected frame function, selected frame | |
431 | line number, etc. | |
432 | ||
433 | Add a "suspend" subcommand of the "continue" command to suspend gdb | |
434 | while continuing execution of the subprocess. Useful when you are | |
435 | debugging servers and you want to dodge out and initiate a connection | |
436 | to a server running under gdb. | |
437 | ||
718bccf1 SS |
438 | Add stab information to allow reasonable debugging of inline functions |
439 | (possibly they should show up on a stack backtrace? With a note | |
440 | indicating that they weren't "real"?). | |
441 | ||
718bccf1 SS |
442 | Modify the naked "until" command to step until past the current source |
443 | line, rather than past the current pc value. This is tricky simply | |
444 | because the low level routines have no way of specifying a multi-line | |
445 | step range, and there is no way of saying "don't print stuff when we | |
446 | stop" from above (otherwise could just call step many times). | |
447 | ||
448 | Modify the handling of symbols grouped through BINCL/EINCL stabs to | |
449 | allocate a partial symtab for each BINCL/EINCL grouping. This will | |
450 | seriously decrease the size of inter-psymtab dependencies and hence | |
451 | lessen the amount that needs to be read in when a new source file is | |
452 | accessed. | |
453 | ||
454 | Do an "x/i $pc" after each stepi or nexti. | |
455 | ||
456 | Modify all of the disassemblers to use printf_filtered to get correct | |
457 | more filtering. | |
458 | ||
459 | Modify gdb to work correctly with Pascal. | |
460 | ||
718bccf1 SS |
461 | Add a command for searching memory, a la adb. It specifies size, |
462 | mask, value, start address. ADB searches until it finds it or hits | |
463 | an error (or is interrupted). | |
464 | ||
465 | # Local Variables: | |
466 | # mode: text | |
467 | # End: |