we have reduce our dependency on many of the functions and defines in a windows mobile porting library (wince shunt).
Created attachment 346495 [details] [diff] [review] patch v.2
windows mobile doesn't have stat(). Is there anything else we can do here Chris?
Comment on attachment 346495 [details] [diff] [review] patch v.2 chris, any ideas on what we can do here?
All it needs to do is find the length of the file. Can you do this with WINCE api calls?
Created attachment 348924 [details] [diff] [review] patch v.3 how is something like this? (not tested)
chris, do you have any test cases I can run this through?
+ int current = ftell(oggz->file, 0, SEEK_CUR); 'current' needs to be a 'long' since that's what ftell returns. We'll need to #ifdef WINCE for this since the 'ftell' implementation loses data (fstat provides access to a 64bit size on systems that support that).
The seeking code should hit this so if you run the video mochitests they'll fail if this doesn't work.
Created attachment 355837 [details] [diff] [review] patch v.4 this compiles (yeah, a nice attribute of a patch), and passes mochitests (another nice attribute)
chris, review ping?
Comment on attachment 355837 [details] [diff] [review] patch v.4 I think the changes to oggz_offset_end should be wrapped in a #if for WINCE, and leave the existing code so that it runs as normal on other platforms. A patch with these changes should also be added to the media/liboggz directory with update.sh modified to apply the patch so it doesn't get lost when I update to newer liboggz versions. The README_MOZILLA will also need to be updated to refer to this bug and the patch. See the way the other patches are handled in those files.
Created attachment 358323 [details] [diff] [review] patch v.5