Re: [Simulavr-devel] Discussion: how to proceed withthedeveloment on GIT repo?

[Onno Kortmann]

Hi,
> To convert the current CVS repo, why would Onno's repo be involved?
Just a very short note here. Thomas already said it, but I think I should 
clearly state it again, as it think there is some confusion about what 
essentially the git repo contains:

There is a branch called cvs-upstream in the github master. It reflects the 
COMPLETE, imported repository of simulavr_XX_. If new changes are in CVS, 
they need to be imported using the available script. As of now, I don't see 
any.

The latest changes from this branch are not yet completely *integrated* into 
the master branch on github (which is where the latest stuff from Thomas and 
me resides), but it is full and complete in the git repo! Nothing is missing 
from simulavr_XX_.

The git repo as it is *can be directly used* as a replacement for the CVS one 
(except for simulavr, but which is easy to care for and as I said I volunteer 
to import the whole history from CVS). It is just that people need to have a 
look. Really.

Before a discussion starts about how to set up a server or similar to do the 
comparison to SVN, *please* try to check out the git repository on github.com 
and get a picture of what is possible and already done with the git 
repository, it should make discussion of SVN superfluus:

1. Clone the full history of simulavrxx in git

$ git clone   git://github.com/onnokort/simulavrxx.git

And then, to look at the various branches of development, go to the new 
simulavrxx directory and issue:

$ git branch -a

gives you a list of all branches available. Do 

$ git checkout -b <a-new-name> <any-branch-from-the-above-list>

to create a local branch where you can look at the particular branch from 
above. The <a-new-name> is an arbitrary identification string.

$ git log

gives the available history for the current branch. It is of course easily 
possible to checkout any past point, but I try not to replace the manual 
here, just to give incentive to start having a look :-)

Moving the whole stuff to savannah is as simple as creating the repository 
(presumably through some online application form?) and simply pushing all 
interesting branches to it. That's it. A couple of simple commands. No 
complicated rsyncs, scps, cps or similar. No configuration files. No file 
permissions on the server. SIMPLE.

As soon as I have some more spare time, I'll clone the history of simulavr 
(without xx) and make a new branch inside the github repository so that 
everything is in place. 

How about that?

Have fun,

Onno