[petsc-dev] please make myself and Satish admin on bitbucket.org/petsc

[Barry Smith]

On Feb 10, 2012, at 6:41 PM, Matthew Knepley wrote:

> On Fri, Feb 10, 2012 at 6:39 PM, Barry Smith <bsmith at mcs.anl.gov> wrote:
> 
> 
>  Not true. Make a clone of bitbucket.org/petsc/petsc.bitbucket.org make a change to it and try to push, you cannot. You need to send email to Sean asking him for write permission to that repository.  So cannot access repositories in petsc that Sean has not explicitly given me access to.
> 
> Of course you can't write to a repository without being given access. Probably because its always someone else
> that does it, you do not remember having someone mail there SSH key, editing the .ssh/authorized_keys file,
> sticking in the correct string with the hacked up script in it.

   The point is that that permission is on a repository by repository bases, so every time Sean (who is the only one who can do it) adds a repository he has to remember to set the permissions for all of us.

> 
> Can you really be making this argument?

    Yes

> Are you purposely not looking at how this works?

    It doesn't work in a reasonable way. It is all designed around hosting one or two repositories in an account and assuming that it is those 2 static repositories forever. Seems to be ok for that. But that is not all of what petsc.cs.itt.edu does, 

    So you could argue; there is not downside to moving petsc-dev and BuildSystem to bitbucket. Absent everything else in the world, fine I buy that. But we are not going to keep two repositories on Bitbucket and everything else on petsc.cs we need to host everything in the same place using the same procedures otherwise moving two repositories to bitbucket brings more headaches of two systems instead of one. And the more I look the more cumbersome the procedures at bitbucket beyond just holding petsc-dev and buildsystem look. 


   Barry

   
> 
>    Matt
>  
> Or try
> bsmith-laptop:Src barrysmith$ hg clone joe ssh://bitbucket.org/petsc/joe
> running ssh bitbucket.org "hg init petsc/joe"
> The authenticity of host 'bitbucket.org (207.223.240.182)' can't be established.
> RSA key fingerprint is 97:8c:1b:f2:6f:14:6b:5c:3b:ec:aa:46:46:74:7c:40.
> Are you sure you want to continue connecting (yes/no)? yes
> Warning: Permanently added 'bitbucket.org,207.223.240.182' (RSA) to the list of known hosts.
> Permission denied (publickey).
> abort: could not create remote repo!
> So cannot make new repositories in petsc, again need to ask Sean to do it for me.
> 
> Interestingly both of these things work fine and out of the box on petsc.cs.itt.edu
> 
> This is why we didn't just blindly switch to bitbucket the other night. Lots of things need to be worked out before we can switch.
> 
> 
> It also appears that if Sean makes it so multiple ones of us can log into the petsc account on bitbucket there will be no record of who made what changes, it will behave like an old fashioned and frowned upon shared account.
> 
> 
>  Barry
> 
> 
> 
> 
> 
>  Barry
> 
> 
> 
> >
> >    Matt
> >
> > >
> > > https://github.com/blog/674-introducing-organizations (blog)
> > >
> > > https://github.com/enthought (example)
> > >
> > >
> > > Bitbucket has a thing called "groups", but it's not really the same.
> >
> >
> >
> >
> > --
> > What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.
> > -- Norbert Wiener
> 
> 
> 
> 
> -- 
> What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.
> -- Norbert Wiener