Nut upsdev - Dec 2011 - Still plugging away at the repo conversion

This is taking longer than I expected, because there is all kinds of
cvs2svn-generated ugliness in your back history that tends to crash 
my tools. Your repo makes a great test case. (Yes, I know, that's like
having someone say "Your dog is fantastically ugly!")

Current state of things is that I have scrapped my precious approach,
which was to put the Subversion tools in a Python harness and replay
the repo history.  The problem with this is that it was hideously 
slow - your 3.3K commits took over 8 hours to process.

I'm now working on a parser for Subversion dumpfiles.  This approach -
use svnsync to grab the history, use svnadmin dump to turn it into
a dumpfile, reposurgeon interprets the dumpfile - is much, *much* 
faster. A full read takes about 10 minutes.

Unfortunately, though my parser works fine on a dozen test repos, it
crashes during postprocessing on yours.  Weird combinations of stream
operations and Unicode in the comments are two issues. I'm working on
it.

A side effect of this project is that I am writing documentation 
for the Subversion dumpfile format that - gasp! - actually documents
it.  Nobody other than the Subversion devs themselves have ever tried
to get this intimate with it before.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>

Society in every state is a blessing, but government even in its best
state is but a necessary evil; in its worst state an intolerable one;
for when we suffer, or are exposed to the same miseries *by a
government*, which we might expect in a country *without government*,
our calamities is heightened by reflecting that we furnish the means
by which we suffer."		-- Thomas Paine

Hi Eric,

2011/12/14 Eric S. Raymond <esr at thyrsus.com>
> This is taking longer than I expected, because there is all kinds of
> cvs2svn-generated ugliness in your back history that tends to crash
> my tools. Your repo makes a great test case. (Yes, I know, that's like
> having someone say "Your dog is fantastically ugly!")
>
> Current state of things is that I have scrapped my precious approach,
> which was to put the Subversion tools in a Python harness and replay
> the repo history.  The problem with this is that it was hideously
> slow - your 3.3K commits took over 8 hours to process.
>
> I'm now working on a parser for Subversion dumpfiles.  This approach -
> use svnsync to grab the history, use svnadmin dump to turn it into
> a dumpfile, reposurgeon interprets the dumpfile - is much, *much*
> faster. A full read takes about 10 minutes.
>
> Unfortunately, though my parser works fine on a dozen test repos, it
> crashes during postprocessing on yours.  Weird combinations of stream
> operations and Unicode in the comments are two issues. I'm working on
> it.
>
> A side effect of this project is that I am writing documentation
> for the Subversion dumpfile format that - gasp! - actually documents
> it.  Nobody other than the Subversion devs themselves have ever tried
> to get this intimate with it before.
>
Thanks for this update, and the time you put into helping us (and others).
Do your best, and tell us back when you're ready.
I'll be releasing 2.6.3 tomorrow, and will then have a Christmas break
(back ~ January 3rd).

cheers,
Arnaud
-- 
Linux / Unix Expert R&D - Eaton - http://powerquality.eaton.com
Network UPS Tools (NUT) Project Leader - http://www.networkupstools.org/
Debian Developer - http://www.debian.org
Free Software Developer - http://arnaud.quette.free.fr/
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.alioth.debian.org/pipermail/nut-upsdev/attachments/20111215/bea67631/attachment.html>