Randy McMurchy wrote: > Then, after Trac proves itself a success as a Bugzilla replacement, > another component of our on-line presence can be migrated over, if it > is deemed that we should go that route.
Let me try to put some of your fears to rest. The source viewing capacity of Trac is so much an inherent part of the Trac system, that I don't think I can even set up the environment *without* having it use that component. (The python setup script bails if you enter an invalid Subversion repository, and I'm not about to start hacking it.) So, the ViewCVS functionality is in by default, and from what I can tell, that part of Trac doesn't make too much of a hit on the server. Next is the Ticketing system, the replacement for Bugzilla. It won't be much of a tracking system at all if we don't look at using this from the start, and as you and others have already expressed, our current Bugzilla setup is horrible to work with. Apart from the fact that I tested Trac on Anduin, which would have helped Trac's processing time, Trac uses SQLite as a backend. I'm inclined to believe that is a mite faster than Bugzilla's MySQL database, but I have no raw data to confirm that. Anyway, yes, I think we should also add in the Bugzilla replacement from the start. The Wiki is another Trac built-in. It's probably more core to Trac than the source browser is. So, no matter what, we'll also have a Wiki. It seems to me that where the concern comes in is in using the Wiki as the main site. I've stated my opinions on this, and have offered (apparently less than compelling) reasons for using Trac as the site. In any case, I *have* taken the concerns and fears of some of the community into consideration, and hopefully, there is a sort of compromise solution here. We can get Trac up and running almost immediately and begin using the Wiki, the Source Browser, and the Ticketing system. We just have to upgrade Subversion, first. Once it's up and we're getting used to that, I would like to start looking at what it would take to port the site over. That part isn't going to happen overnight, there would be some scripting to be done to get it all working totally smoothly, especially in connection with the mirrors. I'm hoping that as I find solutions for the challenges there, and as people get used to using Trac, that the idea of having it be the all-in-one won't sound so... scary, I guess... -- JH -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
