I don't want to give ssh access to everyone on my server. And even if I did,
we can't do exactly the same, savannah provides a huge infrastructure for
managing ssh users.
Well how about a .htaccess with just one global password which only devs know, and then anonymous access.
Example:
Bradley commits, he would use user name genixpro, password d37g239
Kyle commits, he would use user name appleboy, password d37g239
Joe somebody commits, he would use user name anonymous, password guest
That way you cant have bots defacing the server, as it requires some user interaction, and you can still set user name commits by using that central password. Anyone like joe blogs who commits anonymously, would simply have his work checked by someone in their free time, while someone who commit with a user name like bradley, would not need their work checked to see if its malicious.
Would that system work?
(the only flaw would be having password figured out, but thats still not a big deal since you can change the password then notify everyone, or simply decide whether the person who got the password is worth allowing him to commit).
The fact the host si temporary is a good reason not to use it. Yet the account problem remains.
And even if I find time to do it, the account problem is not solved yet (see below).
See above
I don't know how to setup mercurial but I know reading so I can do it, with enough time.
Well, on my system at least, it was as easy as running urpmi mercurial (sometimes called apt-get, yum, smart, or rpm on other systems), then I could make, edit, commit etc locally. A little port forwarding and configuring, then people outside could have used it :D
I don't want to stop anything, just do it right and not in rush.
True. So.....
I have the feeling you do not realise that pushing work on very loaded people
while discarding as non-existant the issue they presentedis at best very
impolite.
I agree, I'm being hasty. And it could take as long as a few month to setup, which is why CVS might still be used. I was just thinking of it as saving time and effort on converting right after alpha23, when no development is taking place, rather than later on down the track when development could be really busy. It could make things difficult for Bradley and yourself.