On Tue, 13 Sep 2011, Daniel Stenberg wrote:
> Quite right. I'll follow up with a fix that keeps the state better to avoid
> this. Although it seems to be a rather rare case in my tests...
See attachment.
But really, we need to put a more generic concept in place - as Henrik
mentioned before - to prevent that this kind of errors sneak in all over.
I'm gonna give this some further thoughts and come back with some kind of
suggestion.
I think my sftp_read and window adjust patches probably can go in anyway, as
they should at least make the situation better.
-- / daniel.haxx.se
_______________________________________________
libssh2-devel http://cool.haxx.se/cgi-bin/mailman/listinfo/libssh2-devel