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

Re: RFC: Untangling the peg revision knot (was: A preliminary study of non-contiguous transformations in the Hilbert space of Alexandrian solutions)

From: Mark Phippard <markphip_at_gmail.com>
Date: Wed, 26 Nov 2008 17:37:35 -0500

On Wed, Nov 26, 2008 at 4:26 PM, Branko Èibej <brane_at_xbc.nu> wrote:
> Mark Phippard wrote:
>> It feels like you are just grasping at straws now, because regardless
>> as to whether users EVER understand this feature or whether we decide
>> to change the default for URL's it will always make sense for @BASE to
>> be the default for a WC.
>>
>
> So you're never confused by "svn log" not showing changes more recent
> than those in your WC? That being just an example of the problem and not
> at all grasping at straws.

Where I feel you are grasping is that I do not see how your example is
related to the peg revision. You are now complaining about the
operative revision range used by the command and that is at the least
a different usability complaint.

A URL without a peg revision is ambiguous, so there is debate over
what to use for a peg revision. A path in a WC is not ambiguous as
that path always corresponds to a specific URL and revision from the
repository. So I do not see how we could use some other value when
operating on the WC,

>> My position is simple. It is easy for two users to have differing
>> opinions on what the best default is when dealing with URL's and to
>> say one is right and the other is wrong is just silly. I would not
>> care if we had chosen some different default, but I do believe that
>> the default should be consistent regardless of the command.
>>
>
> That is a logical position for someone who's focused on SVN as an API.
> An SVN client is not an API, it's a user interface. Anyone who's ever
> dealt with usability will tell you that in this context the definition
> of "consistent" and "obvious" may be quite different from yours. What
> works best in a UI is often not obvious at first; I can't remember if I
> had reservations about the defaults when peg revisions were introduced,
> but I do have now that we've been using them for a few years.

You seem to be taking the position that there is only one right answer
here and it is yours. I think I have provided several compelling
examples of why there is not a single right answer. If we cannot even
agree on that point then I do not see where we can go from here.

> Like in your example from a few posts earlier, about getting revision
> and path from "log -v"; I was quite surprised that you focused on one
> file; the case I see used most often is finding some a bunch of related
> changes on some possibly non-existend branch by doing an "svn ls -v
> http://repos/branches" and picking the branch name from what's in the
> "-v" part of the log output.

I am sure I could come up with plenty of valid examples where your
preferred default is better and I know I can provide more where HEAD
is better. That has been my point from the first post, there there is
no single right answer.

-- 
Thanks
Mark Phippard
http://markphip.blogspot.com/
Received on 2008-11-26 23:37:49 CET

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.