[core] When runing any RuleTst , start/end methods not called

A source code analyzer

Brought to you by: adangel, juansotuyo

#1506 [core] When runing any RuleTst , start/end methods not called

Milestone: PMD-5.3.8

Status: closed

Owner: Andreas Dangel

Labels: core (10)

Module: PMD

Priority: 2-Critical

Type: Bug

Affects version:

Ruleset / Rule: All

Updated: 2016-08-07

Created: 2016-07-16

Creator: blundell

Private: No

When runing any RuleTst , the overrideable methods for each rule are not called.

Look at this screenshot, it should print out start and end for the rule, as you can see it prints nothing but visit is printed. This means any rule that override start and end and use them to update report will not be tested correctly.

1 Attachments

Screen Shot 2016-07-16 at 12.32.17 pm.png

Related

Issues: ~~#1511~~

Discussion

blundell - 2016-07-16

Have I understood this correctly?
I would have expected this in the console:

start
visit
visit
visit
end

Last edit: blundell 2016-07-16

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

blundell - 2016-07-16

It looks like SourceCodeProcessor.java never calls start() or end(), but MonoThreadProcessor & MultiThreadProcessor do?

SourceCodeProcessor.java could be "fixed" at this line:

https://github.com/pmd/pmd/blob/master/pmd-core/src/main/java/net/sourceforge/pmd/SourceCodeProcessor.java#L146

like so

ruleSets.start(ctx); ruleSets.apply(acus, ctx, language); ruleSets.end(ctx);

Can make a PullRequest if you agree?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Andreas Dangel - 2016-08-03
  
  Your expectation about start/end is correct.
  
  As you have seen, MonoThreadProcessor and MultiThreadProcessor call correclty start/end - and then they call the SourceCodeProcessor. That's why the SourceCodeProcessor doesn't do this anymore - and it shouldn't.
  
  So, we have a problem with the test framework.
  
  Looking at the code, there is a difference in when/how often start&end is called: MonoThreadProcessor calls it for each file, while the MultiThreadProcessor calls it only once - and for a ruleset instance, that is never used, as for the threads, new rule instances will be created.
  So, I suspect, this doesn't really work yet....
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Andreas Dangel - 2016-08-06

labels: --> core

summary: When runing any RuleTst , methods not called --> [core] When runing any RuleTst , start/stop methods not called

status: open --> in-progress

assigned_to: Andreas Dangel
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Andreas Dangel - 2016-08-06

summary: [core] When runing any RuleTst , start/stop methods not called --> [core] When runing any RuleTst , start/end methods not called
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Andreas Dangel - 2016-08-06

Related issue [#974]

Related

Issues: ~~#974~~

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Andreas Dangel - 2016-08-07

status: in-progress --> closed

Milestone: New Tickets --> PMD-5.3.8
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Andreas Dangel - 2016-08-07

I've created a follow-up issue: [#1511] in order to make the implementation consistent.

The bug within the test framework is fixed, so that RuleTst now calls start and end.

The bugfix will be available in PMD 5.3.8, 5.4.3, 5.5.2 and later.

Commit: https://github.com/pmd/pmd/commit/cf4938a65b26fdc039aafbf3b5da77baf2ca61e2

Related

Issues: ~~#1511~~

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Log in to post a comment.