Opened 18 years ago
Closed 16 years ago
#2808 closed defect (wontfix)
sourceforge2trac importer does not work (with postgres)
| Reported by: | Owned by: | Jonas Borgström | |
|---|---|---|---|
| Priority: | normal | Milestone: | |
| Component: | general | Version: | 0.9.4 |
| Severity: | normal | Keywords: | sourceforge2trac unicode |
| Cc: | Branch: | ||
| Release Notes: | |||
| API Changes: | |||
| Internal Changes: | |||
Description
The sourceforge2trac contrib script fails due to not using the DB-API correctly and assuming sqlite internals. It also seems to cause segmentation faults whenever it tries to add a ticket with non-ASCII characters.
Attachments (1)
Change History (6)
by , 18 years ago
| Attachment: | sf2trac.ptc added |
|---|
comment:1 by , 18 years ago
comment:2 by , 18 years ago
Well, I guess a simple .encode('utf-8') will work with UNICODE/UTF8 databases, right?
This will probably become a non-issue if/when the unicode branch is merged.
comment:3 by , 16 years ago
| Milestone: | 1.0 |
|---|
Unicode branch merged quite some time ago (0.10), but would be nice to get verification that this ticket can be closed. #3521 with patch proves recent success with import, so hopefully someone can confirm this issue also for non-ascii characters.
comment:4 by , 16 years ago
I'm afraid I don't have either the data or resources to verify this; I ran into this for a one time migration and have not used or touched any sf trackers since.
comment:5 by , 16 years ago
| Keywords: | sourceforge2trac unicode added |
|---|---|
| Resolution: | → wontfix |
| Status: | new → closed |
So, to be reopened if needed.
This patch tries to fix this. The solution is not optimal: it just does a nasty encode("ascii","replace) on all data to strip non-ASCII character. Unless a unicode-aware postgres connection is used this might be the only way to get this working unfortunately.