Hi Joachim,
I think automatic regression notification at least to the author is a good idea.
Probably I can do it in a nearest time. Unfortunately, right now I fail to get your code up.
Also I believe it is a good style to check commits for regressions before pushing them.
But maybe GHC community is less performance oriented.
25.07.2014, 01:41, "Joachim Breitner"
Hi Alexander,
Am Donnerstag, den 24.07.2014, 19:22 +0400 schrieb Alexander Pakhomov:
I've been informed you created http://ghcspeed-nomeata.rhcloud.com. I believe it is necessary for any serious perfomance work. Do you need some help? What have you done and what problems are unsolved? thanks for your interest.
One way to help is to monitor the page for regressions, and notify whoever caused it. I’m doing that from time to time, relying on the Latest Results summary.
Then I don’t find the codespeed software to be perfect. The thread at https://groups.google.com/forum/#!topic/codespeed/yY5-kPrcG94 discusses some of the things I don’t like. So if you feel like hacking Python, just hack away on it and I’ll happily accept pull requests at https://github.com/nomeata/codespeed (branch ghc), or directly upstream.
Eventually, the whole setup should be moved to some dedicated and official hardware, but I’m not in a hurry with this. It’s been running for just a week or two, and I’d like to observe how its behaving.
Greetings, Joachim
PS: I prefer to discuss these things on the appropriate mailing list and in public, if only for the archive. If you agree, simply reply to the ghc-dev list (quoting in full) and I’ll read your reply there.
-- Joachim “nomeata” Breitner mail@joachim-breitner.de • http://www.joachim-breitner.de/ Jabber: nomeata@joachim-breitner.de • GPG-Key: 0xF0FBF51F Debian Developer: nomeata@debian.org