Drupal and Subversion

public
howardowenscom - Wed, 2007-12-19 22:52

One of my developers has put up a blog post explaining how he's using Drupal and Subversion.

http://tinyurl.com/3arfj3

For anyone interested.

Aww, I was hoping it was

eli's picture
eli - Thu, 2007-12-20 02:48

Aww, I was hoping it was some kind of incredibly clever way of dealing with DB changes in SVN.

Still, good advice for people who want to keep their modules running through CVS (or if you're building a project on HEAD!)


Ah - yes - I'm still trying

nicksergeant@drupal.org's picture
nicksergeant@dr... - Thu, 2007-12-20 03:19

Ah - yes - I'm still trying to figure out the best solution for managing DB changes as well. Currently, I do a dump of my entire DB every night, but that's obviously only manageable at the current moment because it's only about 2.5MB.

I've seen some posts floating around about how best to manage DB changes with Drupal - I'll see if I can scrounge them together and post here.

http://www.nicksergeant.com


SQLite

Garrett Albright - Thu, 2007-12-20 16:04

I'm (probably fruitlessly) hoping that Drupal can gain highly-compatible SQLite support in the near future for this reason, among several others. Since SQLite stores its data in a single file which can be located anywhere on the filesystem, it lends itself well to being used with version control systems, especially if they can handle binary diffs (I like Mercurial).

Maybe I'll end up working on it myself once I'm done with my current project…

show me the diff

moshe weitzman's picture
moshe weitzman - Fri, 2007-12-21 02:06

i'd like to see what such a diff looks like. if it looks at all readable, thats pretty compelling argument for sqlite


Not readable

Garrett Albright - Fri, 2007-12-21 16:37

A binary diff is most certainly not human readable, but that's not the point. The point is that the version control system can back up or synch binary files just as easy as they can with text files, because they only have to worry about the parts of the data that has changed. For highly-structured data such as in a database, this can save a lot of space and time.

Systems which cannot handle binary diffs (I don't think CVS can…) have to make entire copies of changed binary files.