Bugzilla – Full Text Bug Listing
| Summary: | Turkish character problems when commenting an issue in JIRA using built-in issue editor | ||||||
|---|---|---|---|---|---|---|---|
| Product: | z_Archived | Reporter: | Oğuz Dağ <oguz.dag> | ||||
| Component: | Mylyn | Assignee: | Steffen Pingel <steffen.pingel> | ||||
| Status: | VERIFIED FIXED | QA Contact: | |||||
| Severity: | major | ||||||
| Priority: | P2 | CC: | blpsilva, ekuleshov, mads.stavang, robert.elves, steffen.pingel | ||||
| Version: | dev | Keywords: | helpwanted | ||||
| Target Milestone: | 2.1 | ||||||
| Hardware: | PC | ||||||
| OS: | Windows XP | ||||||
| Whiteboard: | |||||||
| Attachments: |
|
||||||
|
Description
Oğuz Dağ
Eugene, Steffen: what's the current status of encoding handing with the Jira Connector? It was tricky to get this stuff right for Bugzilla... *** Bug 202235 has been marked as a duplicate of this bug. *** Here are my observations from a quick investigation:
1) encoding on mylyn.eclipse.org was set to UTF-8
- submitting of taks with non-USASCII characters works
2) changed encoding to ISO-8895-1
- Umlauts (which can be represented in ISO-8895-1) were garbled
The JIRA connector explicitly sets the charset in JiraWebIssueService:
post.setRequestHeader("Content-Type", "application/x-www-form-urlencoded; charset=utf-8");
That might need to be set to the repository charset (which can potentially be determined from the HTTP headers).
Committed fix to CVS. - The character set is now determined from the HTTP headers when the repository is validated. - If the encoding can not be determined a warning is displayed in the repository settings dialog and UTF-8 is used as the fallback. - Older Mylyn installations are likely to have the default charset (UTF-8) set which might not match the repository encoding. To avoid using the wrong charset it is dynamically determined during server login until the user once validates the repository. Marking as resolved. Created attachment 78511 [details]
mylyn/context/zip
Great stuff Steffen! BTW, if user will manually set encoding in repo settings, will we use that encoding with your patch? You need to validate the repository once. If you manually set the character encoding after that, it will be used. Great stuff Steffen! Should we consider auto-detection of encoding as something that should be done at the framework level? I don't think we need any additional facilities on the framework level. The detection code is very connector specific and in the case of HTTP trivial as it is read from the headers. A fixed dev build is available from the update site: http://www.eclipse.org/mylyn/downloads/builds.php *** Bug 200430 has been marked as a duplicate of this bug. *** Hi, I have noticed a different behavior on the server validation, as it did change my encoding preferences do ISO-8859-1. However, when i try to synchronize queries with the server, i still get the same error i reported on bug #200430 ("com.sun.org.apache.xerces.internal.impl.io.MalformedByteSequenceException: Invalid byte 2 of 3-byte UTF-8 sequence") Just for the record, i've used the DEV builds present in the following update sites: download.eclipse.org/tools/mylyn/update/dev/e3.3 download.eclipse.org/tools/mylyn/update/dev/extras Thanks for the feedback. I have reopened bug 200430 to investigate the problem. (In reply to comment #11) > A fixed dev build is available from the update site: > > http://www.eclipse.org/mylyn/downloads/builds.php > (In reply to comment #11) > A fixed dev build is available from the update site: > > http://www.eclipse.org/mylyn/downloads/builds.php > Thanks for your effort... I tried, It works fine... |