[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Gnu-arch-users] Re: Questions from a Subversion user
From: |
John Goerzen |
Subject: |
[Gnu-arch-users] Re: Questions from a Subversion user |
Date: |
Mon, 18 Aug 2003 10:53:50 -0500 |
User-agent: |
Gnus/5.1002 (Gnus v5.10.2) Emacs/21.3 (gnu/linux) |
Thanks for the helpful reply. A couple of responses are included below:
Robert Anderson <address@hidden> writes:
> No, and my impression of the consensus is that this is a feature (that
> it does _not_ do this), as it simplifies merge semantics. At the same
Good enough. I'm OK with a paradigm shift :-)
> In arch we say "archive" not repository. And my first impression is to
> say that the size of the archive is essentially irrelevant. It is
> "suitable" for any size archive.
Excellent. Subversion starts to have some problems when you exceed
.5GB.
> arch supports a bunch of transports: ftp, sftp, http, and others. arch
> implements its own locking scheme. I am not sure what you mean by
> "insure integrity."
What I'm talking about is multiple users writing to the repository at
once. FTP (and SFTP, and HTTP) doesn't implement anything like
flock() -- and the best that one normally can do is creating
lockfiles, which have inherent race conditions and are thus
unreliable. It'd be really easy to have two people running a STOR on
the same file at once -- with likely disastrous results.
(I believe WebDAV does not have this problem.)
>> 4. A lot of documentation talks about people making new repositories
>> each year because old ones get "big". Why do people need to make
>> new repositories annually?
>
> Big in the sense of lots of categories and branches, many which may be
> "dead."
OK, that makes a lot more sense. I misunderstood the motivation for
this.
-- John
Re: [Gnu-arch-users] Questions from a Subversion user, Andrew Suffield, 2003/08/18