Re: Migrating to Community-driven Infrastructure - Step 1: Inventory
 

IMHO it's essential for any software development project.

In general true. Can you comment on the amount of server and network resources garage is using? I suspect (entirely speculatively of course) that it's peanuts compared to eg the builder or the apt repositories.

As for admins, I'm fairly sure we can find enough BOFHs in this community. I'm certainly willing provided that the terms are not too onerous and the result scratches at least some of my itches (read: if support for Diablo is dropped, my interest goes to zero).

Re: Migrating to Community-driven Infrastructure - Step 1: Inventory
 

We have to take care for packages where source and/or
bugtracker link to those garage projects. If we ever move
maemo.org to some other place and maintainers of those
package aren't active anymore I volunteer to maintain
orphaned packages.

nicolai

Re: Migrating to Community-driven Infrastructure - Step 1: Inventory
 

Nicolai, that would be a huge work, thanks for volunteering.

In any case, we will need distribution or section maintainers, in opossition to single package mantainers.

Re: Migrating to Community-driven Infrastructure - Step 1: Inventory
 

Thanks for reminding me about it. You did emphasize that before, when you offered moving the itT forum to maemo.org, so thanks for holding to that promise.

Re: Migrating to Community-driven Infrastructure - Step 1: Inventory
 

OBS is actually already well underway, and a good chunk of the work for Fremantle (using SB2) has been done, but still needs some love. We also still need help with documenting (one volunteer so far) and testing (couple of those). Also, if others have experience in OBS config/migration, as noted there are still issues with Diablo, Chinook, and the like, which could use work.

There are meetings every Monday on IRC (FreeNode, #maemo-meeting @15:00UTC). The OBS wiki page has details and minutes of past meetings, as well as full logs for those that want to join in and/or see where things are.

Re: Migrating to Community-driven Infrastructure - Step 1: Inventory
 

Wonko: could you please update first post pointing to Wiki page and adding links to latest Council meeting minutes (this was a discussed topic) and include reminder about Council meetings every Friday at 18:00 UTC?

I'm on the go with limited connection so I don't have possibility to add links at this moment.

Thank you.

Re: Migrating to Community-driven Infrastructure - Step 1: Inventory
 

Is there a mailing list, or other asynchronous communication medium, for those of us who are occupied with unrelated day jobs at that time?

Re: Migrating to Community-driven Infrastructure - Step 1: Inventory
 

I think Quim's point is that do "we" (as in the maemo community) run our own garage/bugs type system, or instead let individual developers choose to use sourceforge or github or whatever? Personally, I think we should just let people use external hosts for this. Although it's less friendly for individuals (they'll have to do research on what is best and setup more accounts etc.) it's a lot less work for maemo community server admins.

Re: Migrating to Community-driven Infrastructure - Step 1: Inventory
 

I think we have to resign from the garage.maemo.org project hosting, there are better alternatives (e.g. sf.net, github, gitorious). Anyone can choose what he likes. A bugs engine for all the projects is nice, but it's a no must-be.

An autobuilder (or counterpart) is a must-be. Why? Imagine you have gcc 4.6 on your scratchbox, from devel. You compile... and kaboom! The app doesn't run as gcc on device is outdated.

Re: Migrating to Community-driven Infrastructure - Step 1: Inventory
 

I respectfully disagree. For one thing, users shouldn't suddenly have to set up a gazillion third-party accounts in order to participate. Having cohesion in the community infrastructure is important. If "anyone can chose what he likes", where do we draw the line? Why not dump the forum for example as well, surely an argument can be made that it's also very expensive to run and there are better alternatives elsewhere. Why not also dump extras, it eats a lot of bandwidth after all and people can just set up their own repositories on various free hosting services around.

Shouldn't we find out how much such a thing would cost? It may well be that we can't afford to host it (in which case we probably can't afford to host the repositories and autobuilder so we may as well give up), but throwing it out without consideration would be a shame.