https://bugs.freedesktop.org/show_bug.cgi?id=67534

--- Comment #42 from Fridrich Strba <[email protected]> ---
(In reply to comment #41)
> The only mildly relevant-looking commit I found that is in libreoffice-4-0-5
> but not in libreoffice-4-0-4 is
> <http://cgit.freedesktop.org/libreoffice/core/commit/
> ?id=f7ff371fdc4c44820a1cfae90c2f1af704f276a2> "bnc#829017 fix issue with
> negative seeks in win32 shell extension," though.

Stephan, the commit fixes basically the fact that on 64-bit systems the
indexing was not working, because unsigned 32-bit integers were casting into
signed 64-bit ones which was producing seeks off the end of the files. It is
quite possible that since now the zip back-end works, it can take some time to
read the content.xml and extract the text to index then. There might be a
slight difference between the thumbnail extraction and the content.xml reading.
Since this is a network drive, it is possible that using unbuffered read might
mean a slowdown and the file being reported as being opened.

TL;DR summary: there might be a problem if the zip streams are extracted one
byte at a time, but that above-mentioned commit is correct and reverting it
would mean that we cannot search odf document anymore.

-- 
You are receiving this mail because:
You are the assignee for the bug.
_______________________________________________
Libreoffice-bugs mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/libreoffice-bugs

Reply via email to