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

Re: svn commit: r31511 - in branches/tree-conflicts: . subversion/include subversion/libsvn_wc subversion/svn subversion/tests/libsvn_wc

From: Julian Foad <julianfoad_at_btopenworld.com>
Date: Thu, 29 May 2008 15:53:23 +0100

Stefan Sperling wrote:
> On Thu, May 29, 2008 at 02:14:55PM +0100, Julian Foad wrote:
>
>>Sorry if this feels like I'm overriding some of your decisions. Let me
>>know if my reasoning makes sense.
>
>
> It does make sense. However, we already know that people want
> tree conflicts to work with TortoiseSVN. Now, the Tortoise developers
> could either invent their own human readable descriptions, or use an
> API we provide (and which our command line client is using).

True. I just think it is highly unlikely that Tortoise or any GUI would want to
use the exact same text that "svn" uses. As a user of Tortoise, I would want to
see the conflicts in a multi-column list or something like that rather than as
long strings of plain text. Even if there was to be a long-text pop-up
description when I hover my mouse pointer over a conflict in the list, I still
think it is unlikely that I'd want the wording of that description to be the
same as what "svn" gives.

But only TortoiseSVN or other client authors could definitively answer this.

It's always a good idea to treat duplication of data with suspicion and try to
eliminate it, as you are doing. I don't know quite how I think this sort of
situation is different, but I do. The same goes for separating the definitions
of the words in the WC entries file from the definitions of the words used for
XML output. In both cases, one thing that makes it right to separate them is
that the information is logically private in its domain and there is no logical
reason why the corresponding information in the other domain should be the same.

> For consistency, I'd like to provide them with an API they can use,
> IFF we decide that having our core libraries providing some form of
> human readable representation of tree conflicts is a good idea.

It will be a good idea for our libraries to provide it IF AND WHEN we know that
somebody wants it.

> So far I had assumed this was the case, but I don't mind reverting
> that decision. In which case I'd favour moving the human readable
> descriptions out of the core libraries entirely, into our command
> line client -- libsvn_subr/cmdline.c might be an option.

In r31511 on the branch I moved it out of the libraries and into the
command-line client: into "subversion/svn/tree-conflicts.[ch]".

Regards,
- Julian

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-05-29 16:53:44 CEST

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