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

Re: [PATCH] svn_wc_ventry(_ex) to eliminate heavily repeated code

From: Malcolm Rowe <malcolm-svn-dev_at_farside.org.uk>
Date: 2007-03-19 22:31:30 CET

On Mon, Mar 19, 2007 at 10:24:32PM +0100, Erik Huelsmann wrote:
> >> As far as the first remark is concerned, I'm looking back at the
> >> adjusted patch and I'm wondering: What *is* the difference between 'No
> >> entry found' and "'%s' is unversioned"? Some errors have the 'no entry
> >> found' (which is also a different result code!) but I can't for the
> >> life of me figure out why we would have different result codes for
> >> these situations, especially because they're detected by the same
> >> condition...
> >>
> >> Anybody any ideas?
> >>
> >
> >Is the first perhaps intended for situations where we believe the entry
> >_should_ have been found?
>
> Well, my point is: when we expect to find none, we check the condition
> and decide *not* to error instead of raising one of these 2...
>
> So, in all other situations, we seem to expect an entry...
>

Sorry, I wasn't clear. Is it perhaps intended to be the difference
between:

1. The user specifies a non-existent file on the command-line.
   libsvn_client passes it down to libsvn_wc, which finds it doesn't
   exist, says "Stupid user" to itself and raises '... is unversioned.'

and

2. We hit a situation where the entry _should_ exist, because e.g. we're
   in the middle of installing a new text-base, and libsvn_wc raises the
   'No entry found' error.

Admittedly, if that is the reason, it's a pretty poor one.

Regards,
Malcolm

  • application/pgp-signature attachment: stored
Received on Mon Mar 19 22:31:44 2007

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