old-cross-binutils/gdb/Projects
2012-06-03 15:36:32 +01:00

114 lines
4.8 KiB
Text

Suggested projects for aspiring or current GDB hackers
======================================================
(You should probably chat with kingdon@ai.mit.edu to make sure that
no one else is doing the project you chose).
Add watchpoints (break if a memory location changes). This would
usually have to involve constant single stepping, but occasionally
there is operating system support which gdb should be able to cleanly
use (e.g. on the 80386, there are 4 debug registers. By ptracing an
address into them, you can get a trap on writes or on reads and
writes).
Rewrite proceed, wait_for_inferior, and normal_stop to clean them up.
Suggestions:
1) Make each test in wait_for_inferior a seperate subroutine
call.
2) Combine wait_for_inferior and normal_stop to clean up
communication via global variables.
3) See if you can find some way to clean up the global
variables that are used; possibly group them by data flow
and information content?
Work out some kind of way to allow running the inferior to be done as
a sub-execution of, eg. breakpoint command lists. Currently running
the inferior interupts any command list execution. This would require
some rewriting of wait_for_inferior & friends, and hence should
probably be done in concert with the above.
Add function arguments to gdb user defined functions.
Add convenience variables that refer to exec file, symbol file,
selected frame source file, selected frame function, selected frame
line number, etc.
Add a "suspend" subcommand of the "continue" command to suspend gdb
while continuing execution of the subprocess. Useful when you are
debugging servers and you want to dodge out and initiate a connection
to a server running under gdb.
Make "handle" understand symbolic signal names.
Work out and implement a reasonably general mechanism for multi-threaded
processies. There are parts of one implemented in convex-dep.c, if
you want an example.
A standalone version of gdb on the i386 exists. Anyone who wants to
do some serious working cleaning it up and making it a general
standalone gdb should contact pace@wheaties.ai.mit.edu.
Add stab information to allow reasonable debugging of inline functions
(possibly they should show up on a stack backtrace? With a note
indicating that they weren't "real"?).
Implement support for specifying arbitrary locations of stack frames
(in practice, this usually requires specification of both the top and
bottom of the stack frame (fp and sp), since you *must* retrieve the
pc that was saved in the innermost frame).
Modify the naked "until" command to step until past the current source
line, rather than past the current pc value. This is tricky simply
because the low level routines have no way of specifying a multi-line
step range, and there is no way of saying "don't print stuff when we
stop" from above (otherwise could just call step many times).
Modify the handling of symbols grouped through BINCL/EINCL stabs to
allocate a partial symtab for each BINCL/EINCL grouping. This will
seriously decrease the size of inter-psymtab dependencies and hence
lessen the amount that needs to be read in when a new source file is
accessed.
Work out some method of saving breakpoints across the reloading of an
executable. Probably this should be by saving the commands by which
the breakpoints were set and re-executing them (as text locations may
change).
Do an "x/i $pc" after each stepi or nexti.
Modify all of the disassemblers to use printf_filtered to get correct
more filtering.
Modify gdb to work correctly with Pascal.
Rewrite macros that handle frame chaining and frameless functions.
They should be able to tell the difference between start, main, and a
frameless function called from main.
Work out what information would need to be included in an executable
by the compiler to allow gdb to debug functions which do not have a
frame pointer. Modify gdb and gcc to do this.
When `attached' to a program (via either OS support or remote
debugging), gdb should arrange to catch signals which the terminal
might send, as it is unlikely that the program will be able to notice
them. SIGINT and SIGTSTP are obvious examples.
Enhance the gdb manual with extra examples where needed.
Arrange for list_command not to use decode_line_1 and thus not require
symbols to be read in simply to read a source file.
Problem in xgdb; the readline library needs the terminal in CBREAK
mode for command line editing, but this makes it difficult to dispatch
on button presses. Possible solution: use a define to replace getc in
readline.c with a routine that does button dispatches. You should
probably see XGDB-README before you fiddle with XGDB. Also, someone
is implementing a new xgdb; it may not be worth while fiddling with
the old one.
# Local Variables:
# mode: text
# End: