Pending breakpoints in MI
Vladimir Prus
ghost@cs.msu.su
Thu Dec 13 16:03:00 GMT 2007
Joel Brobecker wrote:
> Volodya,
>
> You forgot to provide a ChangeLog...
Oops, sorry. Attached now -- there are three ChangeLog
files affected, and I provide bits for each.
>> gdb_breakpoint (char *address, char *condition,
>> int hardwareflag, int tempflag,
>> int thread, int ignore_count,
>> + int pending,
>> char **error_message)
>
> I can't believe I'm saying this, since I absolutely hate tabs, but
> these are the coding rules: You need to use a tab instead of 8 spaces
> in the line above. Compare for instance how the "char **error_message"
> line is indented.
It appears I've failed to re-configure Emacs properly after
rearranging my local environment. I've fixed all instances now
(and one function which got space indenting in the previous patch).
>> --- gdb/testsuite/gdb.mi/mi-pending.exp
>> (/patches/pending_mi_2_code_duplication) (revision 45)
>> +++ gdb/testsuite/gdb.mi/mi-pending.exp
>> (/patches/pending_mi_3_breakpoints) (revision 45) @@ -0,0 +1,75 @@
>> +# Copyright 2003, 2004, 2005, 2007 Free Software Foundation, Inc.
>
> Shouldn't the copyright year be only 2007?
Yes, right.
>> --- gdb/mi/mi-cmd-break.c (/patches/pending_mi_2_code_duplication)
>> (revision 45)
>> +++ gdb/mi/mi-cmd-break.c (/patches/pending_mi_3_breakpoints) (revision
>> 45) @@ -63,7 +63,10 @@ enum bp_type
>> breakpoint. -break-insert -t -h <location> --> insert a temporary
>> hw bp.
>> -break-insert -r <regexp> --> insert a bp at functions matching
>> - <regexp> */
>> + <regexp>
>> +
>> + The -f flag makes GDB create a pending breakpoint if no
>> + locations for breakpoint are found now. */
>
> Actually, I think that the entire comment needs to be rewritten
> in a way that the syntax is not described there at all. Just say
> that this function implements the -break-insert command which, as
> its name suggests, inserts a breakpoint. The syntax needs to be
> documented in the GDB manual. See
> http://www.sourceware.org/gdb/current/onlinedocs/gdb_25.html#SEC259
> for what is currently documented. Let's use this opportunity to
> make sure that every option (and only the supported options) are
> documented.
I've done so. Eli, does the manual changes in the attached patch look
good to you?
Thanks a lot for the review!
Revised patch attached, is it OK?
- Volodya
gdb/
Add support for pending breakpoints in MI.
* gdb.h (gdb_breakpoint): Add parameter pending.
* breakpoint.c (break_command_1): Fix formatting.
(gdb_breakpoint): Add parameter pending. Fix
formatting.
* mi/mi-cmd-break.c (mi_break_insert): Change
comment to refer to manual. Add support for
the -f flag, for pending breakpoint.
gdb/doc
* gdb.texinfo (GDB/MI Breakpoint Commands): Document
the -f option for -break-insert, remove -r option,
and clarify specification of location.
gdb/testsuite
* gdb.mi/mi-pending.exp: New.
* gdb.mi/mi-pending.c: New.
* gdb.mi/mi-pendshr.c: New.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: pending_mi_3_breakpoints.diff
Type: text/x-diff
Size: 9934 bytes
Desc: not available
URL: <http://sourceware.org/pipermail/gdb-patches/attachments/20071213/b79db0a2/attachment.bin>
More information about the Gdb-patches
mailing list