* [rfa] PROBLEMS: document 'constructor breakpoint ignored' bug
@ 2003-06-24 21:07 Michael Elizabeth Chastain
2003-06-25 21:46 ` Andrew Cagney
0 siblings, 1 reply; 6+ messages in thread
From: Michael Elizabeth Chastain @ 2003-06-24 21:07 UTC (permalink / raw)
To: gdb-patches
This patch documents the notorious "constructor breakpoints ignored"
problem in the gdb PROBLEMS file.
Okay to apply this to mainline?
Okay to apply this to the 6.0 branch?
Michael C
2003-06-24 Michael Chastain <mec@shout.net>
* PROBLEMS: Document pr gdb/1091 and pr gdb/1193,
the "constructor breakpoints ignored" bug.
Index: PROBLEMS
===================================================================
RCS file: /cvs/src/src/gdb/PROBLEMS,v
retrieving revision 1.15
diff -c -3 -r1.15 PROBLEMS
*** PROBLEMS 23 Jun 2003 03:28:13 -0000 1.15
--- PROBLEMS 24 Jun 2003 20:20:53 -0000
***************
*** 3,6 ****
--- 3,21 ----
See also: http://www.gnu.org/software/gdb/bugs/
+ gdb/1091: Constructor breakpoints ignored
+ gdb/1193: g++ 3.3 creates multiple constructors: gdb 5.3 can't set breakpoints
+ When gcc 3.x compiles a C++ constructor or C++ destructor, it generates
+ 2 or 3 different versions of the object code. These versions have
+ unique mangled names (they have to, in order for linking to work), but
+ they have identical source code names, which leads to a great deal of
+ confusion. Specifically, if you set a breakpoint in a constructor or a
+ destructor, gdb will put a breakpoint in one of the versions, but your
+ program may execute the other version. This makes it impossible to set
+ breakpoints reliably in constructors or destructors.
+
+ gcc 3.x generates these multiple object code functions in order to
+ implement virtual base classes. gcc 2.x generated just one object code
+ function with a hidden parameter, but gcc 3.x conforms to a multi-vendor
+ ABI for C++ which requires multiple object code functions.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [rfa] PROBLEMS: document 'constructor breakpoint ignored' bug
2003-06-24 21:07 [rfa] PROBLEMS: document 'constructor breakpoint ignored' bug Michael Elizabeth Chastain
@ 2003-06-25 21:46 ` Andrew Cagney
2003-06-26 1:00 ` Daniel Jacobowitz
0 siblings, 1 reply; 6+ messages in thread
From: Andrew Cagney @ 2003-06-25 21:46 UTC (permalink / raw)
To: Michael Elizabeth Chastain; +Cc: gdb-patches
[-- Attachment #1: Type: text/plain, Size: 28 bytes --]
It's C++ so Daniel?
Andrew
[-- Attachment #2: mailbox-message://ac131313@movemail/fsf/gdb/patches#18906999 --]
[-- Type: message/rfc822, Size: 4471 bytes --]
From: Michael Elizabeth Chastain <mec@shout.net>
To: gdb-patches@sources.redhat.com
Subject: [rfa] PROBLEMS: document 'constructor breakpoint ignored' bug
Date: Tue, 24 Jun 2003 16:26:29 -0400
Message-ID: <200306242026.h5OKQThr012996@duracef.shout.net>
This patch documents the notorious "constructor breakpoints ignored"
problem in the gdb PROBLEMS file.
Okay to apply this to mainline?
Okay to apply this to the 6.0 branch?
Michael C
2003-06-24 Michael Chastain <mec@shout.net>
* PROBLEMS: Document pr gdb/1091 and pr gdb/1193,
the "constructor breakpoints ignored" bug.
