Some good suggestions (I think...since I'm not a Subversion user...yet)
and thanks for the reply. However I'm not sure what scenario your
example is showing, but the example that I'm trying to portray is this
(and this is but one example, but the most basic of examples in our
account):
1. A request comes in from our client to change the wording in an html
file (this happens daily, multiple times).
2. Someone from our 'webmaster' team would then 'check out' the .html
file from VSS in the 'system test' environment.
3. The changes are made and the .html file is checked back in. The
customer has a look and approves (or other).
4. The file is then checkout out from the production environment and
the changes are made (or in some cases just overlaid...depending on the
file & situation) and the file is checked back into VSS.
** Only 1 file changed, only 1 file has a new version and a new
timestamp. **
This file was checked out of a 'main' directory that has, say...50 files
in it and 10 subdirectories with 20 files in each of those (so lets say
250 files total). This .html file doesn't need to be 'versioned' with
the rest of the files out there...many of which really don't have
anything to do with this .html file in the example. So I'm still not
seeing the benefit of 'having' to check out all 250 files, version them
all and then check them all back in. Seems like overkill to me, but
then again, perhaps I'm still missing the point somehow.
-Gary
-----Original Message-----
From: Mark Shead [mailto:sheadm@optimalinternet.com]
Sent: Monday, November 14, 2005 1:57 PM
To: List Subversion
Subject: RE: RE: Re: Potential New User questions
It seems strange that your developers would regularly delete their
entire workspace in order to check out files once at a time. I don't
see how they could test any of their changes on their local machines
before doing a commit if they only had one file at a time. Even if
someone is working with .html they are usually going to need the images
folder to be able to see their changes.
One of the reasons that you can't get an individual file (I'm assuming)
is because subversion keeps all the local information about versioning
in .svn in the top level checked out directory. If a user pulls down a
file to just a random directory, the .svn directory is going to possibly
conflict with the .svn directories for other files that have been
individually checked out like that. By dealing with everything on the
directory level, subversion can check to make sure it isn't writing over
another working copy. There are other ways to do this of course, but
using a .svn directory seems to be a very good method that will work on
any operating system.
You *can* get individual files by doing something like this:
svn cat https://svn.something.com/trunk/file.txt > file.txt
However they are no longer under version control.
You also can check back in individual files by doing:
svn commit filename.txt
But they have to be in a working copy for this to work.
Generally if you are dealing with text documents, there isn't much
overhead in downloading a directory instead of just one file. Don't
forget that you can use -N to tell the checkout to just checkout files
and not subdirectories.
If checking out an entire directory seems to inconvenient, you might
give it a try and see if it really takes any more time than what you are
doing now. For example:
svn co http://server/repo/dir
cd dir
vim file.txt
svn commit -m "fixed every bug"
Is two less keystrokes than what you're wanting:
svn co http://sever/repo/dir/file.txt
vim file.txt
svn commit -m "fixed every bug"
If you are trying to checkout files over your cellphone connection or
something like that, the individual file checkout will be faster, but
for most use their isn't much of a speed difference unless you are
dealing with large binary files.
--Mark
-----Original Message-----
From: Ryan Schmidt [mailto:subversion-2005@ryandesign.com]
Sent: Sunday, November 13, 2005 8:59 AM
To: Theisen, Gary
Cc: List Subversion
Subject: Re: Potential New User questions
On Nov 12, 2005, at 00:12, Theisen, Gary wrote:
> And that is primarily what our team would be using Subversion for.
> 'Just' version control, check in check out control, lock file if
> checked out etc. Often times, we are dealing with just 1 .asp or 1
> .html page and don't need to reversion or check out the entire
> directory that that
> 1 file lives in. I'd like to see the option in Subversion, to just
> check out and check in and reversion 1 file (or 2 files, etc) instead
> of said mentioned entire directory.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Wed Nov 16 02:36:44 2005