Upgrade to Checkstyle 5.5

Hi Julien,

I'm able to apply your patch with some modifications. I didn't make a lot of tests, but I definitely can reproduce one bug ( http://sourceforge.net/tracker/?func=detail&aid=3132645&group_id=29721&atid=397078 ). I think that we must wait with upgrade to the next version of checkstyle, even if workaround can be straightforward - just add an empty line to the end of file.

I agree this can be a major regression for people that don't have an empty line at the end of their files.
So let's wait for a fix of checkstyle.

Do you know if the bug has been introduced in checkstyle 5.2 or 5.3 ?

I can reproduce it with 5.2 too.

Ok, so I temporarily remove this issue from release 2.6.

I reproduce the issue with checkstyle 5.2 so it was certainly introduced in a previous version.

>java -jar checkstyle-5.2-all.jar -c sun_checks.xml Test.java
Starting audit...
[...]\Test.java:0: Il manque un caractÞre NewLine Ó la fin du fichier
[...]\Test.java:0: Missing package-info.java file.
[...]\Test.java:1:17: Exception levÚe : unexpected char: 0xFFFF
Audit done.

Also note that Checkstyle from 5.1 to 5.3 has a memory leak (see SONAR-2511).

For information : memory leak was fixed in checkstyle 5.3 - cad2022770b8

For information : parsing exception (unexpected char: 0xFFFF) was fixed in checkstyle 5.4-SNAPSHOT - 90eadf9cf0b2

Could we directly upgrade to Checkstyle 5.4 I need this change

• Enhanced MagicNumber to support ignoring numbers in annotation declarations.

We were waiting for the availability in maven central repository. It should be possible now. We have to check that the memory leak and the parsing exception are fixed.

Checkstyle available on central: http://repo1.maven.org/maven2/com/puppycrawl/tools/checkstyle/5.4/

Is there anybody working on it? Would you like help?

Any help is welcome Julien . The most important thing is too list all rule changes between version 5.1 and version 5.4.

I already did the work for 5.1 to 5.3. Just need to merge my old patch with i18n. I'm working on a new patch.

See my pull request. Warning: not very well tested so may be typos. I'll work on french plugin update.

Nothing to do in french plugin. Translation of checkstyle is not done yet

Guys, looks like there was a typo in my previous comment - parsing exception (unexpected char: 0xFFFF) was not fixed in 5.4, but in 5.5, which is not yet released. Thus and having in consideration SONAR-2585 it might make sense to wait a bit more before upgrade of checkstyle.

Checkstyle 5.5 is out, any updates on this issue?

Thomas, please be just a bit patient - we know about release as we participated in it. This issue is scheduled for 2.13 and we work on it.

Upgraded in 3243252e10.

Looks like we're waiting for 2.13 - but just to note as I don't see it mentioned in the notes and somebody else might benefit from this:

Using try-with-resources in Java 7 seems to trigger Checkstyle "Inner Assignment" on every single use.

For the time being, I've disabled this check. I'm hoping that Checkstyle 5.5 with official Java 7 support won't have this problem?

Hi Mark,

Thanks for this info. However I'm not sure that Checkstyle 5.5 solves this issue, so it would be better to check this.

Hi Evgeny:

You are correct. Checkstyle 5.5 continues to give these warnings. I guess I will take this up with them...

Manually tested Evgeny, but could you list in this JIRA ticket all the changes relating to this migration (mainly new rules)

Freddy, I'm not sure that this is required since Checkstyle provides nice list of changes (in comparison with PMD) - http://checkstyle.sourceforge.net/releasenotes.html
However here is a list of changed rules :

• New rules from 5.2:
  • InnerTypeLast
• Modified rules from 5.2:
  • DeclarationOrder
  • ImportOrder
• New rules from 5.3:
  • OuterTypeFilename
  • NestedForDepth
  • MethodCount
  • OneStatementPerLine
• Modified rules from 5.3:
  • PackageDeclaration
  • MagicNumber
  • AbstractClassName
  • AvoidStarImport
• Modified rules from 5.4:
  • UnusedImports
  • MagicNumber
  • EqualsAvoidNull
  • IllegalThrows
  • RedundantModifier

Thanks Evgeny, Manually tested

Upcoming IT with a checkstyle-ruling