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

Re: more wc-ng node revision differences that affect 'svn diff'

From: Greg Stein <gstein_at_gmail.com>
Date: Sun, 18 Oct 2009 16:24:55 -0400

On Sun, Oct 18, 2009 at 15:12, Stefan Sperling <stsp_at_elego.de> wrote:
> Hey,
>
> The patch below results in test failures, as shown below.
>
> Like other failures I've seen, the tests fail because revisions
> reported by wc-ng for added or deleted nodes don't match up with
> what wc-1 reported.

Right. wc-1 overloaded the entry->revision field to mean different
things, in different states. Very messy.

wc-ng cleans all this up. You only have a revision value for something
direct from the repository. And then if you add/delete that node, then
the revision is SVN_INVALID_REVNUM because an uncommitted change does
not have a revision in the repository. You can still look "under" that
change and get the revision number of the node from the repos, if that
is necessary, but the changed node doesn't really have one.

entry->revision has different semantics, so its value will not match
what wc-ng will tell you.

> Can someone explain the rationale behind these differences?
> Should I simply adjust the expected output or do these tests failures
> exhibit undesired behaviour?

It looks like the expected output of those tests is proper (if you
ignore the "revision 0" to imply a missing left node). I don't
understand why your change is reporting "revision 1" for those nodes.
They're clearly supposed to be r2 or r3.

>...

Cheers,
-g

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2408716
Received on 2009-10-18 22:25:07 CEST

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