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

Re: [PATCH] fix for issue #2556: support working copies on the root of a (virtual) drive

From: Lieven Govaerts <lgo_at_mobsol.be>
Date: 2006-08-19 09:40:34 CEST

My intention was to follow up these patches, which contain only a basic
test scenario, with some extra tests and fixes (where needed). I have
two approaches in mind:
1. checking the code the same way like you did.
2. locally modifying the python test framework to map each created
working copy to a windows virtual drive. Then I'll test all svn actions
on the root of a drive.

FYI: There's also issue #1711, which theses patches partially solve, and
issue #1869, which has been solved in the apr tree. I just touched this
issue because my colleagues want to use virtual drives on Windows, but
I'll try to fix the other issue as well.

Philip Martin wrote:

> Grepping for \[0\].*\'/\' shows that there might be a few places where
> we assume that path[0]=='/' is special; I don't know whether your
> patch invalidates those assumptions. If path is a Subversion FS path
> then it's OK but if path is a WC path then it could be a problem. For
> example:
> - libsvn_client/commit.c:1312 seems to be checking for "/", should it
> be checking for "x:/" as well?

Yes it should. This can be tested by committing multiple directories on
the root of a drive.

> - libsvn_wc/props.c:2110 seems to be checking for "/...", should it be
> checking for "x:/..." as well?

If I read the code correctly, this is the case of a propset on the root
of a drive.

> - libsvn_client/diff.c:401 and 408 seem to be checking for "/...",
> should they be checking for "x:/..." as well?

As far as I see, this can be tested by a diff on two files in the root
folder of a drive.

> I don't know if these are bugs, and if they are I suspect they are
> unrelated to your patch, but someone who groks this Windows stuff
> probably ought to take a look :)

I prefer to commit these two patches first, and then follow up with any
other cases we can find in the code. I'll keep the locations you found
on my TODO list for issue 2556.



To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Aug 19 09:40:27 2006

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.