On Wed, 2004-12-29 at 22:01 -0500, Mark Gilbert wrote:
> On Wed, 2004-12-29 at 20:46 -0500, Hubert Figuiere wrote:
> > Relying on the date for a branch is not the right way. How do you know
> > the date of branching ?
>
> Because I branched it? November 22nd 22:00 CET. That's the date you
> use to get HEAD as of the branchpoint.
So it is not obvious.
> > cvs co -rbranchpoint my_module is so simple to recover the tree "at the
> > branchpoint"
>
> cvs co -D my_date my_module, to exclude for simplicity the other ways to
> "recover the tree". Precisely the command I'd use before tagging the
> bloody branchpoint if it'd make you happy.
What if you forgot that date? What if someone else wants to do it ?
Apparently you forget that AbiWord coding is not MG's but "a group of
developper".
>
> > and a "cvs rtag -b -rbranchpoint BRANCH my_module" is more complicated
> > to tag ?
>
> I didn't say that, only that it's unnecessary.
It is not. See above.
>
> > You still haven't offered ANY serious alternative to what I'm offering
>
> Was I joking?
See above.
> > You are telling me, if I read correctly in beetween lines, that there is
> > some cron script doing that ? So are are telling me that we might shoot
> > ourself in the foot ? BTW, I asked clearly THAT question, never got a
> > reply. Where is that script run from ?
>
> Yes, because developers have better things to do than repetitive
> administrative toil, like code. We have machines to do that sort of
> work for us now, and error handling to make sure that they don't get out
> of line. Seeing as multiple conflicts have arisen and it's worked
> perfectly as described, I wouldn't be so quick to smash the stocking
> frame.
So obviously you have NO experience in that kind of task. Obviously not.
> And no, I am not telling you that "we might shoot ourself in the foot,"
> you said that. IF, you have a rather more technical explanation of the
> foot-shooting bonanza, I'd be glad to hear it.
Yep. Control what you do, and at least you'll save from some mistake.
Excuse me, but letting a program messup with my code this way just
scares me, even more because it does it silently. Why do you think
there is not merge and commit command in CVS ? Not to make it bad. Just
because a merge should be reviewed ALWAYS.
Since we don't have non-regression and unit testing, there are at least
3 things: build, test and diff review.
> > 2/ you had posted a mail to the mailing list explaining that.
>
> It's not my branch. Uwog emailed the list about it, said to blame me
> for the name of it which is true enough, and I thought nothing more of
> it, because I had no idea whatsoever that my choice of priorsync would
> be so biblically controversial. Looks like I won't make that mistake in
> the future.
No. It is the lack of explanation about what has been done, and the
fscking arrogance you take it Mr Genius.
> > No. By arbitrary I mean "undefined". only a symbolic tag or a date can
>
> In the prior section of that manual you love to shake you'll find that
> HEAD is indeed explicitly described as a symbolic tag.
No. It is STICKY symbolic tag. It designate a branch, not a tree state.
> Now, all that said, I wish to point to nothing more than that it works.
> It worked, it's working, and it'll work some more. Me mum always taught
> me that in the open source world, work's what matters.
Apparently she taught you well how to piss off people seriously. You
just won.
And if you think I'm becoming offensive, yes, I am. The other day I did
that mail, sent it to anyone, in the hope that either someone who tell
'go ahead' or 'I have my idea, let's talk about it to see what is the
best'. And the other you seems to act like a dictator. Remember dude.
You are no-ones boss here.
Hub
-- Crazy French - http://www.figuiere.net/hub/Received on Thu Dec 30 04:18:48 2004
This archive was generated by hypermail 2.1.8 : Thu Dec 30 2004 - 04:18:48 CET