ERROR: Error during SonarQube Scanner execution

[Zakaryae Amekkaoui]

Hi everyone,

I'm trying to scan a project, but i get this error :

ERROR: Error during SonarQube Scanner execution java.lang.IllegalArgumentException: 26298 is not a valid line for pointer.

I tried to specify the encoding in sonar-project.properties ( sonar.sourceEncoding=UTF-8), but it doesn't work.

I'm using windows8-64.

Thank you,

Have a nice day

Zack.

[Scott B.]

Hi.

You didn't say what is the SonarQube version, what programming language you're analyzing, what is the plugins and versions installed on your instance... Nobody will help you.

[Zakaryae Amekkaoui]

Hi Scott,

Thanks for your response, i'm using version 6.7.1 for SonarQube and sonar-scanner-3.0.3.778.

The programming language is NCL.

This is the rest of the error message  :

INFO: 168/168 source files have been analyzed

INFO: ------------------------------------------------------------------------

INFO: EXECUTION FAILURE

INFO: ------------------------------------------------------------------------

INFO: Total time: 4:21.153s

INFO: Final Memory: 87M/1611M

INFO: ------------------------------------------------------------------------

ERROR: Error during SonarQube Scanner execution

java.lang.IllegalArgumentException: 662742 is not a valid line for pointer. File

 sources/AGAS1.EXP has 3974 line(s)

        at org.sonar.api.internal.google.common.base.Preconditions.checkArgument

(Preconditions.java:145)

        at org.sonar.api.batch.fs.internal.DefaultInputFile.checkValid(DefaultIn

putFile.java:309)

        at org.sonar.api.batch.fs.internal.DefaultInputFile.newPointer(DefaultIn

putFile.java:246)

        at org.sonar.api.batch.fs.internal.DefaultInputFile.selectLine(DefaultIn

putFile.java:269)

        at org.sonar.plugins.ncl.NclSquidSensor.saveViolations(NclSquidSensor.ja

va:213)

        at org.sonar.plugins.ncl.NclSquidSensor.save(NclSquidSensor.java:123)

        at org.sonar.plugins.ncl.NclSquidSensor.execute(NclSquidSensor.java:106)

        at org.sonar.scanner.sensor.SensorWrapper.analyse(SensorWrapper.java:53)

        at org.sonar.scanner.phases.SensorsExecutor.executeSensor(SensorsExecuto

r.java:88)

[Scott B.]

Oh boy... this is a custom plugin, right? Is your plugin or it's from a third party? Anyway, it seems a bug in the plugin.

In older SonarQube versions, to read the contents of a org.sonar.api.batch.fs.InputFile, the only option was to use the default encoding (defined in the sonar.sourceEncoding property).

Since SonarQube ~6.3, the platform got smarter and started to detect the file encoding automatically. But it has a downside: when you read the file using an encoding diffent than the detected by SonarQube you can hit this "xxxx is not a valid line for pointer" exceptions.

An easier option is to fix the encoding of your source code, to leave all files with the same encoding. Then define sonar.sourceEncoding and you'll be ok.

If you want to fix the plugin, I'm sorry to say: you have a long journey ahead. The only definitive way to fix the plugin is to NOT read the file manually (like "Files.toString(inputFile.file(), charset)") and start to use the InputFile.contents method (avaliable in sonar-plugin-api 6.2).

Just remembering how hard I worked to fix this in a custom plugin gives me headhaches.

[Zakaryae Amekkaoui]

Yeep it's a custom plugin for a proprietary language.

I tried the first solution but it doen't work, so i've to try the second one.

Thanks very much Scott for your response.

Have a nice day.

[Zakaryae Amekkaoui]

Hi Scott again,

I tried the second solution, but i stell have the error "java.nio.charset.malformedinputexception input length = 1".

My solution is like : 

@Override

public void visitFile(@Nullable AstNode astNode) {

       String pth = getContext().getFile().toString(); 

       try { 

              Stream<String> lignes = Files.lines(Paths.get(pth));

              lignes.forEach(line -> {

           .....

           });

      }

}

Do you have any idea for this problem ?

Thanks.

Have a nice day.

Zack

[Julien HENRY]

Hi Zack,

Using Files.lines() without specifying encoding means you are assuming all files are encoded in UTF-8. We have an algorithm to detect encoding in SonarQube, and so the safest way to be sure you are processing the input file with the same encoding as the one detected by SonarQube is to use InputFile::charset() API.

Regards,

Julien

[Zakaryae Amekkaoui]

Hi Julien,

Thank you for your response, can you please tell me how can i use this API ? i didin't found an example for that.

Regards.

Zack

[Julien HENRY]

Hi,

In your NclSquidSensor, you can access InputFiles.

See for example:

https://github.com/SonarSource/sonar-custom-plugin-example/blob/76cc0682cae67e33d21f98a56d05b423c0e3ca7e/src/main/java/org/sonarsource/plugins/example/measures/SetSizeOnFilesSensor.java#L45

++

Julien Henry | SonarSource

Developer

https://sonarsource.com

--
You received this message because you are subscribed to a topic in the Google Groups "SonarQube" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/sonarqube/__TXqiQaEYo/unsubscribe.
To unsubscribe from this group and all its topics, send an email to sonarqube+unsubscribe@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/sonarqube/b825e649-5e43-429d-bf31-32cc030af526%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Zakaryae Amekkaoui]

Hi,

Thanks a lot Julien, in my NCLSuiqSensor i have this code :

  @Override

  public void execute(SensorContext context) {

    FileSystem fileSystem = context.fileSystem();

    FilePredicates predicates = fileSystem.predicates();

    List<SquidAstVisitor<LexerlessGrammar>> visitors = new ArrayList<>(checks.all());

    visitors.add(new FileLinesVisitor(fileLinesContextFactory, fileSystem));

    NclConfiguration configuration = new NclConfiguration(fileSystem.encoding());

    scanner = NclAstScanner.create(configuration, visitors);

    Iterable<java.io.File> files = fileSystem.files(

      predicates.and(

        predicates.hasType(InputFile.Type.MAIN),

        predicates.hasLanguage(Ncl.KEY),

        inputFile -> !inputFile.absolutePath().endsWith("mxml")

      ));

    scanner.scanFiles(ImmutableList.copyOf(files));

    Collection<SourceCode> squidSourceFiles = scanner.getIndex().search(new QueryByType(SourceFile.class));

    save(context, squidSourceFiles);

  }

So i tried to change the configuration to UTF-8, come thing like that  :

   NclConfiguration configuration = new NclConfiguration(Charset.forName("UTF-8"));

But it doesn't work. Do you have any suggestion ?

Thanks.

Have a nice day.

Regards.

Zack.

To unsubscribe from this group and all its topics, send an email to sonarqube+...@googlegroups.com.

[Julien HENRY]

I'm sorry but I'm not very confident with SSLR/sonar-channel library. You can maybe look at open source SonarSource code analyzers to see how they are managing to pass the InputFile charset to the SourceCode layer.

Julien Henry | SonarSource

Developer

https://sonarsource.com

To unsubscribe from this group and all its topics, send an email to sonarqube+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sonarqube/4691b2cf-0cda-49ec-9e64-b03165100f07%40googlegroups.com.

[Zakaryae Amekkaoui]

Okey thanks a lot Julien :)

Regards

Zack