|
From: | Markus Schiltknecht |
Subject: | Re: [Monotone-devel] Time to think about a release? |
Date: | Tue, 25 Mar 2008 19:10:17 +0100 |
User-agent: | Mozilla-Thunderbird 2.0.0.9 (X11/20080110) |
Hi, Zack Weinberg wrote:
Hm.. does it really make sense to piggyback those to db-compaction? That branch already contains way too many changes, IMO, so I'd vote for doing that in a separate branch.I had two reasons: it gets awkward to have multiple outstanding branches with schema changes,
Oh, yeah, I understand. That's tedious. However, landing db-compaction first would leave only the IDNA change as outstanding :-)
and I'm not sure the change I have in mind touches the schema hash.
..while still wanting a schema migration step, that is? Maybe we should add a schema version number or a cuddly name or something, which we can change to force a new database schema hash, without really having to touch the schema.
I'd like to understand why a regenerate_caches run is necessary, at least. It really oughtn't.
You are right, it shouldn't be needed. I've fixed that now, including an upgrade to the schema_migration test. No regeneration of rosters is needed anymore. (I just forgot to unhex the rosters and roster_deltas checksum).
With that fix, db-compaction is successfully running all unit tests, now (with the given exceptions, see below).
We have test failures on mainline again? ugh.
Unfortunately, yes. Those are: ssh_agent and non_workspace_keydir. The later one being pretty new. But I didn't look into it.
Regards Markus
[Prev in Thread] | Current Thread | [Next in Thread] |