Thanks for a answer.
There is some problems using your idea.
1) RSYNC is an Unix only program, it can't be used by Windows. Both of my
servers are windows based.
2) The servers aren't on the same network so I have to use somthing like FTP
to transfer files.
I thought to use post commit script to send the file over the ftp, but the
post comit scrip is provided only with the folloing arguments "the path to
the repository, and the new revision number that was created" (accoring to
Is there any way to know wich files were changed and to transfer them
maintaing the directory sructure?
On 3/21/06, Ryan Schmidt <subversion-2006Q1@ryandesign.com> wrote:
> On Mar 21, 2006, at 10:42, MrBr @ GMail wrote:
> > I want to update my web server with the newest revision after
> > committing a revision.
> > The problem is that keeping a working copy on the web server in not
> > an option.
> > There should be only the source code and nothing more.
> > The web server is off site.
> > The SVN server is windows+svn+Apache
> So your repository server and your web server are two different
> machines. No problem. In your post-commit hook, you could use "svn
> export" to create an export of the HEAD of the repository, somewhere
> on the repository server. You could then use rsync to send only the
> relevant changes over to the web server. You would need to either do
> this entire task before exiting the post-commit hook, which may take
> some time and annoy your committers, or you need to spawn off another
> task in the post-commit hook, but if you do that, you must handle
> concurrency issues. (What happens if two people commit changes very
> close together such that the second commit happens while the update
> launched by the first commit is still running?)
> That would work well on Linux/Unix/*BSD/OSX machines; unfortunately I
> have no knowledge of what will or will not work on a Windows machine.
Received on Tue Mar 21 11:30:02 2006