[MECLIPSE-683] UTF8 Byte Order Marks in POM files are not ignored correctly

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Won't Fix
• Affects Version/s: 2.8
• Component/s: Core : Dependencies resolution and build path (.classpath)
• Assignee: Unassigned
• Reporter: Martin Müller
• Created: 16/Mar/11 05:54
• Updated: 25/Nov/14 14:47
• Resolved: 25/Nov/14 14:47

Description

This is the same problem as in the maven sql plugin (http://jira.codehaus.org/browse/MSQL-33).

If one builds with
% mvn eclipse:clean eclipse:eclipse -Declipse.workspace=WORKSPACE_LOCATION
and pom.xml files contain the utf8 byte order mark (http://en.wikipedia.org/wiki/Byte_order_mark) the projects in the workspace are not recorgnized correctly in the eclipse workspace. A warning indicates that the procect is not recognized:

[WARNING] could not read workspace project:\Projekte\Workspaces\Trunk\.metadata\.plugins\org.eclipse.core.resources\.projects\projectname
org.codehaus.plexus.util.xml.pull.XmlPullParserException: only whitespace content allowed before start tag and not \uef (position: START_DOCUMENT seen \uef... @1:1)
at org.codehaus.plexus.util.xml.pull.MXParser.parseProlog(MXParser.java:1519)
at org.codehaus.plexus.util.xml.pull.MXParser.nextImpl(MXParser.java:1395)

This leads to a external dependencies instead of correct workspace dependencies.

Testcase can be easily created with e.g. Copy XML Editor, which defaults to save the byte order mark in UTF-8 files (see options).