pmd-cpd report about jsinspect HOT 18 CLOSED

Do you have any documentation on the expected XML format? And are tokens mandatory? Because jsinspect specifically avoids considering tokens as part of its parsing.

from jsinspect.

I don't know of any documentation. The best I can give you is the original pmd-cpd source code (http://sourceforge.net/p/pmd/code/ci/master/tree/pmd-core/src/main/java/net/sourceforge/pmd/cpd/XMLRenderer.java) and the jenkins reader (https://github.com/jenkinsci/dry-plugin/blob/master/src/main/java/hudson/plugins/dry/parser/DuplicateCode.java).

Looking at my existing cpd jenkins reports (generated from phpcpd), it doesn't look like the jenkins parser cares about the value of the tokens field. It uses the lines to determine the severity of the violation.

from jsinspect.

Great, thanks! Will look into it next week.

from jsinspect.

@trichins Do you have any example output you could provide? Just curious how it usually displays "codefragments" - assuming they're not identical, would a diff be displayed?

from jsinspect.

All the output that I've seen just shows a code snippet. I think it is based around the idea that the code will be identical between the two sections, otherwise it wouldn't be a copy/paste violation. But I don't think it actually cares what is put in there. I use it to help me track down where the violation is so a diff would work just as well.

from jsinspect.

Hi @trichins, I've committed something that could work at: 0ab568b

Could you confirm whether or not this output is compatible with your setup?

<?xml version="1.0" encoding="utf-8"?>
<pmd-cpd>
<duplication>
<file path="spec/fixtures/intersection.js" line="1"/>
<file path="spec/fixtures/intersection.js" line="7"/>
<file path="spec/fixtures/indentation.js" line="1"/>
<file path="spec/fixtures/indentation.js" line="7"/>
<codefragment>
- spec/fixtures/intersection.js:1,5
+ spec/fixtures/intersection.js:7,11

+  function intersectionB(arrayA, arrayB) {
+    arrayA.filter(function(n) {
+      return arrayB.indexOf(n) != -1;
-  function intersectionA(array1, array2) {
-    array1.filter(function(n) {
-      return array2.indexOf(n) != -1;
     });
   }

- spec/fixtures/intersection.js:1,5
+ spec/fixtures/indentation.js:1,5

   function intersectionA(array1, array2) {
     array1.filter(function(n) {
       return array2.indexOf(n) != -1;
     });
   }

- spec/fixtures/intersection.js:1,5
+ spec/fixtures/indentation.js:7,11

   function intersectionA(array1, array2) {
     array1.filter(function(n) {
+      return array2.indexOf(n) == -1;
-      return array2.indexOf(n) != -1;
     });
   }
</codefragment>
</duplication>
</pmd-cpd>

Of note, the <duplication> element is missing the tokens attribute, since we don't count them, as well as lines, as it may vary between files.

from jsinspect.

Also, you can generate the test output above by running the following from master:
./bin/jsinspect -r pmd spec/fixtures/intersection.js spec/fixtures/indentation.js

from jsinspect.

Closing for now. Let me know if you run into any issues with this format!
Thanks.

from jsinspect.

Sorry it's taken me such a long time to get back to you. I just tested it and it works pretty well. Two issues that it has is:

1 - the <duplication> node needs at least a "lines" attribute. It is used by jenkins to determine the severity of the duplication. The value is the number of duplicate lines.

2 - The path attribute in the <file> node should be an absolute path. I'm seeing two diffferent paths being written, one that starts with "./" and one that doesn't. Both are relative. I've worked around this by running sed on the output.

from jsinspect.

Hi @trichins thanks for the reply. Lines doesn't make much sense in this case, but I suppose I could report the max number of lines between the different instances. As for the paths, what are they used for? Just linking to the files?

from jsinspect.

max number of lines will work.

jenkins tries to find the file and display it with the lines highlighted when you click on the report. Unfortunately, the finding part happens during the initial parsing of the xml output and if can't find the file, it fails the entire job.

from jsinspect.

:( That's unfortunate. Expect fixes for both tonight or tomorrow. Thanks again!

from jsinspect.

@trichins Mind taking a look at the latest release when you get the chance? I decided to have lines be the total num of lines spanned by each parent node in a "duplication". And while I added absolute paths for the file elements, I kept relative paths in the diffs, as they're generally cleaner to read.

./bin/jsinspect -r pmd spec/fixtures/intersection.js spec/fixtures/indentation.js

<?xml version="1.0" encoding="utf-8"?>
<pmd-cpd>
<duplication lines="20">
<file path="/Users/danielstjules/GitHub/jsinspect/spec/fixtures/intersection.js" line="1"/>
<file path="/Users/danielstjules/GitHub/jsinspect/spec/fixtures/intersection.js" line="7"/>
<file path="/Users/danielstjules/GitHub/jsinspect/spec/fixtures/indentation.js" line="1"/>
<file path="/Users/danielstjules/GitHub/jsinspect/spec/fixtures/indentation.js" line="7"/>
<codefragment>
- spec/fixtures/intersection.js:1,5
+ spec/fixtures/intersection.js:7,11

+  function intersectionB(arrayA, arrayB) {
+    arrayA.filter(function(n) {
+      return arrayB.indexOf(n) != -1;
-  function intersectionA(array1, array2) {
-    array1.filter(function(n) {
-      return array2.indexOf(n) != -1;
     });
   }

- spec/fixtures/intersection.js:1,5
+ spec/fixtures/indentation.js:1,5

   function intersectionA(array1, array2) {
     array1.filter(function(n) {
       return array2.indexOf(n) != -1;
     });
   }

- spec/fixtures/intersection.js:1,5
+ spec/fixtures/indentation.js:7,11

   function intersectionA(array1, array2) {
     array1.filter(function(n) {
+      return array2.indexOf(n) == -1;
-      return array2.indexOf(n) != -1;
     });
   }
</codefragment>
</duplication>
</pmd-cpd>

from jsinspect.

The paths are working.

As for the lines count, I'd prefer if it was the max line count of each parent node. But since this is just a hint to jenkins on which one is the worst offender, a sum kind of works as well.

from jsinspect.

Agreed. 10 lines of code copy-pasted 100 times should be ranked more negatively than a single 30-line copy-paste. Glad things are working. :)

from jsinspect.

@danielstjules It seems that the <file ... path is coming out as absolute, while the paths shown inside the <codefragment> are relative.
The examples above are using relative for both.

Would it be possible to use only relative paths in both cases?

stefanjudis/grunt-jsinspect#21

from jsinspect.

jenkins tries to find the file and display it with the lines highlighted when you click on the report.

Absolute paths are required in <file ...> for linking to the respective from the Jenkins report. And I opted to use relative paths within the code fragments as I find them a bit more legible :)

from jsinspect.

Ok, thank you for the clarification

from jsinspect.