|
Boost Testing : |
Subject: Re: [Boost-testing] trunk tests failing
From: Belcourt, Kenneth (kbelco_at_[hidden])
Date: 2009-08-08 16:05:09
On Aug 8, 2009, at 1:52 PM, Belcourt, Kenneth wrote:
> # /incoming/trunk/processed/merged/Debian-Sid.xml: run
> # target doesn't exists, building
> # Merging "/incoming/trunk/processed/Debian-Sid.xml" with
> expected results...
> # xsltproc -o "/sierra/Src/kbelco/boost/reporting/trunk/
> incoming/trunk/processed/merged/Debian-Sid.xml" --param source
> "'trunk'" --param expected_results_file "'/sierra/Src/kbelco/boost/
> reporting/boost/status/expected_results.xml'" --param
> failures_markup_file "'/sierra/Src/kbelco/boost/reporting/boost/
> status/
> explicit-failures-markup.xml'" "/sierra/Src/kbelco/boost/reporting/
> boost/tools/regression/xsl_reports/xsl/v2/add_expected_results.xsl"
> "/
> sierra/Src/kbelco/boost/reporting/trunk/incoming/trunk/processed/
> Debian-Sid-trimmed.xml"
>
> Unknown test type "run_mpi"
>
> Any ideas?
Okay, looks like someone added support for running / reporting MPI,
but didn't update all the files used by the report processing. Not
sure who updated this but either we should revert the change, as trunk
tests haven't run since Wednesday (and remove the Debian-Sid results),
or fix this issue. I won't be able to look at it until Sunday evening
at the earliest.
It seems like these files are missing references to run_mpi.
./regression/xsl_reports/test/generate_test_results.py: types =
[ "compile", "compile_fail", "link", "link_fail", "run", "run_fail",
"run_pyd" ]
./regression/xsl_reports/xsl/v2/
add_expected_results.xsl: <xsl:when test="$type='run'
or $type='run_fail' or $type='run_pyd'">
./regression/xsl_reports/xsl/v2/result_page.xsl: <xsl:when
test="$test_type='run_fail'"> <xsl:text>rf</xsl:text> </xsl:when>
./regression/xsl_reports/xsl/result_page.xsl: <xsl:when
test="$test_type='run_fail'"> <xsl:text>rf</xsl:text> </xsl:when>
./regression/xsl_reports/test_results.xsd: <xs:enumeration
value="run_fail"/>
Because I don't update the Boost installation in the test reporting
directory, if someone fixes this problem, please let me know so I can
update. Trunk testing will remain broken until we revert or fix this
issue, sorry.
-- Noel