On Mon, 23 May 2011, Luca Barbato wrote:
> On 5/21/11 2:35 PM, Martin Storsjö wrote:
>
> > Hmm, I guess we should map error codes from winsock errors to AVERRORs
> > here, too, as we do internally in ff_neterrno() now. That's mostly
> > important for EAGAIN though, and I don't think we should hit that one
> > here.
>
> The two that are remapped should be taken care by the timeout loop I think.
Yeah, in normal cases. I was wondering what would happen if we'd get those
errors returned here for some strange reason - but earlier, we've just
been returning AVERROR(EIO) for that case anyway, so we shouldn't have any
code relying on tracing them back to exactly which error code they were,
for this codepath.
Patch ok then.
// Martin
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
