[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: svn commit: r1637184 - in /subversion/trunk/subversion: libsvn_fs_fs/ tests/libsvn_fs/

From: Branko ─îibej <brane_at_wandisco.com>
Date: Fri, 07 Nov 2014 05:25:11 +0100

On 07.11.2014 00:44, Stefan Fuhrmann wrote:
>
> Furthermore, hot upgrades are a new feature in FSFSv7: So, if you have
> an issue with that, by all means propose that we forbid hot
> upgrades, as
> that will not be a regression compared to FSFSv6. Do not rip out a
> completely orthogonal, important feature instead.
>
>
> Brane, you might confuse this with 'svnadmin pack'. The thing
> about upgrade is that we have a nasty FS API limitation that
> is a missing concept of read sessions / locks. Everyone can
> open a repo, keep the svn_fs_t, upgrade the repo and then
> *write* to it using the old FS pointer with the old format info.
> IOW, we must make sure that writers can happily ignore any
> format info as long as they are not trying to write in an even
> newer format than specified on disk.

[...]

Thanks for clarifying. This is more or less what I meant ... IOW, with
v7, someone actually made an effort to make sure we'd fail consistently
in some scenarios.

-- Brane
Received on 2014-11-07 05:28:03 CET

This is an archived mail posted to the Subversion Dev mailing list.