That looks interesting, but it doesn't work for me (I'm not that
familiar with shell scripting, so forgive my newbie
misunderstandings...)
I can get
find . -type d -name '*.svn*' -exec svn rm {} \;
launched from the root of the working copy to work though. Is this
equivalent? (Like I said, I REALLY don't know enough about shell
scripting; looking at the command below, I don't see how grep figures
out which files to work on, and I don't know those sed commands at all)
Cem Karan
Guy who really needs to spend some time learking more about shell
scripting...
> -----Original Message-----
> From: Roland Besserer [mailto:roland@motorola.com]
> Sent: Wednesday, December 08, 2004 1:47 PM
> To: users@subversion.tigris.org
> Subject: Re: Committing .svn dirs and your're toast?
>
>
>
> Well, I didn't mean 'manually' in the literal sense :-)
>
> With 97 directories to remove, doing it your way would have
> caused 97 revision increments - nasty. I did it slightly
> different by piping the check-in message (essentially a svn
> list output generate by a post commit script) into:
>
> grep '.svn/$' | sed '/A /$REPOPATH/' | xargs svn rm
>
> which runs a single 'svn rm' command to remove all 97
> directories. Well, depending on your implementation of xargs
> and the total size of the argument list, xargs may split it
> into multiple invokations (3 in my case because the $REPOPATH
> for each file name argument is pretty long).
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Thu Dec 9 14:03:27 2004