Status: Resolved (View Workflow)
Hudson 1.366, Warnings plugin 3.7, Doxygen 1.7.1
In Doxygen 1.7.1 the case of the warnings changed (Error -> error, Warning -> warning). The doxygen parser in the warnings plugin is case sensitive so the warnings/errors are not logged.
The reason for the change is in doxygen issue 621695 (https://bugzilla.gnome.org/show_bug.cgi?id=621695)
What do you expect me to do? Create a small Java console application and copy pasting the regexp code and matching our doxygen lines? Or something else?
That depends on much time you have
The optimal solution would be to check out the source code and try to improve the regular expressions in the current DoxygenParser class. Do you already have some Java experience? An option would be to split the parser into two parts: the first one scans single line warnings and the second one multi-line ones.
If that is not possible, you can attach at least some (raw) warning logs that are typical. Then I can uses these to improve the parsers...
Unfortunately I was not notified that you had answered my last question. In the meantime, I have found a good workaround for this bug. Doxygen allows you to designate a file so that the doxygen output will be written there. But the normal stderr also works when you specify this file. These are the details:
- The WARN_LOGFILE tag can be used to specify a file to which warning
- and error messages should be written. If left blank the output is written
- to stderr.
WARN_LOGFILE = doxygen.output
Then in the warnings plugin instead of having the MSBuild and doxygen parsers for the console you just specify the MSBuild parser.
In the "Scan workspace files" section of the Warnings plugin configuration, you specify the doxygen.output file with the doxygen parser.
It works like a charm. And you don't need to modify the regular expression to take into account the "[exec]" stuff that running doxygen through NAnt adds.
Seems that there still is some work to do.