[dev] [commits] Horde-Hatchery branch master updated. d3fcd5cdd11ada51d44c03fccd529ba9beed543d

[Ben Klang]

On Jan 13, 2010, at 12:33 PM, Michael M Slusarz wrote:

> If you mean the fact that there would not be a default index.php at the base of the webroot, e.g. http://www.example.com/index.php(since horde would be installed in http://www.example.com/horde/), I think this is completely irrelevant.  It is my strongly held belief that there is no reason why we should make the git repo a "drop-in" installation of horde.  If someone does wants to run using the git repo (realistically, this will be a handful of people - and those people will mostly be horde devs), they either have to accept that the apps won't live underneath horde or they will have to manually symlink the directories.  Again, not a big deal.


I disagree with this statement.  While I don't necessarily feel that a git clone should work 100% out-of-the-box, it should be close.  By making the repository close to a workable Horde install we encourage casual developers and minimize our own work on new installs.  Right now there is a fairly high barrier to begin working on Horde, even for someone like me who is intimately familiar with the process.  I can only imagine what it would be like for someone who may want to fix a bug but has only ever installed Horde from the tarballs.

That being said, there are naturally some concessions that we will need to make to the development and packaging processes that will conflict with Horde working fresh from a git checkout.  But in my opinion, the closer we can align those needs, the better.

/BAK/
-- 
Ben Klang
Alkaloid Networks LLC
ben at alkaloid.net
404.475.4850
http://projects.alkaloid.net