Re: [Boost-testing] Test results processing transition..

Subject: Re: [Boost-testing] Test results processing transition..
From: Rene Rivera (grafikrobot_at_[hidden])
Date: 2015-01-20 11:22:57

On Thu, Jan 15, 2015 at 12:33 PM, Rene Rivera <grafikrobot_at_[hidden]> wrote:

>
> First run (total) ~28.0 min
> Second run (total) ~13.0 min
>
> First run (git) ~6.5 min
> Second run (git) ~1.5 min
>
> First run (build_results) ~19.0 min
> Second run (build_results) 10.5 min
>
> Hence, the core of building the results is most of the time. But there's
> still some considerable overhead. I did notice that the result build tool
> is compiled every time for some reason. And it's not supposed to be :-( As
> for the core result building.. From eyeing it, it seems the slow part is
> the XML reading. And then there's git.. We could probably reduce that time
> by only getting selected portions of Boost (enough to build the tools).
>

I've now changed the script so that: (a) it only fetches the Boost subrepos
needed to build the report tool, and (b) only attempt to build the report
tool (instead of all the regression tools).

The times are now approximately:

First run (total) ~ 25 minutes
Second run (total) ~14.5 min

First run (git) ~2.5 min
Second run (git) ~0.8 min

First run (update_tools) ~2.0 min
Second run (update_tools) ~0.8 min

First run (build_results) ~21.0 min
Second run (build_results) ~13.0 min

-- 
-- Rene Rivera
-- Grafik - Don't Assume Anything
-- Robot Dreams - http://robot-dreams.net
-- rrivera/acm.org (msn) - grafikrobot/aim,yahoo,skype,efnet,gmail