On Wed, 4 Apr 2012, Alexander Lamaison wrote:
> Somehow, this fixes the problems on Windows where writing would block after
> around 90MB. I'm not sure I understand the logic but it works!
Lovely. I'm not entirely sure why this fixes the problem we were chasing (on
IRC). It might be as trivial as LIBSSH2_CHANNEL_MINADJUST being too small so
that the remote at some point wants more space to be available before it sends
more.
Right now I went with "safe and seems to work" rather than "fully analyzed and
optimized".
The way the channel window updates can be done without "force" set (to avoid
small unnecessary updates) should also be less relevant these days as we do
less spurious updates than we used to do previously as the code flow has been
changed a bit recently when it comes to how the window updates are done.
Meaning there might not even be a noticable impact on performance at all.
I'm aiming for a 1.4.1 release ASAP.
-- / daniel.haxx.se _______________________________________________ libssh2-devel http://cool.haxx.se/cgi-bin/mailman/listinfo/libssh2-develReceived on 2012-04-04