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

Re: RFC -- Configurable Diff Path

From: Erik Huelsmann <ehuels_at_gmail.com>
Date: 2007-07-19 23:12:45 CEST

On 7/17/07, Mark Phippard <markphip@gmail.com> wrote:
> On 7/16/07, Jeremy Whitlock <jcscoobyrs@gmail.com> wrote:
> > While working on issue 2723, which will update the core diff API
> > to have a method argument to specify the format (absolute or relative)
> > of the Index path output, has brought up conversation about having the
> > "svn diff" subcommand support an argument to override the default
> > handling of the Index path output. As of now, here are the tested
> > scenarios:
> >
> > Scenario 1
> > ---------------
> > svn diff RELATIVE_PATH_TO_WC (Outputs RELATIVE_PATH_TO_WC/MODIFIED_PATH)
> >
> > Scenario 2
> > ---------------
> > svn diff ABSOLUTE_PATH_TO_WC (Outputs ABSOLUTE_PATH_TO_WC/MODIFIED_PATH)
> >
> > Scenario 3
> > ---------------
> > svn diff -rOLD:NEW URL_TO_PATH (Outputs
> > PATH_RELATIVE_TO_REPO_ROOT/MODIFIED_PATH)
> >
> > Now...why does this matter? Well, all of these scenarios end up with
> > the Index path not being -p0 compliant, which is the standard for most
> > project patches.
>
> I do not know about this, but I will take your word for it. I do not
> recall a lot of complaints about the format of svn diff output though,
> so I'd be careful in changing the command line. This request was all
> about JavaHL. In that environment we have not control over the CWD so
> we always have to use absolute paths. This of course makes the
> patches that we produce using Subclipse more difficult for the user to
> apply.

Right. My primary concern would be to resolve the issue in a way that
the APIs are available for Subclipse to start working sanely.

> Now, I am not against expanding the scope of this issue to include
> changing the command line as well. I just think you need to be
> careful about doing so unnecessarily. I would rather see the change
> happen in the libraries and API and let a change to the command line
> UI come out of an actual request for it.

After the APIs have been introduced (including the changes to the C++
and Java parts of JavaHL), the command-line client could be extended
to generate a diff which chops off the part of the path which was used
as the base of the diff operation. That would leave all of the other
outputs the same and I'd like to have the extra option.

The second point definitely isn't a must-have though. The first point *is*.

bye,

Erik.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Jul 19 23:11:51 2007

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.