Destroyed chars in XMI version control file

[q2]

My configuration:
- EA 13 (1310),
- Windows 10,
- projects under subversion version control (svn),
- Preferences->Xml Specification->Code page = UTF-8
- Types of local model DB: eap, feap, MySQL

I have found that characters from outside basic ascii scope (in my particular case Polish chars like ą, ź, ć are destroyed during XMI file genereation).
Finally such example source text "Zażółć gęślą jaźń" in notes of any EA element after check in and check out is destroyed to such text "Zażółć gęślÄ… jaźń". Subsequent check-ins produce quite new longer strange text.

I have found temporary solution - just to switch to default windows-1252 code page (Preferences->Xml Specification->Code page). In such case "foreign" chars are stored as XML entities and all works OK...

Another observations:
1. Manual export of package to some XMI (i.e. other than svn folder) makes correct UTF-8 file!
2. On configuration like below:
- EA 13 (1310),
- Windows 2003,
- projects under subversion version control,
- Preferences->Xml Specification->Code page = UTF-8
- Types of local model DB: feap
XMI svn files are generated correctly.

[qwerty]

You'll get coders all around crying for not using ASCII. Probably we never should have stopped talking Bongos which already had been a valid binary code.

Anyhow, most likely the issues come from your EAP files. You need to turn on Jet4 which is not default. Just google or try the forum search here (good luck) if you need details on how to do that (it's in the options and you need the drivers which are under some non-free license).

q.

[Geert Bellekens]

I have recently discovered that exporting XML Schema's using UTF-8 did the same thing with french characters like çàé.
Changing the option in the generator dialog to unicode made it work as expected, however that also had an effect on the generated xsd files, which was unwanted.
IIRC UTF-8 should be able to handle extended ascii the same way unicode does, so I was a bit surprised that using UTF-8 resulted in issues like that.
Could very well be a bug.

Geert

[q2]

I do not think so this is eap file and Jet4 issue because:

1. Manual export of package to XMI file with code page option set UTF-8 works OK. Version control "check-in" with this option set to UTF-8 produces wrong characters in svn file.
2. The wrong behavior happens not only on MS Access eap but with using feap and MySQL as well.

Anyway to be sure I have made some exercises with Jet 4 and downloaded from Sparx unicode eap file but without success