Bug 40938 - ice in create_tmp_var
Summary: ice in create_tmp_var
Status: RESOLVED WORKSFORME
Alias: None
Product: gcc
Classification: Unclassified
Component: c++ (show other bugs)
Version: 4.5.0
: P3 normal
Target Milestone: 4.5.0
Assignee: Not yet assigned to anyone
URL:
Keywords:
Depends on: 40866
Blocks:
  Show dependency treegraph
 
Reported: 2009-08-02 08:46 UTC by David Binderman
Modified: 2009-11-15 19:56 UTC (History)
2 users (show)

See Also:
Host: x86_64-suse-linux
Target:
Build:
Known to work:
Known to fail:
Last reconfirmed:


Attachments
C++ source code (144.30 KB, text/plain)
2009-08-02 08:48 UTC, David Binderman
Details

Note You need to log in before you can comment on or make changes to this bug.
Description David Binderman 2009-08-02 08:46:33 UTC
I just tried to compile Suse Linux package libqca2-2.0.2-2.2
with the gcc 4.5 mainline snapshot 20090730
and the compiler said

qca_core.cpp:1880:3: internal compiler error: in create_tmp_var, at gimplify.c:504
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

Preprocessed source attached. No special flags required.
Comment 1 David Binderman 2009-08-02 08:48:15 UTC
Created attachment 18284 [details]
C++ source code
Comment 2 Paolo Carlini 2009-08-02 09:33:20 UTC
PR40866 ?
Comment 3 David Binderman 2009-08-02 09:44:53 UTC
(In reply to comment #2)
> PR40866 ?

Maybe - there are many ways to fail a sanity check.

I notice that the line where it crashes is of the form

    gcc_assert(!X && Y);

It might be better to split this into two gcc_assert statements,
so the user would be better informed which X or Y is wrong.

    gcc_assert( !X);
    gcc_assert( Y);

Comment 4 Richard Biener 2009-08-02 09:47:11 UTC
No, the asserts are enabled in release builds, so they should be combined to
reduce executable size.
Comment 5 Paolo Carlini 2009-08-02 11:36:24 UTC
Still, you can certainly help the project by temporarily splitting the two checks in your local tree, rebuilding, and telling us if both PRs fail on the same one or not...
Comment 6 Paolo Carlini 2009-08-02 11:45:40 UTC
Or alternately playing a bit with the debugger, of course ;)
Comment 7 Paolo Carlini 2009-11-15 17:44:23 UTC
Works for me now (r154190). If you can still see something wrong, please reopen.
Comment 8 David Binderman 2009-11-15 19:19:31 UTC
(In reply to comment #7)
> Works for me now (r154190). 

I have no idea what r154190 means, but I can confirm
that the snapshot of 2009112 seems to work ok,
even at optimisation level -O3 -march=native.

This looks fixed to me.


Comment 9 Paolo Carlini 2009-11-15 19:56:24 UTC
r154190 is the subversion version I built and tested, this project doesn't use cvs anymore.