some db cursors are not explicitly closed
Bug #158386 reported by
Eleanor Berger
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Launchpad itself |
Triaged
|
Low
|
Unassigned |
Bug Description
Every now and then we need to get a low-leve DB-API cursor (instead of using SQLObject). After we use such a cursor, we should explicitly close it (the effects of neglecting to are unknown to me, but the DB-API documentation indicates that this is mandatory), but there are some places in the code where we don't do that right now.
This bug is for identifying all these places and correcting them.
See, for example: lib/canonical/
Changed in launchpad: | |
assignee: | nobody → launchpad |
Changed in launchpad: | |
assignee: | launchpad → nobody |
Changed in launchpad-foundations: | |
status: | New → Triaged |
importance: | Undecided → Low |
visibility: | private → public |
Changed in launchpad: | |
importance: | Low → High |
To post a comment you must log in.
Hi Tom,
Right, I think it will be clearer that way, but I couldn't find a mention in http:// www.python. org/dev/ peps/pep- 0249/, is it missing or am I not paying enough attention ?
Check the very specific soyuz-callsites:
{{{ buildmaster/ buildergroup. py:386 archivepublishe r/domination. py:326 ftpmaster- tools/initialis e-from- parent. py:152 ftpmaster- tools/sync- source. py:472 ftpmaster- tools/sync- source. py:565
lib/canonical/
lib/canonical/
scripts/
scripts/
scripts/
}}}
Obviously there are other callsites related to soyuz in launchpad/database but they are more trivial to sort.
The diff attached address these occurrences.