I would also like to see a more structured release program. It's kind of scary to tell people that they should "just use the latest CVS code".
For testing and development, this isn't a bad thing - as long as the trunk codebase generally *compiles* and *runs* more often than not. I've had no problems compiling the latest snapshots, but running it seems to lead me down a path of frustration - nothing seemed to work as the 0.4.0 release did.
For production, a stable release cycle really would be nice. Particularly with a patch history along the stable tree until a next release (0.4.1, 0.4.2, etc, until the 0.5.0 development tree is deemed adequate for a feature freeze to 0.5.0).
That's where consultants earn their money. They
should be preforming some kind of quality control. You build the code, get it to work, test it and ONLY
then install it at a customer's site for final
testing. If you don't have a consultent then you do
this kind of work yourself.
In the OpenSource world, consultants typically implement stable releases of packages - typically for bug reporting if for no other reason. It's difficult to explain to someone why your CVS checkout from 3 months ago exhibits some unexpected flaw - who out there is going to have the same snapshot installed to compare notes?
You are right. It would be stupid to install a new untested CVS download on a working PBX system.
Granted. It would also be stupid to dictate a code-fork of the Asterisk source base just to have a "stable" reference tree.
I'm partial to the Mozilla approach. New development in trunk, branch stable releases with fixes rolled back into trunk as appropriate.
How many folks are running 2.6.0preX series Linux kernels on their production servers? This isn't much different, really. I don't mind rolling in a half dozen backports to patch up a 2.4 tree to something I can use, and *depend* on.
In the end, I'm really just happy to have something of Asterisk's quality in the OpenSource community. Despite a few minor quirks, 0.4.0 really does seem to work quite well.
-- - Ian C. Blenke <[EMAIL PROTECTED]> (This message bound by the following: http://www.nks.net/email_disclaimer.html)
_______________________________________________ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users
