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

Re: RFC: svn_io_file_flush_to_disk

From: Greg Hudson <ghudson_at_MIT.EDU>
Date: 2004-05-01 20:56:56 CEST

On Sat, 2004-05-01 at 14:50, D.J. Heap wrote:
> I'm a bit paranoid about the retrying code not being good enough on a
> reader-heavy system -- but you certainly know better than I how long
> and/or often 'current' will be held by readers, so I'll drop it if you
> think it's reasonable.

In general, a reader will open the "current" file for just long enough
to read it (and it's short), and then go off and do something much more
expensive. So, even if there are thousands of concurrent readers, I
don't think they would block a writer trying to update the current file.

If it becomes a practical problem, we can do something awful, like have
multiple "current" files where the reader looks for the highest-numbered
one to decide which one is real. But I think we should wait for it to
become a practical problem first.

To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat May 1 20:57:14 2004

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

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.