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

[address@hidden]

Michael Hennebry wrote:
> On Sat Sep 26 3:11 , ThomasK address@hidden> sent:
>
>> as Michael Hennebry asked too in his posting a few days before: how to
>> proceed with all the changes, which are made now on Onno's GIT repo at
>> github.com?
>
> I think the thing to do is to go from CVS to git
> and make Onno's git repo the official repository.

I really should have been more clear.
I should have written "and make a copy of Onno's git repo the official 
repository.".

On Sun Sep 27 13:32 , Joerg Wunsch  sent:

>As Weddington, Eric wrote:
>
>> I don't understand the reasonsing for this. The purpose of Savannah,
>> is to provide central project management, including a revision
>> control system. I have no overwhelming objection to using git, but I
>> would like to understand the reasoning why the current CVS
>> repository is unacceptable, or why SVN is unacceptable that it
>> requires that we use git.

Changing from CVS certainly is not required.
My suggestion was based on the following premises:
There has been a lot of work done that for some reason
did not find its way into the Savannah repository.
Otto's git repository is the Savannah repository + the additional work.
git and subversion are both better VCSs than CVS.
For those who are good at them, git is better than CVS and subversion.
For whatever reason, there has been discussion of changing the VCS.

Were it just me, subversion might be best.
CVS has never brought me joy, git is still something I've just read about,
and subversion is something I use at work.

>git is available on savannah as well.  I don't really know why it's
>really that much better though, but I wouldn't object if that is the
>general consensus among the active developers.  It's those people who
>have to work with it, so it's their decision.
>
>But like Eric, I'd like to point out that any active developer should
>simply be a registered member within the savannah group, so he gets
>VCS write access, regardless of which VCS is chosen.

Clicking on "request for inclusion" brought up an "unknown signer" message.
Firefox did not specify the signer.

Even with commit privileges, I probably wouldn't use them.
I find it way too easy to make a mess.

>> Personally, I do have a problem making the "official" repository
>> offsite away from Savannah.
>
>Me too.  Anything else would be yet another code fork.

My mistake.  I'd intended to specify a copy of Onno's repository.

>Also, please don't forget that the historic simulavr code base should
>not be just dropped under the carpet when migrating.  It is part of
>the project, there are at least still some users of it (among others,
>avr-libc's test suite has not been migrated to simulavrxx), and the
>minimum requirement I'd like to see for simulavrxx before even
>considering to completely deprecate the old code base would be that
>simulavrxx starts making official releases of some kind, within a sort
>of regular interval.  Amonng other things, I'd need such releases to
>make the code go into the FreeBSD ports tree.  Just checking out a VCS
>snapshot doesn't give me a warm feeling about the status of the
>project.  (I don't care though whether these releases are named
>"simulavr" or "simulavrxx"; if the former, then please start with
>version 1.0, to make it well distinguishable from the old 0.1.x
>versions.)

--
Michael Hennebry
address@hidden
"War is only a hobby."


---- Msg sent via CableONE.net MyMail - http://www.cableone.net