Re: [Boost-testing] Frequency/timing of regression summary generation?

From: Aleksey Gurtovoy (agurtovoy_at_[hidden])
Date: 2007-11-10 21:57:51

Rene Rivera wrote:
> Just FYIs...
> 
> Aleksey Gurtovoy wrote:
>> Report generation times are greatly affected by:
>>
>> 1) the overall number of failing test cases (an output page is generated 
>> for each of these).
> 
> Which step in the process are those generated?

'make_links_action' in 'boost_wide_report.py'.

[..]

>> -- not to mention general factors such as processor load.
> 
> Report generation is all this machine does currently. But what seems to 
> be taking most of the CPU time is:
> 
> ===
> Writing document summary_.html
> #     Making summary user  report...
> #     xsltproc -o 
> "/home/grafik/devroots/Boost/boost-regression/trunk/all/user/summary.html" 
> --param run_date "'2007-11-08T02:33:52Z'"  --param explicit_markup_file 
> "'/home/grafik/devroots/Boost/boost-regression/boost/status/explicit-failures-markup.xml'" 
>   --param source "'trunk'"  --param mode "'user'" 
> "/home/grafik/devroots/Boost/boost-regression/boost/tools/regression/xsl_reports/xsl/v2/summary_page.xsl" 
>  
> "/home/grafik/devroots/Boost/boost-regression/trunk/all/extended_test_results.xml"
> Writing document summary_.html
> ===
> 
> It's up to 80 CPU minutes just on that, and onwards.
> 

I'm afraid that's about what we can expect with the current reporting 
scripts and number of toolsets (45?) we test against on the trunk.

-- 
Aleksey Gurtovoy
MetaCommunications Engineering