Postgres sequence not reset after add_last_activity_to_build upgrade script

[osimons]

A table is copied back and forth, but the sequence is not updated afterwards so the sequence gets reset to 1.

This was the bitten_build_id_seq.

[rblank]

I assume you know about ​#T8575? Fixed only in 0.12.1, but you could borrow the sequence update line for PostgreSQL.

[osimons]

Already got one, thanks: source:/trunk/bitten/upgrades.py#L35​

No magic involved, so you still have to remember to actually call it... :-)

[hodgestar]

Tests of sequence values post upgrade commited to trunk in r924. Fix committed to trunk in r925. Both backported to 0.6.x in r926.

Note: After discussion with osimons I fixed the relevant upgrade function rather than bumping the schema version and creating a whole new upgrade function. The grounds for this decision was that there hasn't actually been a release since the bug was introduced.