Comment #18 on issue 23362 by craig.schlenter: glibc assert in  
WebCore::CSSParser::~CSSParser() (double free) -- gcc 4.4-specific?
http://code.google.com/p/chromium/issues/detail?id=23362

A quick summary of the current strict aliasing and gcc 4.4 situation (yes I  
see it is
probably not the problem as per comment 17 so this is mostly FYI):

1. If you build with gcc 4.4 you _must_ define a gyp variable  
gcc_version=44 (you can
do this in ~/.gyp/include.gypi and various other ways). This tweaks the v8
compilation flags to include -fno-strict-aliasing (originally just for  
dtoa.c but I
believe there are other aliasing issues in v8 too now). It also adds  
-fno-tree-vrp
which stopped some crashes in the v8 test suite.

2. Webcore.gyp currently uses -fno-strict-aliasing unconditionally i.e.  
regardless of
gcc version.

3. The change to common.gypi to add a -fno-strict-aliasing option was added  
I believe
to avoid a strict aliasing issue in stl-tree.h (in glibc) that is triggered  
in one of
the unit_tests. If you only compile the chrome target or #if 0 out that  
test, this is
unnecessary AFAIK. I think fta requested this because he builds unit tests  
etc. too.

I build in release mode with library=shared_library and gcc_version=44 and  
msnbc.com
works for me FWIW. My version of gcc is "gcc (GCC) 4.4.0 20090310 (Red Hat  
4.4.0-
0.24)" that I somehow managed to build from source on F9. I believe karmic  
and F11
etc. might have slightly more modern versions than that though.

Perhaps those people seeing the issue could try building webkit at a lower
optimisation level than the rest of the code and see if that works.

--
You received this message because you are listed in the owner
or CC fields of this issue, or because you starred this issue.
You may adjust your issue notification preferences at:
http://code.google.com/hosting/settings

--~--~---------~--~----~------------~-------~--~----~
Automated mail from issue updates at http://crbug.com/
Subscription options: http://groups.google.com/group/chromium-bugs
-~----------~----~----~----~------~----~------~--~---

Reply via email to