Title | P4DTI not known to support unicode |
Status | closed |
Priority | essential |
Assigned user | Gareth Rees |
Organization | Ravenbrook |
Description | Perforce 2002.1 has a unicode mode, in which files of type "unicode", and all metadata, are stored in UTF-8. Then this UTF-8 data is translated to and from the client character set on every access. The P4DTI has not been tested in this mode, and we don't know whether this works. |
Analysis | Unicode-mode server requires unicode-mode on clients, and vice versa. In fact it does work (tested against Perforce 2005.1) if P4CHARSET is set when running the P4DTI (forcing the client into unicode mode). We need to do better. Bugzilla can handle UTF8 data, and that's the recommended hack for Bugzilla given non-ASCII data. So change P4DTI to switch automatically between unicode (UTF8) and non-unicode client modes, when a unicode mode error is detected, and to generate an error if this switching doesn't fix the problem. Change test suite to test this behaviour with both unicode and non-unicode servers (by adding a test which runs the 'normal' test against a unicode server). Test with non-ASCII data entered both from Perforce and from Bugzilla. |
How found | customer |
Evidence | <http://info.ravenbrook.com/mail/2002/01/18/12-00-09/0.txt > |
Observed in | 1.3.3 |
Created by | Gareth Rees |
Created on | 2002-02-01 14:56:02 |
Last modified by | Nick Barnes |
Last modified on | 2005-10-21 15:49:11 |
History | 2002-02-01 GDR Created. 2005-10-21 NB Wrote analysis, figured out what to do, did it. |
Change | Effect | Date | User | Description |
---|---|---|---|---|
155470 | closed | 2005-10-19 18:19:33 | Nick Barnes | Numerous changes to bring P4DTI up-to-date with Bugzilla 2.20 and unicode Perforce servers. |