[
https://issues.apache.org/jira/browse/LUCENE-6732?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Uwe Schindler updated LUCENE-6732:
----------------------------------
Description:
Today I enabled warnings analysis on Policeman Jenkins. This scans the build
log for warnings by javac and reports them in statistics, together with source
file dumps.
When doing that I found out that someone added again a lot of "invalid" license
headers using {{/\*\*}} instead a simple comment. This causes javadocs warnings
under some circumstances, because {{/\*\*}} is start of javadocs and not a
license comment.
I then tried to fix the validate-source-patterns to detect this, but due to a
bug in ANT, the {{<containsregexp/>}} filter is applied per line (although it
has multiline matching capabilities!!!).
So I rewrote our checker to run with groovy. This also has some good parts:
- it tells you wwhat was broken, otherwise you just know there is an error, but
not whats wrong (tab, nocommit,...)
- its much faster (multiple {{<containsregexp/>}} read file over and over, this
one reads file one time into a string and then applies all regular expressions).
was:
Today I enabled warnings analysis on Policeman Jenkins. This scans the build
log for warnings by javac and reports them in statistics, together with source
file dumps.
When doing that I found out that someone added again a lot of "invalid" license
headers using {{/**}} instead a simple comment. This causes javadocs warnings
under some circumstances, because {{/**}} is start of javadocs and not a
license comment.
I then tried to fix the validate-source-patterns to detect this, but due to a
bug in ANT, the {{<containsregexp/>}} filter is applied per line (although it
has multiline matching capabilities!!!).
So I rewrote our checker to run with groovy. This also has some good parts:
- it tells you wwhat was broken, otherwise you just know there is an error, but
not whats wrong (tab, nocommit,...)
- its much faster (multiple {{<containsregexp/>}} read file over and over, this
one reads file one time into a string and then applies all regular expressions).
> Improve validate-source-patterns in build.xml (e.g., detect invalid license
> headers!!)
> --------------------------------------------------------------------------------------
>
> Key: LUCENE-6732
> URL: https://issues.apache.org/jira/browse/LUCENE-6732
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Uwe Schindler
> Assignee: Uwe Schindler
>
> Today I enabled warnings analysis on Policeman Jenkins. This scans the build
> log for warnings by javac and reports them in statistics, together with
> source file dumps.
> When doing that I found out that someone added again a lot of "invalid"
> license headers using {{/\*\*}} instead a simple comment. This causes
> javadocs warnings under some circumstances, because {{/\*\*}} is start of
> javadocs and not a license comment.
> I then tried to fix the validate-source-patterns to detect this, but due to a
> bug in ANT, the {{<containsregexp/>}} filter is applied per line (although it
> has multiline matching capabilities!!!).
> So I rewrote our checker to run with groovy. This also has some good parts:
> - it tells you wwhat was broken, otherwise you just know there is an error,
> but not whats wrong (tab, nocommit,...)
> - its much faster (multiple {{<containsregexp/>}} read file over and over,
> this one reads file one time into a string and then applies all regular
> expressions).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
