At 15:56 Uhr +0100 25.11.2002, David wrote:
I am pretty sure we can do it... not that I like the idea of doing it.-----BEGIN PGP SIGNED MESSAGE----- Hash: RIPEMD160On Montag, November 25, 2002, at 03:30 Uhr, David R. Morrison wrote:Well, I cannot relate to that, but Ia m sure, that it is an honour. There are issues with Sourceforge which will always hamper our further development. Questions like "can we run perl cgi-scripts" or not, will never arise on our own Server.Seems to me that sourceforge is serving our needs quite well, and I for one am proud that we were selected as their project of the month.
And actually, having a team that works hard to maintain a server for me, so that I can concentrate of the actual issues, instead of having to bother maintaingin a server, seems like a boon to me.
Furthermore mirroring binary distribution sets will always be a pain, I do not know if you read the request of a large university to mirror or binary packages. This is not easily possible, since there is no rsync provided by sourceforge and special arrangements would have to be made.So you did actually ask the SourceForge stuff? Or are you guessing? We did allow for rsync mirroring our bindist in the past, when it was still on the SF.net servers, BTW.
I'd be interested to get the link to the suppor request tracker item where the SF.net team rejected the request for rsync of our stuff.
Personally I dislike the single location with sourceforge and the fact that you are somewhat bound to their limit system.Ah, that's what it boils down to (like with 99% of the people I met who diss SF.net) - personal dislike :-)
We can, if we think it's useful, still have an own server, to host email addresses. Or other things, we have to... But right now, I am not willing to give up on SF.net that easily. I am very very happy with the services they have been and are still providing to us.I can only agree and that is my strongest point. I wish to organise fink and that requires a lot of voluntary work. Work which no developer likes to do, because it is tedious and simply annoying, a lot like the paper work we all like to avoid if possible. I will try to find volunteers and I will try to sort them into groups with the permission of all of you. I will try to coordinate them and I will even try to recruit them. Yet a fink server, where I can create my own little word for them, using squiremail or horde, giving them their own email, etc, will make it a lot easier on you and me.That being said, the biggest need that we have IMO is for more volunteers, since there are a number of tasks that aren't get done as regularly or as systematically as they might. We could have a "monitor the package submission tracker" committee, a "build distributions regularly" committee, a "sort the submitted bugs and assign them" committee, and so on.
I have watched projects in the past from SF.net, claiming they did not get all they needed. Dillo for example - just take a look at their homepage and have a good laugh about the help plea on it.
A server we host on our own also means we have to maintain it on our own. Somebody has to pay the bandwidth, for example. You know what was the main reasons SF.net asked us to move the bindist from the webservers? No, not the disk space (a couple GB), it was the fact that we incurred multiple GB/day of load. So if we host our own bindist server, be prepared to pay for that.
One option I see, if it turns out that indeed our bindist can't be rsynced to external mirrors, though: we could setup our own server to host a master copy of the bindist; we'd still rsync it to the SF.net servers, and mirrors could use it, but we'd try to restrict access to it otherwise (e.g. by letting the apt config files point to the SF.net mirrors by default, as before).
Yes it is a personal issue, and I wonder on what you base your claim that " many are a bit hesitant with sourceforge". I have only encountered a handful of people for whom it was true, in all cases when pressed it boiled down to personal dislike.I think there are enough interested people, but many are a bit hesitant with sourceforge, just like me. I hated the idea of becoming part of their net, just to submit to fink. But once more, that is a personal issue and I cannot speak for others.The problem is not with resources like mirrors, but with finding enough interested people.
I have a personal like for them, and not just because I am personal friend to some of the staff, but that does figure in, too, I admit that. It also means I never worry about getting good support (and those people I talked to on complained about bad support from SF.net so far always were basing it about experiences they made over a year ago before SF.net restructured their QA).
As for mirrors, I think it is important to gather an infrastructure now and not when it is nearly impossible to let those who are interested grow with us. It make a difference if you ask someone "Hey want to mirror 500MB" and they grow with us , instead of asking "hey , want to mirror 4Gig".Ah come one, that argument is a bit nonsense. That doesn't make such a big difference, if they are concerned about the bandwidth used to create the mirror, they can easily spread out the mirror process to last several days/weeks/months.
I agree it won't hurt us to get mirrors, but it's not as pressing ultimate important, in my eyes.
Max
--
-----------------------------------------------
Max Horn
Software Developer
email: <mailto:[EMAIL PROTECTED]>
phone: (+49) 6151-494890
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
Fink-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-devel
