Bugzilla – Full Text Bug Listing
| Summary: | Encoding of generated Entities is always ISO8859-1 | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | [WebTools] Dali JPA Tools | Reporter: | Burghard Britzke <bubi> | ||||||
| Component: | General | Assignee: | Pascal Filion <pascal.filion> | ||||||
| Status: | RESOLVED FIXED | QA Contact: | |||||||
| Severity: | normal | ||||||||
| Priority: | P3 | CC: | bubi, ding870, mail, mathias.griepentrog, neil.hauge, olafzieger, smaragdraichu | ||||||
| Version: | 2.3.3 | ||||||||
| Target Milestone: | 3.0 M7 | ||||||||
| Hardware: | All | ||||||||
| OS: | All | ||||||||
| Whiteboard: | |||||||||
| Attachments: |
|
||||||||
|
Description
Burghard Britzke
I tried the test case and it seems to be working properly, i.e. I have the file enconding set to ISO-8859-1 in the IDE preferences (General->Workspace). I created a Java project and set the encoding to UTF-8. I generated the entities and I used äöüß as the file name for one of the classes, the Java editor is showing the right encoding. However, there is one setting that overrides the IDE and project levels encoding and it's the Content Types in the IDE preferences. In the IDE preferences under General->Content Types, Text->Java Source File, if the Default encoding is set, then it will be picked up. If I don't set the project level encoding and the IDE level encoding is set to ISO-8859-1, I do get the issue, but if I change the content type for Java Source File to UTF-8, then the problem goes away. (In reply to comment #1) we repeated the issue on minimal four different installations and two different OS platforms For our installations the (General->Workspace) setting mostly has been set to UTF-8 and the (Project->Properties->Resource) setting has been set to "Inherited from container (UTF-8)". The Content Type for Java Source File is not set (e. g. empty). following effects we recognized for different settings (here only for mac os x but I will try it for winxp tomorrow, too) In all cases the filenames of the Source Files are encoded correctly! Workspace->Project->Content Type->Effect UTF-8->inherited->empty->Class Names with � (Raute with Question Mark in it) UTF-8->inherited->UTF-8->Class Names with � (Raute with Question Mark in it) ISO-8859-1->inherited->empty->Class Names missing the äöü <--- THIS IS WHERE YOU GOT AN ISSUE TOO ISO-8859-1->UTF-8->empty-> Class Names with � (Raute with Question Mark in it) ISO-8859-1->UTF-8->UTF-8-> Class Names with � (Raute with Question Mark in it) You told that you got the issue if IDE level encoding is ISO-8859-1 ... but for this settings a class name encoding of ISO-8859-1 is not an issue but the expected result. We can repeat the issue for various settings and at least for mac os x and winxp. But let's take a closer look at our config: even if I think it does not matter: we use derby 10.7.1.1, EclipseLink 2.1.2 (Helios) Dali Java Persistence Tools Version 2.3.3.v201010220000 (Build id: 20100915173744) - mentioned above Here are the results for our WinXP tests: File names are all OK with umlauts. But in the Project Explorer View Class Names are cut before the Umlaut (We never started a class name with an umlaut in our tests). In the Editor View the class names are displayes as follows: UTF-8->inherited->empty->Class Names with � (empty square) UTF-8->ISO-8859-1->empty->Class Names OK UTF-8->inherited->ISO-8859-1->Class Names OK ISO-8859-1->inherited->empty-> OK ISO-8859-1->UTF-8->empty-> Class Names with � (empty square) ISO-8859-1->empty->UTF-8-> Class Names with � (empty square) On WinXP we have tested with Version 2.3.1.v201006300000 (Build 20100730021206) Repeated the Tests with WinXP and Dali Java Persistence Tools Version 2.3.3.v201010220000 (Build id: 20100915173744) Same as Version 2.3.1 on WinXP which we tested yesterday THE RESULTING ENCODING IS ISO-8859-1 no matter which encoding is choosen by either Workspace->Preferences, Project->Properties or Content Type->Text->Java Source File. You can repeat the error even if you try to generate Entities with UTF-16 encoding. I am trying to understand the problem. Are you saying the incorrect display of the class name is in the Project Explorer view? Which would means it's not displaying the file name coming from the file system correctly. (In reply to comment #5) > I am trying to understand the problem. Are you saying the incorrect display of > the class name is in the Project Explorer view? Which would means it's not > displaying the file name coming from the file system correctly. In the -->Project Explorer View<-- the filenames are displayed correctly. The Class Names are cut before the umlauts. In the -->Editor View<-- the Class Names are always displayed as if they where ISO-8859-1 encoded. This means: if the file encoding is set to UTF-8 by inheriting from the workspace settings or project properties, the generated files are displayed in a wrong manner. If you change the encoding settings for that file after generation, it is displayed correctly. I think this is a secure sign that the class names (the text in the java source file) is encoded in ISO-8859-1 but it should be encoded as preset in the project properties or in the workspace settings. Created attachment 192486 [details]
Using the right encoding when generating a Java file
No encoding was set when converting a String into byte[], JDK uses the system's default encoding. The fix was to use Eclipse's encoding for proper convertion.
Created attachment 192539 [details]
Using the right encoding when generating a Java file
This patch only has the fix, the previous patch was also deleting trailing spaces.
Committed to head. |