Community
Participate
Working Groups
If an imported BPEL resource begins with an <?xml?> processing instruction, the content type describer for BPEL is not invoked by the Eclipse framework. Instead, the *.bpel file is described as a plain XML file and the XML editor is opened instead of the BPEL Desginer.
Fixed. This was caused by content type IDs in org.eclipse.bpel.common.ui and org.eclipse.bpel.compare.ui plugins conflicting with the one defined in the org.eclipse.bpel.ui plugin.
This never showed up before because the New File Wizard doesn't generate the <?xml?> processing instruction. Should it?
I can confirm that the fix solves the problem. But I still have one question: If BPEL designer does not write a XML header in its BPEL files than it does save them UTF-8 encoded (since this is a default), right?
(In reply to comment #3) > I can confirm that the fix solves the problem. > But I still have one question: > If BPEL designer does not write a XML header in its BPEL files than it does > save them UTF-8 encoded (since this is a default), right? Yes, as far as I know. Oddly enough, BPEL Designer uses a org.eclipse.wst.sse.ui.StructuredTextEditor to read the .bpel file. The StructuredTextEditor reader is highly fault tolerant (which has caused me a lot of grief!) and tries very hard to make sense of its input. This usually results in a more or less valid DOM which is then used to construct the BPEL EMF model. I have not dug that deeply into the StructuredTextEditor so I'm not familiar with how (or if!) it handles different encodings. My hope is that whatever XML parser is behind the scenes, it knows about the various encodings. A quick search of bugzilla for "encoding" or "utf" returns Zarro Boogs found, so either this is not an issue or it has not been discovered yet ;) If you have the time and inclination, I would sure appreciate it if you could test out different encodings (UTF-16 big/little endian, iso-8859-1 (Latin-1), shift-jis, etc.) and let me know if there are problems. In the meantime, I'll close this bugzilla. If you find an encoding problem, please open a new one.
