Community
Participate
Working Groups
Build output ErrorParsers usnig ErrorPattern might cause performance issues. The matching of the pattern is done via matcher.find() This will try to apply the pattern on any position on the built output line and is very unperformant compared to matcher.matches() There might be regular expressions that need to be reworked, but what I found most of the regex could be applied using matcher.matches()
Do you intend to develop a patch and submit it into gerrit? If you do run unit tests. Error parsers are covered by unit tests pretty well so if the tests pass you can be confident
CQ:WIND00-WB4-3624 Just wondering, how is this problem different than what we got fixed with https://bugs.eclipse.org/bugs/show_bug.cgi?id=419207 ?
The difference is that the matcher.find pattern in but 419207 was matched and afterwards a check whether for a match of the whole line was done. e.g. matcher.find() && matcher.group(0).length()==line.length() this is 100% equal to matcher.matches(). In this report the matcher.find() isn't checked for the whole line. So depending on the regular expression used it might be the case that it matches a substring, if this is reworked to use matcher.matches() the regular expressions will need some double checking.