curl-library
autobuild page (was Re: Array Index size_t vs int)
Date: Sat, 29 Sep 2012 15:32:41 +0200 (CEST)
On Sat, 29 Sep 2012, Marc Hoersken wrote:
>> Finally I'll work on fixing it so that it doesn't re-parse all output on
>> every invoke, which is the basic reason it runs so long.
>
> Thanks for fixing the autobuilds table. It worked fine during the last days!
> :-)
Well thank you for pointing out the flaws, it made me finally get my act
together and to these changes!
I made exactly the changes I mentioned before. Now the script only has to
parse each build log once and subsequent invokes re-use the saved output, thus
all it has to do now is to parse all new build logs that have appeared since
last time. Also a much less waste of CPU cycles.
I also simplified it more and now the script produces a single build page for
all logs it cared about instead of the previous multi-page output. I believe I
set the current limit to 400 build rows, but there's no reason we can't grow
that further in case of need other than the mere size of the HTML page.
We continuously remove all logs older than 15 days but it still makes up to
almost 3GB of logs right now.
-- / daniel.haxx.se ------------------------------------------------------------------- List admin: http://cool.haxx.se/list/listinfo/curl-library Etiquette: http://curl.haxx.se/mail/etiquette.htmlReceived on 2012-09-29