I've sent Nicolas a copy of the user map file.

Only difference is I changed Charlie's email address to
[email protected] since he is concerned about spam bots.  I've
updated your Google docs file as well.

I am fairly certain all the remaining unmapped handles are former SSL
employees (summer hires, etc.) and would all be defunct email addresses
anyways.

----- Rom

-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Rom Walton
Sent: Friday, July 27, 2012 11:49 AM
To: Oliver Bock
Cc: [email protected]
Subject: Re: [boinc_dev] Git Migration

By their very nature, the workspaces concept is useless in a DVCS
environment.

I could see the various workspaces be considered branches off of master
and then deleted for the final product.  Although, to be honest, I'm not
sure it really matters. The only branches that are important at this
stage of the game are master, 6.8, and 7.0a. So many of the earlier
branches and tags wouldn't build properly because of the missing
binaries or CVS corruption. I don't think we should stress too much
about orphaned nodes.

I guess I would view the acid test as to whether 'blame' works on 6.8,
7.0a, and trunk.  If so the tree is in good enough shape.  I'll check
out the boinctaskbar.cpp file which we brought in a change from DavidB's
workspace and see if it causes problems in boinccleaned.git.

> Did you test/compare both approaches?

>From what I could tell the SVN meta data in the comments section were
different.  You guys are the experts. :)

I want to get this wrapped up.

----- Rom


-----Original Message-----
From: Oliver Bock [mailto:[email protected]]
Sent: Friday, July 27, 2012 5:19 AM
To: Rom Walton
Cc: [email protected]; David Anderson (BOINC)
Subject: Re: Git Migration

On 7/26/12 18:59 , Rom Walton wrote:
> Which of the two baselines should we be considering our Git central 
> repository on?

Not easy to tell since I haven't had the time to have closer look at
Nicholas' approach and results.

I used the proven standard approach while Nicolas used a fairly new
tool. That doesn't necessarily mean anything about the result of
course...

I think my approach covers all requirements and I responded to all
comments made so far. The only potentially missing thing is the
integration of the personal workspaces Nicolas asked for. I asked you
and David for your opinion but haven't got any response so far. If that
is vital to the migration (and can't be done in separate repos like
boinc-alpha) I think we have to use Nicolas' conversion, if not I think
mine is probably safer.

> Basically both baselines have the basics, Nicolas went a little too 
> extreme on removing precompiled binaries and his current demo tree 
> broke the -test_app functionality for make_project.  It also appears 
> it didn't take Nicolas very long to migrate it all so stuff like that
can be fixed.

How about user mapping? Has this been completed by now?

> which tool did a better job?

Did you test/compare both approaches? Did anyone apart from Nicolas and
me (probably Charlie) test anything? The ML is almost completely silent
about this.


Best,
Oliver

_______________________________________________
boinc_dev mailing list
[email protected]
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.
_______________________________________________
boinc_dev mailing list
[email protected]
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.

Reply via email to