RFC new template keywords (wcmodified, wcmodifieddate) + sampleimpl.

Dominik Psenner dpsenner at gmail.com
Tue May 31 10:03:47 CDT 2011


>> My point is: Wouldn't releasing software built from modified working
>copies
>> change a support session into a nightmare as nobody is able to tell which
>> code's running for real?
>
>I agree completely, and I would NEVER release such a hack to any QA or
>PRODUCTION network. But then I can be quite the CM enforcer ;-)
>
>But I have found that a surprisingly large number of developers treat
>development network and testbeds as a different story, and as part of a
>drive to increase CM best practices, I'm trying to improve all
>identifiability, so "for any artefact on any system we can say where it
>came from".
>
>In corporate environments, it is sometimes amazing what some developers
>will try and get away with (time pressures, lack of clue, laziness, the
>list seems to go on and on). Or the other situation I've encountered is,
>developer's looking at one piece of code and running completely the
>wrong binary, because for example an scp / ftp failed and they weren't
>paying attention. Guess who tends to have to figure that out for them :-(
>
>I suppose I'm trying further improve already good local CM practices,
>and long the way, make the same improvements easier for others to
>implement.

Once more a diff between real and ideal is quite huge and I won't ever
understand CEOs forcing their subordinates into those practices. :-) At the
moment I can only think of better education as the only real solution. Going
further on with this discussion would lead to concrete deployment scenarios
that IMHO should not be part of a VCS. Certainly one can use a VCS to deploy
something, but that doesn't mean the VCS is aware of that.



More information about the Mercurial-devel mailing list