subversion-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From David Burleson <>
Subject Re: Ideal Subversion Setup
Date Thu, 17 Dec 2009 09:53:45 GMT
Hi Ryan,

Thanks for the reply. So there might be some hope.

Let me elaborate on our setup. I've decided the first thing to do is to 
cut out the inbetween step of dragging files over. So I have setup a 
folder on our local linux webserver and setup apache web server to use 
it as a virtual host (to do testing). This works great. Now I need to 
sort out the symlinks for the images directory. We create the symlinks 
manually through command line (ssh). We are using TortoiseSVN as a way 
to create the checkout to be used on our Windows computers. So....

1. Linux Server: create a directory for development area
2. Linux Server: setup apache to use the newly created directory as a 
virtual host
3. Windows PC: Browse to the newly created directory via the network
4. Windows PC: Right click on the directory and select SVNcheckout and 
checkout the repository
5. Linux Server: Create a symlink in the directory to another directory 
on the linux server do I commit that symlink as a symlink and not commit the 
contents within it?

And, what is the best way to preserve folder/dir permissions using chmod 
into a repository?

Cheers for the help so far!


On 17/12/2009 03:33 AM, Ryan Schmidt wrote:
> On Dec 16, 2009, at 10:00, David Burleson wrote:
>> I have been using subversion with TortoiseSVN for a couple of years now. I work in
a team of 3 web developers on multiple websites. Im starting to wonder if the way we use subversion
and version control is the correct way. So, I have posted to ask advice on the best way to
use Subversion in a web development team with multiple projects.
>> We currently each have our own development area on a local web server for each project.
We also each have our own SVNcheckout of each project.
> Sounds good to me!
>> Once we have checked out/updated our 'repo' we drag the contents over to our development
area to work. Once we are finished working, we drag the files back over to our 'repo' and
commit it.
> Where you say "repo" you actually mean "working copy" (or "checkout"); the "repo" or
repository is what you check out from and commit to, and there is only one of them, on your
Subversion server.
>> I have a feeling that the better process is to make your development area the SVNcheckout.
> I would completely agree with that. You're making a lot of unnecessary error-prone work
for yourselves by manually moving files around.
>> My only problems is how subversion and TortoiseSVN handle symlinks and file/dir permissions.
We have a couple of symlinks for folders like 'images' so we don't have to duplicate the directories
on the webserver and some folders which we upload files too have 777 permissions and what
not. I don't know if subversion or TortoiseSVN will pass these own and treat the symlink as
a symlink and not a folder, and pass the permission into the 'repo' too.
> Subversion handles symlinks just fine, but Windows does not. (There is no such thing
as a symlink on Windows.) So I'm not clear on how you're using symlinks on Windows at all,
unless you're using cygwin. In which case you should use a cygwin Subversion client, not a
Windows Subversion client like TortoiseSVN.
> Subversion repositories do not store or retain file or directory permissions or ownership.
The only exception is the execute bit on files, which can be preserved by setting the svn:executable
property of that file to any value (which Subversion automatically does if you add a file
that's executable).

View raw message