[physfs] 7z support working?

[Jeremy]

Think I found the problem.I had added LzmaStateDecode.* to the project as
well as the other files, and it compiles and runs, but crashes, presumably
because LzmaStateDecode.h also defines a CLzmaDecoderState struct like
LzmaDecode.h Removing the LzmaStateDecode.* files from the project at least
gets the test working again, so it'll probably work in my app now as well,
since the crash was the same.

J


On Feb 3, 2008 1:32 PM, Dennis Schridde <devurandom at gmx.net> wrote:

> Am Sonntag, 3. Februar 2008 19:04:54 schrieb Jeremy:
> > Is there something special you have to do in order to get 7z support
> > working? I set up the PHYSFS_SUPPORTS_7Z define, and everything seems
> > peachy until I attempt to mount a 7z archive, and it crashes when
> > attempting to use an uninitialized struct.
> Which struct?
>
> > Using the test_physfs.c and
> > latest from svn. Even in the app I'm using an older version of physfs
> in,
> > it won't mount a 7z file. It doesn't crash, but there appears to be a
> bug
> > reading the header. If I fix that it gets farther and then crashes.
> Can I have a look at that fix? And which bug do you mean?
>
> > Hopefully I'm just forgetting to do something and this support isn't
> just
> > broken in SVN.
> It works great here...
>
> Can you provide an equivalent of "bt full"? And examine the involved
> variables?
> The line 220, it seems to talk about, is:
>    outStream[nowPos++] = dictionary[dictionaryPos] = dictionary[pos];
> Without knowing what nowPos, dictionary, dictionaryPos, pos and the
> associated
> sizes are, it might get difficult to fix.
>
> And is it just me, or are msvc backtraces more difficult to read than gdb
> ones? It also seems as if it is missing information, like files...
>
> --Dennis
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://icculus.org/pipermail/physfs/attachments/20080203/2605c96c/attachment.htm>