Re: Potential problems in GSMimeParse and libobjc.so.1.0.0

[Richard Frith-Macdonald]

On Thursday, March 20, 2003, at 11:02  am, Manuel Guesdon wrote:

> Hi,
>
> FYI, using valgrind, I get the following warnings on GSMimeParser
>
> ==425== Conditional jump or move depends on uninitialised value(s)
> ==425==    at 0x40E71CF6: _i_GSMimeParser_Private__unfoldHeader 
> (GSMime.m:2431)
> ==425==    by 0x40E6E369: _i_GSMimeParser__parse_ (GSMime.m:1158)

Thanks ... I think I've fixed that in CVS.

> and on libobjc v 1.0.0
>
> ==425== Invalid read of size 4
> ==425==    at 0x40F2060F: sarray_free (in /usr/lib/libobjc.so.1.0.0)
> ==425==    by 0x40F20939: sarray_free (in /usr/lib/libobjc.so.1.0.0)
> ==425==    by 0x40F2295D: __objc_update_dispatch_table_for_class (in 
> /usr/lib/libobjc.so.1.0.0)
> ==425==    by 0x40F22989: __objc_update_dispatch_table_for_class (in 
> /usr/lib/libobjc.so.1.0.0)

I sent Nicola (as he's the maintainer for objc in gcc) a fix for that 
one a few weeks ago.
I've been using the a patched libobjc since then ... here's the diff if 
you are interested.

diff
Description: Binary data