subversion-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Bert Huijben" <>
Subject RE: Compressed Pristines (Summary)
Date Wed, 04 Apr 2012 20:59:33 GMT

> -----Original Message-----
> From: Ashod Nakashian []
> Sent: woensdag 4 april 2012 13:28
> To: Greg Stein
> Cc:
> Subject: Re: Compressed Pristines (Summary)
> >________________________________
> > From: Greg Stein <>
> >My vote is compress for files > N bytes (and store on disk), and stick
> others into a new pristine.db. Maybe compress before inserting into
> Not sure. Add a few heuristics for skipping compression on certain files
> types.
> >Shouldn't be too bad to start with that.
> >Cheers,
> >-g
> >
> I feel this is indeed what we're closing on, at least for an initial
working demo.
> But I'd like to hear more agreements before committing to this path. I
> some did show support for this approach, but it's hard to track them in
> noise.
> So to make it easier, let's either voice support to this suggestion and
> to an implementation, or voice objection with at least reasons and
> alternative action. Silence is passive agreement, so the onus on those
> opposing ;-)

The combination of some compression (gzip or plain deflate), maybe some
specific uncompressed files and the small files in SQLite sounds good.

I think the cutoff point for SQLite should be below what we can have in
memory as we can't keep database transactions open to back up streams. 32K
(and a few times higher) should be safe to read in memory at once.


View raw message