I'm wanting to put my website which includes the my html files and PHP
folders from the net (eg Wordpress - weblog system) under Subversion
control.
The html side is straightforward (just import current site), but I want
to keep the locally created HTML and the vendor supplied code in their
own repository folders.
I've set up a repository as per the Subversion book for the vendor drops
of Wordpress and various Wordpress plugins, storing release versions as
tags and my locally tweaked version in (as suggested) "current" instead
of the usual "trunk".
The REPOSITORY:
==============================================================
/Website/Trunk/index.html <--- Original html website in here
+ /images/ <--/
+
/Vendors/Wordpress/current <- v1.5 with my configuration
+ /tags/Release-1.5-Strayhorn
+
+ /WP-Amazon/current/wp_amazon.php <<<1> see website
+ /tags | |
+ | |
+ /WP-Amazon/current/wp_stickypost.php <<<2> layout below
+ /tags
==============================================================
WEBSITE: (would upload an exported version of this)
/-index.html <-- WC of Original HTML website,
+ /images <-- now from svn://repo/website/Trunk
+
/wordpress/index.php <-- WC of /repo/Vendors/Wordpress/current
+ /wp_plugins/wp_amazon.php <--<1> Single Files from
+ /wp_stickpost.php <--<2> different vendors
^^^^^^^^^^^^^^^^^
How do I manage this???
Mixing Files from different Repository branches
===============================================
Wordpress plugins are ".php" files which must all be in
the "/wordpress/wp_plugins" directory (eg "wp_amazon.php" above)
If a file is simply copied from from a working copy of
"wp_amazon/current" into "/wordpress/wp_plugins" it:
1. loses any association with its original repo (and history)
2. becomes part of the main "repo/wordpress/current"
(instead of "/repo/vendorDrop/wp_amazon/current")
===> Without checking out a 12 plugins to their own working copies
and then manually (or with scripts) copying the files
(eg wp_amazon.php --> wordpress/wp_plugins/),
how can this be automated?
If Subversion would let you could check out individual files (and
remember where they came from) this wouldn't be a problem. Generally
complete folder checkout is tidier, but how do I handle a directory
(wp_plugins) containing files of "mixed heritage"?
I'd really appreciate some suggestions on how to organise thing so that
both the HTML, the PHP directories (eg. Wordpress & plugins) can
maintained and updated individually. I'd like to be able to try plugins
and back out (sounds like a branch) or update a plugin and maintain
local mods, preferably without a lot of manual intervention.
Nested Subversion working copies
================================
The main site folder (svn://repo/website) has another svn controlled
folder inside it (the WordPress folder). These are different, one is
from a vendor drop and the other is locally created, so how can I
maintain the separation but avoid subversion getting confused and
trying to include changes in /wordpress when I do a commit from top (/)
of the website. This should commit the html content to /repo/website,
but not do anything changes in /wordpress - these will be committed
separately.
==> On thought is to use externals:
- moving all the html files down one level (say into /site)
- setting up the / level of the site to pull the /site and
/wordpress folders in as svn:externals
- use server redirects so all the html files are still found
(so links from search engines don't give 404 errors)
Any other ways come to mind?
Thanks
Giovanni - www.reflections.co.nz
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Sat Mar 5 02:48:36 2005