Fix off issue #22 <?m2e execute onConfiguration?> fails if pom is big…

[barclay-reg]

… enough (8kb)

• fixed pom parser, as logic to check for x-m-l letters does not respected input tokenization
  ** all other occurences of <? are still ignored, but now: also when they appear behind READ_CHUNK_SIZE

[khmarbaise]

As far as I know there already configuration possibilities for m2e like this:

<pluginManagement>
  <plugins>
    <plugin>
     <groupId>org.eclipse.m2e</groupId>
     <artifactId>lifecycle-mapping</artifactId>
     <version>1.0.0</version>
     <configuration>
       <lifecycleMappingMetadata>
         <pluginExecutions>
           <pluginExecution>
             <pluginExecutionFilter>
               <groupId>some-group-id</groupId>
               <artifactId>some-artifact-id</artifactId>
               <versionRange>[1.0.0,)</versionRange>
               <goals>
                 <goal>some-goal</goal>
               </goals>
             </pluginExecutionFilter>
             <action>
               <ignore/>
             </action>
           </pluginExecution>
         </pluginExecutions>
       </lifecycleMappingMetadata>
     </configuration>
    </plugin>
  </plugins>
</pluginManagement>

[barclay-reg]

yes - this was standard up to m2e 1.6, but the new notation (<?m2e) has some advantages - especially in big pom.xml files. Ironically this big pom.xml cannot be parsed at the moment. It's supported by XML standard, so the parser is - in this niche - non standard compliant

[rfscholte]

I'm missing a unittest to confirm the fix. Would be nice if you can provide that as well.

[rfscholte]

Seems be be related to https://issues.apache.org/jira/browse/MNG-6216

[barclay-reg]

@rfscholte tests added

[rfscholte]

Thanks. But is seems that if I only apply the test, it is already running succesfully...

[barclay-reg]

you are correct - test is fixed now, so it fails for the old code. i was trapped by the buffer resizing, which accidentally let the test never fail (buf was alway big enough to avoid the ArrayIndexOutOfBoundException)

[rfscholte]

Exactly the reason why unittests are required :) Thanks for the patch!