Re: USER32: cursoricon

[Alexandre Julliard]

Vitaly Lipatov <lav(a)etersoft.ru> writes:
> Changelog:
>     return fail if file format is unknown

This will make it fail for all normal cursor/icon files. These don't
have a header so you can't check a magic number.

-- 
Alexandre Julliard
julliard(a)winehq.org

USER32: cursoricon

[Vitaly Lipatov]

В сообщении от 11 Январь 2005 16:48 Alexandre Julliard написал(a):
> Vitaly Lipatov <lav(a)etersoft.ru> writes:
> > Changelog:
> >     return fail if file format is unknown
>
> This will make it fail for all normal cursor/icon files. These don't
> have a header so you can't check a magic number.
> Well, have you any suggestions about read access error in next lines?
>   for (i=0; i < entries; i++)
>       size += bits->idEntries[i].dwDIBSize + (fCursor ? sizeof(POINT16): 0);
> I guess there is incorrect number of entries (idCount).
> Can I check it for some reasonable values?
> I will try to extract test file for this bug...

-- 
Vitaly Lipatov, ALT Linux Team
Russia, Saint-Petersburg, www.etersoft.ru

USER32: cursoricon

[Alexandre Julliard]

Vitaly Lipatov <lav(a)etersoft.ru> writes:
> Well, have you any suggestions about read access error in next lines?
>   for (i=0; i < entries; i++)
>       size += bits->idEntries[i].dwDIBSize + (fCursor ? sizeof(POINT16): 0);
> I guess there is incorrect number of entries (idCount).
> Can I check it for some reasonable values?

Sure, check that it doesn't go beyond the end of the file.

-- 
Alexandre Julliard
julliard(a)winehq.org