Re: [Boost-testing] [boost] What's up with the trunk test summary?

Subject: Re: [Boost-testing] [boost] What's up with the trunk test summary?
From: Rene Rivera (grafikrobot_at_[hidden])
Date: 2008-09-07 17:29:40

K. Noel Belcourt wrote:
> 
> On Sep 7, 2008, at 2:12 PM, Rene Rivera wrote:
> 
>> K. Noel Belcourt wrote:
>>>
>>> On Sep 7, 2008, at 1:12 PM, Rene Rivera wrote:
>>>
>>>> K. Noel Belcourt wrote:
>>>>> What can I do on this end to help?
>>>>
>>>> In the interim, I see that I'm uploading results beta.boost.org okay.
>>> But I'm unable to figure out where the developer and output directories
>>> are unzipped and live on that system. It seems that they should be under
>>> /home/www/live.boost.org, but I don't see any results there.
>>> Specifically, I'm trying to find the trunk summary.html, toc.html, or
>>> any of the Sandia results?  Where do they live?
>>
>> They don't live anywhere... The pages are read directly from the ZIP
>> archive.
> 
> Well then let me explain what concerns me.
> 
> When I browse the release results, there are html files for each 
> library, graph.html, date_time.html, etc... but my trunk directory 
> doesn't contain any of these files, none.  Seems to me that the 
> library.html files should be generated by build_results.sh, but I need 
> some help understanding where the code lives that should generate these 
> files, since it seems to be failing.

OK, the html files should be generated to results-root/trunk/all/*. The 
build_results.sh doesn't directly generate any html. The work is done 
with the python script and the xslt transformations. And at minimum 
there are files missing from the ZIP that I get when the results are 
generated. For example I have a "$/trunk/all/developer/toc.html" which 
corresponds to the left side navigation which is missing from the ZIP.

To outline the process:

a) build_results.sh gets the current Boost trunk to have access to the 
rest of the tools

b) b_r.sh calls boost_wide_report.py with the appropriate options based 
on the simple branch tag.

c) b_w_r.py checks for and downloads any new test result ZIP archives 
into "$/trunk/incoming/trunk"

d) b_w_r.py uncompresses the test archives into 
"$/trunk/incoming/trunk/processed"

e) b_w_r.py runs an initial xsltproc processing step that converts the 
tests to the "$/trunk/incoming/trunk/processed/merged" dir.

f) b_w_r.py runs a bunch more xsltproc scripts that generate the thml 
files for the various kinds of reports it can generate. How many of 
these it runs are selectable, and the b_r.sh script indicates them based 
on the branch it's running reports for. It's what this case 
<http://svn.boost.org/trac/boost/browser/trunk/tools/regression/xsl_reports/build_results.sh#L95> 
is doing.

c) b_w.sh creates the ZIP archive of the html files generated

d) b_w.sh uploads and makes live the new ZIP archive

Hopefully that's enough to navigate the output dirs. I can probably 
exampling parts of the python script. But I can't explain any of the 
xslt stuff... It's a terribly confusing grey box to me.

-- 
-- Grafik - Don't Assume Anything
-- Redshift Software, Inc. - http://redshift-software.com
-- rrivera/acm.org (msn) - grafik/redshift-software.com
-- 102708583/icq - grafikrobot/aim,yahoo,skype,efnet,gmail