Hi Jörg,
2006/6/12, Jörg Schaible <[EMAIL PROTECTED]>:
Hi Stefan, Stefan Hübner wrote on Sunday, June 11, 2006 9:25 AM: > Hi all, > > an issue has been created on this over at > http://jira.codehaus.org/browse/MECLIPSE-56, already. Fixing it > decently seems to be quite a hassle though, because maven doesn't care > much about pom.xml's encoding. at the moment the developers seem to > consider rewriting the whole parsing stuff from scratch. well, this might be another effect of the encoding problem, but this has nothing to do with Eclipse in special. As stated, the deployed POM will have the entities (like "ö") resolved and the XML is simply invalid.
[snip] you're absolutely right, Jörg. The encoding misery is not limited to Eclipse at all. This is just one symptom of many. Maven doesn't read POMs using their declared encoding properly. It's certainly able to resolve all sorts of entities but only with the effect, that those characters will lead to new problems during POM-exports. Maven simply uses the local system's default character set while parsing or writing a POM. The same probably happens while writing POMs. To me it seems, that lots of those different sorts of character set problems will only be resolved, if Maven starts to fully take into account the POM's declared encoding. You might have a look at http://jira.codehaus.org/browse/PLX-214 for further discussions. jm2c, Stefan