curcuru 02/01/18 11:35:23
Modified: test/java/xdocs/sources xalantest.xml
Added: test/java/xdocs/sources/tests faq.xml
Log:
Add new FAQ for testing docs; in progress, needs editing!
Revision Changes Path
1.4 +4 -0 xml-xalan/test/java/xdocs/sources/xalantest.xml
Index: xalantest.xml
===================================================================
RCS file: /home/cvs/xml-xalan/test/java/xdocs/sources/xalantest.xml,v
retrieving revision 1.3
retrieving revision 1.4
diff -u -r1.3 -r1.4
--- xalantest.xml 18 Dec 2001 23:12:38 -0000 1.3
+++ xalantest.xml 18 Jan 2002 19:35:23 -0000 1.4
@@ -19,6 +19,10 @@
<separator/>
+ <document id="faq"
+ label="FAQ"
+ source="tests/faq.xml"/>
+
<document id="run"
label="Running Tests"
source="tests/run.xml"/>
1.1 xml-xalan/test/java/xdocs/sources/tests/faq.xml
Index: faq.xml
===================================================================
<?xml version="1.0" standalone="no"?>
<!DOCTYPE s1 SYSTEM "sbk:/style/dtd/document.dtd">
<s1 title="Frequently asked questions">
<s2 title="In progress">
<p>We need to start a FAQ for testing Xalan, unfortunately the faq
doctype
doesn't properly work in this tree, so I'm using an s1/s2 doc instead
temporarily.</p>
<ul>
<li><link anchor="viewresults">Pretty-printing results</link></li>
</ul>
</s2>
<s2 title="Prepare to run tests">
<p>CVSROOT=:pserver:<em>user</em>@cvs.apache.org:/home/cvs etc.<br/>
or CVSROOT=:pserver:[EMAIL PROTECTED]:/home/cvspublic etc.<br/>
cvs checkout xml-xalan/java<br/>
cvs checkout xml-xalan/test<br/>
cd xml-xalan/java<br/>
build jar<br/>
cd ../test<br/>
build jar<br/>
build -projecthelp - to get a list of targets<br/>
</p>
</s2>
<s2 title="Run conformance tests">
<p>The conf set of tests can be run in a variety of ways, especially with
each different flavor.
For a list of flavors, see
xml-xalan/test/java/src/org/apache/qetest/xslwrapper/TransformWrapperFactory.properties</p>
<ul>
<li><source>build conf</source> (default: trax.systemId)<br/><br/></li>
<li><source>build conf -Dconf.flavor=trax.sax</source> (uses SAX
IMPORTANT!)<br/><br/></li>
<li><source>build conf -Dconf.flavor=trax.dom</source> (uses dom
IMPORTANT!)<br/><br/></li>
<li>trax.file - File object instead of systemId<br/><br/></li>
<li>trax.stream - InputStreams - IMPORTANT!<br/><br/></li>
<li>trax.localPath - uses local paths instead of URLs,
experimental<br/><br/></li>
<li>trax.systemId3 - does systemId transform three times in a row,
experimental<br/><br/></li>
<li>process - uses command line class - IMPORTANT!<br/><br/></li>
</ul>
</s2>
<s2 title="Run contrib/perf tests">
<p>Same as running conf tests, except substitute 'contrib' for 'conf'
everywhere,
or 'perf' instead. Note that 'perf' uses a custom Testlet testing
algorithim
that iterates several times and outputs custom perf elements.</p>
<ul>
<li><source>build contrib
-Dcontrib.flavor=trax.sax</source><br/><br/></li>
<li><source>build perf -Dperf.flavor=trax.stream</source><br/><br/></li>
</ul>
</s2>
<anchor name="xsltmark"/>
<s2 title="Run XSLTMARK perf tests">
<p>Xalan has it's own more detailed automation harness for running
stylesheet tests and capturing performance metrics. You can run the
XSLTMARK
set of stylesheets using Xalan's harness to see more detailed performance
data.</p>
<ul>
<li>Copy xsltmark/testcases to xml-xalan/test (so it's
xml-xalan/test/testcases)<br/><br/></li>
<li>Copy xsltmark/testcases/default.conf to
xml-xalan/test/xsltmark.filelist<br/><br/></li>
<li>cd xml-xalan/test<br/><br/></li>
<li>Search-and-replace xsltmark.filelist 't='
't=testcases/'<br/><br/></li>
<li>Search-and-replace xsltmark.filelist 'e='
