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

Re: File descriptor leak of rep-cache.db in 1.6.x

From: Dan Villiom Podlaski Christiansen <danchr_at_gmail.com>
Date: Thu, 3 Jun 2010 22:19:09 +0200

On 3 Jun 2010, at 19:27, Dan Villiom Podlaski Christiansen wrote:

> Some statistics:

…which were wrong. Jelmer was a bit surprised that Subvertpy should be 2x slower than the SWIG bindings, and he was right: I had some debugging code left over which called ‘gc.collect()’ regularly.

BDB repository w. Subvertpy:
Virtual private memory: 46 MB
CPU time: 2m47s

FSFS repository w. Subvertpy:
Virtual private memory: 415 MB
CPU time: 2m48s

BDB repository w. SWIG:
Virtual private memory: 408 MB
CPU time: 3m10s

FSFS repository w. SWIG:
Virtual private memory: 410 MB
CPU time: 3m07s

> Based on this I would say that BDB repositories appear to not leak much when using Subvertpy. Everything else leaks quite a lot. Granted, there is a significant CPU overhead in Subvertpy, but I believe its much cleaner API and better memory handling compensate for that. Besides, many people using hgsubversion will have their conversions I/O bound rather CPU bound.

More accurately: There appear to be no downsides to using Subvertpy.

--
Dan Villiom Podlaski Christiansen
danchr_at_gmail.com

  • application/pkcs7-signature attachment: smime.p7s
Received on 2010-06-03 22:19:58 CEST

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