On Wed, 28 Apr 2010, Alexander Lamaison wrote:
>> First, I disagree that these things "are almost always useless".
>
> I would love to see evidence of their benefit.
It of course depends entirely on the situation where it is made and for what
target etc, but I'm primarily involved in embedded systems on less powerful
CPUs and I've seen DRAMATIC improvements by simple changes such as these.
I'm however not suggesting that this change has any measurable impact on
libssh2 performance. As I explained, I saw several benefits with my edit.
> This is great. But can we have it as a function? Is there any reason not
> to write:
If you think it is important enough, then please proceed and push that change.
I don't feel strongly enough about it to keep it like that against anyone's
expressed will.
-- / daniel.haxx.se _______________________________________________ libssh2-devel http://cool.haxx.se/cgi-bin/mailman/listinfo/libssh2-develReceived on 2010-04-28