'e=testcases/'<br/><br/></li>
<li><source>build perf
-Dperf.fileList=xsltmark.filelist</source><br/><br/></li>
<li><source>java -classpath blah org.apache.xalan.xslt.Process -in
results-perf/results.xml -xsl PerfScanner.xsl -out
results-perf/PerfReport.html</source><br/><br/></li>
<li>(where blah includes
xml-apis.jar;xalan.jar;xercesImpl.jar)<br/><br/></li>
</ul>
</s2>
<anchor name="viewresults"/>
<s2 title="View results in HTML">
<p>Every test creates a results.xml (or TestName.xml) file of all the
results
the test has performed - no need to examine the console, since all the
info will
be here. We have a couple of prototype stylesheets to view the results
in a
semi-pretty HTML style.</p>
<ul>
<li><source>build perf</source><br/>
(Creates results-perf/results.xml)<br/></li>
<li><source>java -classpath blah org.apache.xalan.xslt.Process -in
results-perf/results.xml -xsl FailScanner.xsl -out
results-perf/FailReport.html</source><br/><br/></li>
<li><source>java -classpath blah org.apache.xalan.xslt.Process -in
results-perf/results.xml -xsl PerfScanner.xsl -out
results-perf/PerfReport.html</source><br/><br/></li>
<li><source>build alltest</source><br/>
(Creates a <b>LOT</b> of results in results-alltest/**)<br/></li>
<li><source>java -classpath blah;testxsl.jar
org.apache.qetest.xsl.ResultScanner results-alltest</source><br/>
(This uses ResultScanner.xsl to style <b>all</b> results in the whole
tree under
results-alltest into a single ResultReport.html in the current directory;
it currently
uses FailScanner.xsl to only include fail results)<br/></li>
</ul>
</s2>
<s2 title="Include/Exclude tests">
<p>Run just a subset of tests, or exclude tests using simple command line
options.</p>
<ul>
<li><source>build conf -Dconf.category=axes;boolean</source> - Only run
those directories<br/><br/></li>
<li><source>build conf -Dconf.excludes=axes107.xsl;boolean12.xsl</source>
- Skip those explicit xsl filenames<br/><br/></li>
</ul>
</s2>
<anchor name="conf.one"/>
<s2 title="Run a single conf test">
<ul>
<li><source>build conf.one -Dconf.test=axes44
-Dconf.flavor=trax.stream</source><br/>
This will run just the axes44.xsl test, using the normal Testlet
algorithim <b>and</b>
using whatever flavor you choose (which makes it easy to see if single
tests
run properly using SAX, DOM, streams, whatever)<br/></li>
</ul>
</s2>
<s2 title="Run all API tests">
<ul>
<li><source>build api -DtestClass=TransformerAPITest</source><br/></li>
<li>This runs all the available specific API tests, which includes all
the
API tests in the smoketest, plus other tests that fail due to known
bugzilla reports.<br/></li>
<li>Note that in the case of <source>testClass</source>, you do
<b>not</b>
prefix the name of the option with conf,api,perf,etc.<br/><br/></li>
</ul>
</s2>
<s2 title="Run a LOT tests">
<ul>
<li><source>build alltest</source> (runs
all,alltest.other,alltest.conf,alltest.contrib)<br/></li>
<li><source>build alltest.other</source> runs all API tests, extensions,
bugzilla, threading, and perf tests<br/></li>
<li><source>build alltest.conf</source> run all flavors of conf
tests<br/></li>
<li><source>build alltest.contrib</source> run all flavors of contrib
tests<br/></li>
<li><source>build alltest.features</source> run all conf and contrib
tests,
matrixed over all flavors, twice: once with the feature incremental set
to true,
once with optimize set to false (note: not all features matter with all
flavors,
but we run them all anyway. Several meg of output!)<br/></li>
</ul>
</s2>
</s1>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