Index: PROBLEMS
===================================================================
RCS file: /cvs/src/src/gdb/PROBLEMS,v
retrieving revision 1.15
diff -c -3 -r1.15 PROBLEMS
*** PROBLEMS 23 Jun 2003 03:28:13 -0000 1.15
--- PROBLEMS 24 Jun 2003 20:20:53 -0000
***************
*** 3,6 ****
--- 3,21 ----
See also: http://www.gnu.org/software/gdb/bugs/
+ gdb/1091: Constructor breakpoints ignored
+ gdb/1193: g++ 3.3 creates multiple constructors: gdb 5.3 can't set breakpoints
+ When gcc 3.x compiles a C++ constructor or C++ destructor, it generates
+ 2 or 3 different versions of the object code. These versions have
+ unique mangled names (they have to, in order for linking to work), but
+ they have identical source code names, which leads to a great deal of
+ confusion. Specifically, if you set a breakpoint in a constructor or a
+ destructor, gdb will put a breakpoint in one of the versions, but your
+ program may execute the other version. This makes it impossible to set
+ breakpoints reliably in constructors or destructors.
+
+ gcc 3.x generates these multiple object code functions in order to
+ implement virtual base classes. gcc 2.x generated just one object code
+ function with a hidden parameter, but gcc 3.x conforms to a multi-vendor
+ ABI for C++ which requires multiple object code functions.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [rfa] PROBLEMS: document 'constructor breakpoint ignored' bug
2003-06-25 21:46 ` Andrew Cagney
@ 2003-06-26 1:00 ` Daniel Jacobowitz
2003-06-26 12:49 ` Paul Koning
0 siblings, 1 reply; 6+ messages in thread
From: Daniel Jacobowitz @ 2003-06-26 1:00 UTC (permalink / raw)
To: Andrew Cagney; +Cc: Michael Elizabeth Chastain, gdb-patches
Sure, since it's not going to be fixed in time. Go for it.
On Wed, Jun 25, 2003 at 05:02:28PM -0400, Andrew Cagney wrote:
> It's C++ so Daniel?
>
> Andrew
> Date: Tue, 24 Jun 2003 16:26:29 -0400
> From: Michael Elizabeth Chastain <mec@shout.net>
> Subject: [rfa] PROBLEMS: document 'constructor breakpoint ignored' bug
> To: gdb-patches@sources.redhat.com
>
> This patch documents the notorious "constructor breakpoints ignored"
> problem in the gdb PROBLEMS file.
>
> Okay to apply this to mainline?
>
> Okay to apply this to the 6.0 branch?
>
> Michael C
>
> 2003-06-24 Michael Chastain <mec@shout.net>
>
> * PROBLEMS: Document pr gdb/1091 and pr gdb/1193,
> the "constructor breakpoints ignored" bug.
>
> Index: PROBLEMS
> ===================================================================
> RCS file: /cvs/src/src/gdb/PROBLEMS,v
> retrieving revision 1.15
> diff -c -3 -r1.15 PROBLEMS
> *** PROBLEMS 23 Jun 2003 03:28:13 -0000 1.15
> --- PROBLEMS 24 Jun 2003 20:20:53 -0000
> ***************
> *** 3,6 ****
> --- 3,21 ----
>
> See also: http://www.gnu.org/software/gdb/bugs/
>
> + gdb/1091: Constructor breakpoints ignored
> + gdb/1193: g++ 3.3 creates multiple constructors: gdb 5.3 can't set breakpoints
>
> + When gcc 3.x compiles a C++ constructor or C++ destructor, it generates
> + 2 or 3 different versions of the object code. These versions have
> + unique mangled names (they have to, in order for linking to work), but
> + they have identical source code names, which leads to a great deal of
> + confusion. Specifically, if you set a breakpoint in a constructor or a
> + destructor, gdb will put a breakpoint in one of the versions, but your
> + program may execute the other version. This makes it impossible to set
> + breakpoints reliably in constructors or destructors.
