[p4] $Log$ et al
jbowles at best.com
Thu Oct 21 13:40:55 PDT 1999
I don't know if it'll post from this account, but I'll try.
Mike and Marc bring up good points w.r.t. submit/sync triggers,
but I have to dump my $0.02 into the mix.
I agree with Mike that there's more powerful ways to make a really
handy "massage the content before I give it to the server or before
it gives it to me" mechanism. Don't mislead yourself into thinking
you're talking about sync/submit, however: every bit of [file]
content would need this mechanism, which means that "p4 print" and
"p4 diff" would need to use this mechanism also. Otherwise, Marc's
"reformat for consistent tab usage" (that works when you sync a
file) would cause 'p4 diff' to look screwy.
The one thing I haven't seen addressed (I'd stopped receiving the
list when I went on vacation) is the issue of portability and the
other one that tags along for this discussion, security. In a
homogenous development environment in which you can guarantee that
each workstation has access to the same tools, these mechanisms
can work and it'd be hard for a devious developer to get around
the triggers/mechanisms we'd *all* put into place to prevent them
from checking in lousy code.
But in a heterogenous environment, which is one of Perforce's
strengths, I can't see how to guarantee that the SAME script with
the SAME semantics and the SAME security is run on a Macintosh as
a VMS machine or an OS/360 client.
Honestly, I'm not sure I can about the security parts. We could
just say "we're talking about a friendly developer." (Paul Tsongas
got a group of reporters to laugh when he said something similar.
Bear with me.)
But getting consistent behavior on each client platform is one of
the great things about Perforce. This proposed feature would be a
very useful one if we could propose how to crack *that* nut, also.
More information about the perforce-user