I am NOT modifying the commitment transaction in a hook script. I am
merely rejecting a commitment transaction that doesn't meet my
specifications. It is up to the user to "fix' the problem, then
preform another commit.
I do this right now with non-revision properties. For example, we have
Unix shell scripts, but burn the CDs for these Unix shell scripts on a
Windows PC. If I do a checkout from the PC, the eol character is CRLF
making these scripts not work on Unix. To get around this problem, I
have a hook script that prevents a commitment from taking place if a
Unix shell script doesn't have the property svn:eol-style = LF.
I'd like to extend this to revision properties. In fact, my hook
script works great. For example, I tell it to reject any transaction
where the revision property bugnum is not set to a numeric value. And,
sure enough, that transaction is rejected!
The problem is that a user can't set the revision property bugnum
until the transaction has actually been committed. My question is why
not? A user can set a property on a file that isn't even in the
archive (when you do a "svn add" for the first time). Why not allow a
user to set a revision-property before that revision is committed to
On 6/20/05, Andrew Thompson <email@example.com> wrote:
> David Weintraub wrote:
> > However, the "svn propset" command doesn't allow me to set the value
> > of a revision property until the revision itself has been committed.
> > Is this an oversight? Is it considered a bug, or is there a reason why
> > it was decided that users should not set revision properties before a
> > commit takes place?
> Click: http://svnbook.red-bean.com/en/1.1/ch05s02.html#svn-ch-5-sect-2.1
> Scroll down to the pretty pink Warning box.
> See also:
> (just a few random related links)
> Andrew Thompson
To unsubscribe, e-mail: firstname.lastname@example.org
For additional commands, e-mail: email@example.com
Received on Tue Jun 21 04:35:47 2005