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

Re: svn commit: r1463721 - in /subversion/trunk/subversion/libsvn_client: checkout.c client.h commit.c copy.c export.c externals.c revert.c switch.c update.c

From: Philip Martin <philip.martin_at_wandisco.com>
Date: Thu, 11 Apr 2013 14:21:48 +0100

Julian Foad <julianfoad_at_btopenworld.com> writes:

> I suppose I was thinking I couldn't be sure that the same path was
> used in each possible case.  Looking at it again, if dst_path is not a
> URL then it would be reasonable to assume that all relevant changes
> are on the same filesystem as dst_path.  Even more so if we use the
> child of dst_path instead, in the copy_as_child case; then the path
> used would be, I think, the same as it was when the
> sleep_for_timestamps calls were distributed among the leaf functions
> before my patch.

The path is used to estimate the timestamp resolution, it dosen't matter
which path in the destination is used.

The other thing to consider is a move between working copies: the source
might be on a filesystem with 1s sleeps while the destination is on a
filesystem with 1ms sleeps. Even then it should be OK to use dst_path
because the source is only getting the delete so there are no timestamps
of concern on that filesystem.

-- 
Certified & Supported Apache Subversion Downloads:
http://www.wandisco.com/subversion/download
Received on 2013-04-11 15:22:23 CEST

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