Re: [boost] Should regression summary consider both pass and warn as passing?

From: Rene Rivera (grafik666_at_[hidden])
Date: 2003-05-26 10:16:30

[2003-05-26] Reece Dunn wrote:

>It should also be possible, if the regression engine outputs an XML table
of 
>information, to support all of these different versions via different XSLT 
>code (simplified via a parameter to the stylesheet).

The regression test don't currently output an easy to use XML... Beman has
it on his todo list. At which point we plan to replace the current Perl
script with a C++ cgi to generate the summary.

>e.g.
><compiler name = "GNU C++" version = "3.2.2">
>   <library name = "Interval Library">
>      <results file = "..." type = "pass|warn|fail"/>
>   </library>
></compiler>
>
>It is then possible to get the XSLT code to generate statistics (number of 
>tests, number passed, failed and generated warnings, percentages). It is 
>also possible to do this for each compiler and compiler/library pair. It
may 
>also be possible to compare a library across different compilers.
>
>This would have the benefit of being able to select how the results are 
>displayed (possibly via a form, transforming the XML data on the server, 
>then sending it back to the client).

Seems nice. But I'm not sure running XSLT transforms on the fly in the
SourceForge web servers is such a hot idea ;-) I don't even know if it's
possible. Re, if they even have the needed packages loaded into the web
server machines.

>SIDE NOTE: I have ASP code that uses the Microsoft XML parser to transform 
>an XML document using a specified XSLT file, passing the remaining
arguments 
>sent to the ASP page to the stylesheet as parameters. Let me know if this
is 
>of interest, and I will post it.

The SourceForge web servers are all running Linux, so I don't ASP code would
help in this case.

-- grafik - Don't Assume Anything
-- rrivera_at_[hidden] - grafik_at_[hidden]
-- 102708583_at_icq