|
Boost-Build : |
From: Roland Schwarz (roland.schwarz_at_[hidden])
Date: 2007-08-19 07:44:44
Rene Rivera wrote:
> I object... I don't want to loose the trunk changes ;-)
The issue as I see it, is not to loose changes from trunk, but from RC_1_34.
Triggered by Renes objection I tried to find the correct base point for
three way compare, since RC_1_34_0_freeze indeed is not.
Thomas Witt announced existence of RC_1_34_0 branch on
2006-03-21, which funnily is the date of begin of spring. ( Maybe
because trees tend to get new branches? ;-) )
The last commit mail before is on 2006-03-20, and the first after is on
2006-03-22. The first commit to RC_1_34_0 branch I can see also is on
2006-03-22.
So I think a safe base for three-way compare would be to use CVS
checkout by date: 2006-03-21.
I did this, and unfortunately things show up to be much more complicated.
So I want to raise the question why it wouldn't be a better base to use
merged_to_RC_1_34_0 instead. At least as I understand this was the
reason why this tag was created at all, not?
If I do this, 3-way merge reduces to 7 manual merges necessary, which
look easy to resolve.
Opinions?
Roland aka speedsnail
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk