Uploaded image for project: 'Jenkins'
  1. Jenkins
  2. JENKINS-56510

Provide generic issue parser for custom tools

    XMLWordPrintable

Details

    • 5.0.0 (analysis-model and warnings-ng)

    Description

      I have a handful of custom tools to generate custom reports on my code base and/or the compiled results. Right now I am abusing the output format of PMD to get close to the output I want. However, it would be nice if I could generate a file that mapped directly to the fields used by the plugin.

      This might also be useful for people who do not have admin access to write groovy parsers. They could do some local analysis, then generate this file format.

      Current Serialization of an Issue instance using XStream:

      <?xml version="1.1" encoding="UTF-8"?>
      <issue>
        <category>category</category>
        <type>type</type>
        <severity>HIGH</severity>
        <message>message</message>
        <lineStart>1</lineStart>
        <lineEnd>2</lineEnd>
        <columnStart>3</columnStart>
        <columnEnd>4</columnEnd>
        <lineRanges>
          <lineRange>
            <start>5</start>
            <end>6</end>
          </lineRange>
        </lineRanges>
        <id>63d61b1f-0cac-4e31-8bb9-f390ed1acfe8</id>
        <description>description</description>
        <reference>reference</reference>
        <origin>origin</origin>
        <moduleName>module-name</moduleName>
        <packageName>package-name</packageName>
        <fileName>file-name</fileName>
        <fingerprint>fingerprint</fingerprint>
      </issue>
      

       
      Current Serialization of a Report instance using XStream:

      <?xml version="1.1" encoding="UTF-8"?>
      <analysisReport>
        <elements class="linked-hash-set">
          <issue>
            <category>category</category>
            <type>type</type>
            <severity>HIGH</severity>
            <message>1</message>
            <lineStart>1</lineStart>
            <lineEnd>2</lineEnd>
            <columnStart>3</columnStart>
            <columnEnd>4</columnEnd>
            <lineRanges>
              <lineRange>
                <start>5</start>
                <end>6</end>
              </lineRange>
            </lineRanges>
            <id>c3b984a3-7f67-4332-953b-27d3943f232c</id>
            <description>description</description>
            <reference>reference</reference>
            <origin>origin</origin>
            <moduleName>module-name</moduleName>
            <packageName>package-name</packageName>
            <fileName>file-name</fileName>
            <fingerprint>fingerprint</fingerprint>
          </issue>
          <issue>
            <category>category</category>
            <type>type</type>
            <severity reference="../../issue/severity"/>
            <message>2</message>
            <lineStart>1</lineStart>
            <lineEnd>2</lineEnd>
            <columnStart>3</columnStart>
            <columnEnd>4</columnEnd>
            <lineRanges>
              <lineRange>
                <start>5</start>
                <end>6</end>
              </lineRange>
            </lineRanges>
            <id>fbf2fee0-292f-4991-bd06-d8c5b13ace93</id>
            <description>description</description>
            <reference>reference</reference>
            <origin>origin</origin>
            <moduleName>module-name</moduleName>
            <packageName>package-name</packageName>
            <fileName>file-name</fileName>
            <fingerprint>fingerprint</fingerprint>
          </issue>
        </elements>
        <infoMessages>
          <string>info</string>
        </infoMessages>
        <errorMessages>
          <string>error</string>
        </errorMessages>
        <duplicatesSize>0</duplicatesSize>
      </analysisReport>
      

       

      Attachments

        Issue Links

          Activity

            drulli Ulli Hafner added a comment - - edited

            This is a good idea! If the parser would be part of warnings-ng (and not analysis-model) then Jenkins XML reading logic (XStream) could be used, see AnalysisResult.java.

            drulli Ulli Hafner added a comment - - edited This is a good idea! If the parser would be part of warnings-ng (and not analysis-model) then Jenkins XML reading logic (XStream) could be used, see AnalysisResult.java .

            That is an interesting idea.

             Could that cause problems since it would be de-serializing XML generated by a different tool?  I don't have any experience with XStream.  Does it handle missing fields well?  I wouldn't expect the external tools to populate all the possible fields of an issue.

            derrickgw Derrick Gibelyou added a comment - That is an interesting idea.  Could that cause problems since it would be de-serializing XML generated by a different tool?  I don't have any experience with XStream.  Does it handle missing fields well?  I wouldn't expect the external tools to populate all the possible fields of an issue.
            drulli Ulli Hafner added a comment -

            Yes, it handles missing fields quite well. And the serialization fallbacks are under our control, see https://github.com/jenkinsci/warnings-ng-plugin/blob/master/src/main/java/io/jenkins/plugins/analysis/core/model/IssueStream.java.

            drulli Ulli Hafner added a comment - Yes, it handles missing fields quite well. And the serialization fallbacks are under our control, see https://github.com/jenkinsci/warnings-ng-plugin/blob/master/src/main/java/io/jenkins/plugins/analysis/core/model/IssueStream.java .
            hobbyte Raphael Furch added a comment -

            drulli  I don't know exactly what to do here. Can you give me a hint please? 

            hobbyte Raphael Furch added a comment - drulli   I don't know exactly what to do here. Can you give me a hint please? 
            drulli Ulli Hafner added a comment - - edited
            1. We need a new tool class e.g. AnalysisReport (subclass of ReportScanningTool) in the warnings-ng project, see Detekt for an example. This class creates a new parser in the createParser method.
            2. We need a new parser class e.g. ReportParser (extending IssueParser). The parse method of this class should use the same code as AnalysisResult.readIssues(). The code in AnalysisResult already reads a report with several issues, so we basically need to reuse or extract that part.
            drulli Ulli Hafner added a comment - - edited We need a new tool class e.g. AnalysisReport (subclass of ReportScanningTool ) in the warnings-ng project, see Detekt for an example. This class creates a new parser in the createParser method. We need a new parser class e.g. ReportParser (extending IssueParser ). The parse method of this class should use the same code as AnalysisResult.readIssues() . The code in AnalysisResult already reads a report with several issues, so we basically need to reuse or extract that part.

            I have similar needs, I have made a proposal for a JSON based format in JENKINS-57098.

            jmini Jeremie Bresson added a comment - I have similar needs, I have made a proposal for a JSON based format in JENKINS-57098 .

            People

              hobbyte Raphael Furch
              derrickgw Derrick Gibelyou
              Votes:
              0 Vote for this issue
              Watchers:
              4 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: