[Poppler-bugs] [Bug 62735] Should not try to fopen a file we already have open

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=62735

--- Comment #14 from Thomas Freitag <Thomas.Freitag at alfa.de> ---
(In reply to comment #13)
> Created attachment 77289 [details] [review]
> add a GooFile class to encapsulate file read access using offsets
> 
> Hopefully, I have made the necessary additions for autotools and CMake to
> generate the 'pread64' and 'lseek64' definitions. I also removed the - as
> far as I can see - by now superfluous calls to 'stat' where file streams are
> constructed since we can just use 'GooFile::size'.
> 
> @Albert: I will run regtest on this version overnight (looking good so far),
> but I think this is basically ready for review.
> 
> @Thomas: I used Poppler's default regtest which AFAIK is non-threaded since
> the default pdftoppm is. I did do a bit of smoke testing using qpdfview and
> multi-threading.

I just ran Your patch through my regression test with the multi-threaded
version of pdftoppm, and I just got the same failed tests than in
https://bugs.freedesktop.org/show_bug.cgi?id=50992#c124 (beside the failed test
caused by tiling patterns, because we solved that in the meantime).
This also helps on systems where a virus scanner is installed: because we open
the PDF now only once, the virus scanner scans it only one time.
Also it seems to be a little bit faster, but this could just be a subjective
feeling.
I also compiled it under Windows and made a few tests without any problems.

Congratulation!!!!

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/poppler-bugs/attachments/20130402/9fe62c50/attachment.html>