You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Ralf Stephan <ra...@ark.in-berlin.de> on 2006/08/25 14:39:18 UTC

cannot commit property data

Hello,
there is one behaviour of subversion which I think is a bug, or
at least, it should be documented in the FAQ and the book, as it
can confuse people a lot. To reproduce:

1. check out a fresh copy of any repo
2. change something in a directory, i.e., svn add file
3. svn ci file
4. svn propedit svn:ignore . (the same directory you changed in 2)
5. svn ci .
Output: Commit failed. Out of date.
    Comment: I know it's not correct but it happened to me when I knew
    exactly that noone else had committed between 3 and 5.

Why should svn need to get some info back from the server it already
has? I know I should always svn up before any commit but when you
are the only committer you don't expect such an error.

But even if this is my error, it should be clearer what I did wrong,
e.g., by giving an informative error message like "The server has
received a directory related commit earlier, and you better do a
svn up first before committing something to the directory."

And even if that is not possible then this case should be handled in
the book and the FAQ. Why? Because I'm not the first to get bitten
by it:

http://svn.haxx.se/dev/archive-2003-04/1238.shtml

So please, people, do something such that the next victim of this
misdesign will not sit there and, like me, ask himself "What was
wrong with that?"

Thanks for your time,

ralf

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: cannot commit property data

Posted by Marcus Rueckert <da...@web.de>.
On 2006-08-27 16:26:32 +0200, Ralf Stephan wrote:
> So, I didn't find it when I needed the info. I would have expected
> such information in the chapter about props, and I'm proposing to
> put it there, too. If these are the only exceptions, people won't
> complain about reading about them twice.
> 
> > The error message, I hope, at least says *which* item is out of date.
> > Does it not?
> 
> It does. So I studied the property chapter.

why only for properties?
files/directories can be out of date too.

darix

-- 
           openSUSE - SUSE Linux is my linux
               openSUSE is good for you
                   www.opensuse.org

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: cannot commit property data

Posted by Ben Collins-Sussman <su...@red-bean.com>.
Seems like it would be a good idea to add a cross-reference to the
directory property limitation info in chapter 7.  Thanks for the
suggestion.

It's hard to write a book which (a) reads straight through  from start
to finish, logically building on itself, and (b) acts as a reference
manual, so that people can just open up to any point in the book and
get a full picture of what's going on.  In some sense, those are
contradictory goals, so we're always in a constant state of
compromise.

On 8/27/06, Ralf Stephan <ra...@ark.in-berlin.de> wrote:
> You wrote
> > On 8/25/06, Ralf Stephan <ra...@ark.in-berlin.de> wrote:
> >
> > >Why should svn need to get some info back from the server it already
> > >has? I know I should always svn up before any commit but when you
> > >are the only committer you don't expect such an error.
> >
> > Please read chapter 2 in the svn book, 'basic concepts'.  In
> > particular, see the section at the end about mixed revision working
> > copies.  You cannot commit a propchange to a directory unless its
> > local revision is equal to HEAD.  That's one of the only two
> > restrictions that exist:
>
> So, I didn't find it when I needed the info. I would have expected
> such information in the chapter about props, and I'm proposing to
> put it there, too. If these are the only exceptions, people won't
> complain about reading about them twice.
>
> > The error message, I hope, at least says *which* item is out of date.
> > Does it not?
>
> It does. So I studied the property chapter.
>
>
> Regards,
> ralf
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: cannot commit property data

Posted by Ralf Stephan <ra...@ark.in-berlin.de>.
You wrote 
> On 8/25/06, Ralf Stephan <ra...@ark.in-berlin.de> wrote:
> 
> >Why should svn need to get some info back from the server it already
> >has? I know I should always svn up before any commit but when you
> >are the only committer you don't expect such an error.
> 
> Please read chapter 2 in the svn book, 'basic concepts'.  In
> particular, see the section at the end about mixed revision working
> copies.  You cannot commit a propchange to a directory unless its
> local revision is equal to HEAD.  That's one of the only two
> restrictions that exist:

So, I didn't find it when I needed the info. I would have expected
such information in the chapter about props, and I'm proposing to
put it there, too. If these are the only exceptions, people won't
complain about reading about them twice.

> The error message, I hope, at least says *which* item is out of date.
> Does it not?

It does. So I studied the property chapter.


Regards,
ralf

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: cannot commit property data

Posted by Ben Collins-Sussman <su...@red-bean.com>.
On 8/25/06, Ralf Stephan <ra...@ark.in-berlin.de> wrote:

> Why should svn need to get some info back from the server it already
> has? I know I should always svn up before any commit but when you
> are the only committer you don't expect such an error.

Please read chapter 2 in the svn book, 'basic concepts'.  In
particular, see the section at the end about mixed revision working
copies.  You cannot commit a propchange to a directory unless its
local revision is equal to HEAD.  That's one of the only two
restrictions that exist:

http://svnbook.red-bean.com/nightly/en/svn.basic.in-action.html#svn.basic.in-action.mixedrevs

>
> But even if this is my error, it should be clearer what I did wrong,
> e.g., by giving an informative error message like "The server has
> received a directory related commit earlier, and you better do a
> svn up first before committing something to the directory."

Again, if you don't know what the phrase "out of date" means, you
might want to read the 'basic concepts' chapter.  It's fundamental
terminology used all over subversion.  Users are going to see that
phrase appear when trying to commit files as well.

The error message, I hope, at least says *which* item is out of date.
Does it not?

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org