[jira] [Created] (RAT-211) Generated rat-output.xml must be well-formed, even if BinaryGuesser fails

["Konstantin Kolinko (JIRA)" <ji...@apache.org>]

Konstantin Kolinko created RAT-211:
--------------------------------------

             Summary: Generated rat-output.xml must be well-formed, even if BinaryGuesser fails
                 Key: RAT-211
                 URL: https://issues.apache.org/jira/browse/RAT-211
             Project: Apache Rat
          Issue Type: Bug
            Reporter: Konstantin Kolinko


This issue was originally reported by Infrastructure team while running RAT over Apache Tomcat source code, see thread
"Files to exclude from buildbot rat tests" (started 2016-02-15) at dev "at" tomcat.apache.org mailing list. (1)

The issue:
===========
1. Buildbot at ASF is configured to run RAT tool over tomcat-trunk, tomcat-8, tomcat-7 source code.

2. Tomcat has *.bmp, *.dia files in its source code (images used by Windows installer, diagrams in documentation) that RAT failed to recognize as binary.

3. RAT generated rat-output.xml file that included header-sample fragments of those *.bmp and *.dia files. Those fragments are actually binary garbage.  The result is that a broken XML file was generated.

4. XSLT transformation from rat-output.xml into rat-output.html failed.

I have not seen the actual error printed by XSLT processor, but I confirmed that the file is broken by downloading rat-output.xml and opening it in Firefox. Firefox reported a syntax error.

Workaround:
===========
rat-excludes.txt file in Tomcat source code was updated to exclude
**/*.bmp
**/*.dia

References:
===========
1. "Files to exclude from buildbot rat tests" (started 2016-02-15) at dev "at" tomcat.apache.org mailing list.
http://markmail.org/message/rhrm54ch5omjalt4

2. Apache Tomcat links to Buildbot resuls:
http://tomcat.apache.org/ci.html#Buildbot

3. Apache Tomcat source code
http://tomcat.apache.org/svn.html

Notes:
- RAT excludes files in Tomcat source code are at
res/rat/rat-excludes.txt

- I know that Buildbot uses Ant to run RAT. The Ant project file for that is not in Tomcat sources, but in Infrastructure configuration (I do not have a link). It can be seen in "shell_5 RAT Report Complete" step during build run. E.g. here:
https://ci.apache.org/builders/tomcat-trunk/builds/1061

- I do not know what version of RAT is used by that build slave on Buildbot.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)