> >NPU checks out a source tree. If the working copy is
> >pristine, NPU says: "Ok, these are all of my files as I'm used
> >to. There is nothing there that I don't understand."
>
> NPU then performs any of the operations that don't entirely
> work with the "separate trees" arrangement, but do work with
> the "meta-data in working copy" arrangement, such as copying
> the working copy to another location or computer.
>
> SVN "breaks."
>
> NPU calls PU and says "I did thus and so, and it broke.
> What's wrong? How do I fix it?"
I see your point now. However, let me offer that in all of the
SCM user training that I have conducted, I have so far been able
to get across the message that users cannot simply move/copy
working directories. During my stints as de facto SCM admin, I
don't remember any support requests to this effect. However,
this is just my experience, and your mileage may vary.
My problem with .svn/ directories is more than the exposure of
SVN internals to the user. It is also an impedient to PUs who
wish to run text processing tools on their codebases. I like
being able to do a 'find . -type f | xargs ...' without having to
worry about special meta files being picked up. The crux here is
that I, and presumably others, already have a bunch of scripts
that would have to be retrofitted to accommodate the special
directory.
> PU says "Hmm ... I do that all the time, it works for me ... oh,
> wait, where's your meta-data stored?"
>
> NPU's eyes roll into the back of his head, his palms get sweaty, and
> he considers a career in underwater basket weaving!
I suppose that the reference to UBW was inevitable. ;-)
Kumaran
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed Aug 27 19:19:38 2003