> +
> + gcc 3.x generates these multiple object code functions in order to
> + implement virtual base classes. gcc 2.x generated just one object code
> + function with a hidden parameter, but gcc 3.x conforms to a multi-vendor
> + ABI for C++ which requires multiple object code functions.
>
--
Daniel Jacobowitz
MontaVista Software Debian GNU/Linux Developer
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [rfa] PROBLEMS: document 'constructor breakpoint ignored' bug
2003-06-26 1:00 ` Daniel Jacobowitz
@ 2003-06-26 12:49 ` Paul Koning
2003-06-26 13:44 ` Daniel Jacobowitz
0 siblings, 1 reply; 6+ messages in thread
From: Paul Koning @ 2003-06-26 12:49 UTC (permalink / raw)
To: drow; +Cc: ac131313, mec, gdb-patches
>>>>> "Daniel" == Daniel Jacobowitz <drow@mvista.com> writes:
Daniel> Sure, since it's not going to be fixed in time. Go for it.
As a workaround, I patched my local copy of gdb to use the "verbose"
demangler mode, so the various flavors of constructors/destructors DO
have different names. They are *weird* names, but at least you can
refer to their entry points by name that way.
It's a trivial patch. It changes the way constructors are named in
gdb, but that doesn't seem to be a big deal if it changes them from
conventionally named but unuseable things into strangely named but
useable...
Interested in that approach, as a stopgap -- rather than document the
restriction?
paul
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [rfa] PROBLEMS: document 'constructor breakpoint ignored' bug
2003-06-26 12:49 ` Paul Koning
@ 2003-06-26 13:44 ` Daniel Jacobowitz
0 siblings, 0 replies; 6+ messages in thread
From: Daniel Jacobowitz @ 2003-06-26 13:44 UTC (permalink / raw)
To: Paul Koning; +Cc: ac131313, mec, gdb-patches
On Thu, Jun 26, 2003 at 08:49:47AM -0400, Paul Koning wrote:
> >>>>> "Daniel" == Daniel Jacobowitz <drow@mvista.com> writes:
>
> Daniel> Sure, since it's not going to be fixed in time. Go for it.
>
> As a workaround, I patched my local copy of gdb to use the "verbose"
> demangler mode, so the various flavors of constructors/destructors DO
> have different names. They are *weird* names, but at least you can
> refer to their entry points by name that way.
>
> It's a trivial patch. It changes the way constructors are named in
> gdb, but that doesn't seem to be a big deal if it changes them from
> conventionally named but unuseable things into strangely named but
> useable...
>
> Interested in that approach, as a stopgap -- rather than document the
> restriction?
No, I'm not. Using verbose mode will break a number of other things.
The verbose output is even less machine-parseable than the non-verbose
output.
--
Daniel Jacobowitz
MontaVista Software Debian GNU/Linux Developer
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [rfa] PROBLEMS: document 'constructor breakpoint ignored' bug
@ 2003-06-26 16:27 Michael Elizabeth Chastain
0 siblings, 0 replies; 6+ messages in thread
From: Michael Elizabeth Chastain @ 2003-06-26 16:27 UTC (permalink / raw)
To: ac131313, drow; +Cc: gdb-patches
drow> Sure, since it's not going to be fixed in time. Go for it.
Committed to HEAD.
Committed to gdb_6_0-branch.
Michael C
2003-06-24 Michael Chastain <mec@shout.net>
* PROBLEMS: Document pr gdb/1091 and pr gdb/1193,
the "constructor breakpoints ignored" bug.
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2003-06-26 16:27 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-06-24 21:07 [rfa] PROBLEMS: document 'constructor breakpoint ignored' bug Michael Elizabeth Chastain
2003-06-25 21:46 ` Andrew Cagney
2003-06-26 1:00 ` Daniel Jacobowitz
2003-06-26 12:49 ` Paul Koning
2003-06-26 13:44 ` Daniel Jacobowitz
2003-06-26 16:27 Michael Elizabeth Chastain
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox