Thanks and sorry again for the delay, I had to experiment with it a bit and we had code freeze anyway. I debugged the remaining test failure and it's not related to your patches. A few other things I noticed:
- We should hold CS while accessing content read position in HTTPREQ_SetFilePointer. - The call to read_req_file from HTTPREQ_SetFilePointer looks suspicious. I think you added it to make QueryDataAvailable test happy, but it seems that we should make HTTPREQ_QueryDataAvailable aware of cache file reads instead. - Error handling seemed a bit suspicious at the times and I think it hid some problems like SetFilePointer failures in tests (we should probably use SetFilePointerEx instead).
While experimenting, I simplified the patch and fixed above problems. What do you think about this version https://gitlab.winehq.org/jacek/wine/-/commits/wininet-file-pointer/ ?