Re: import inconsistency

[Paul Edwards]

"Pierre Asselin" <pa@invalid.invalid> wrote in message 
k99bcb.sda.ln@brick.verano.sba.ca.us">news:k99bcb.sda.ln@brick.verano.sba.ca.us...
> >> I have no idea which inconsistency you're referring to.
>
> > A tinpot branch who gets in first, causes the file to stay in the
> > Attic forever, which means exception coding forever.
>
> Maybe the solution is for import to change
>
>     1.1 (dead, empty)
> 1.1.2.1 (live)
>
> into
>
>     1.1 (dead, empty)
> 1.1.2.1 (live)
>     1.2 (live)
> 1.2.1.1 (live, no delta from 1.2)
>
> with the RCS -b branch pointing to 1.2.1 instead of the usual 1.1.1 .
>
> Is this feasible?  Or is 1.1.1 hard-coded all over the place?
>
> Paul, the imported file would now be live on the trunk (Attic or no Attic)
> instead of dead on the trunk.  Would that solve your problem, the one
> that forces you in to "exception coding" ?

I wouldn't use the words "forces me into exception coding" to
describe my feelings.  CVS is going through logic paths that
are specific to that file.  You can see reports that CVS even
has problems with a trailing "/" in $CVSROOT, so I don't
like to know that particular files in my repository are going
through unusual code.  I wouldn't mind so much if the files
in the Attic were dead and unimportant, but actually they are
every bit as live as the other ones.

I know that CVS works on normal files, because if there was
any problem with normal files, everyone else would have
complained long before me and the problem would be fixed
already.

The problem is when I am doing things that other people are
not doing.  E.g. the poor bugger who was using a case-insensitive
file system and had a file in the Attic.  He was asking for trouble,
and he got it.  I am asking for trouble with good, live files in the
Attic.  I don't want to get it.

On the other hand, if EVERYONE's files went into the Attic on
import, then we'd be in business.  Although the fact that everyone
else except me works on the head is a real bummer.  Most of the
problems I have encountered (e.g. rdiff -s being useless) are
because most people are not actually using the branch feature,
except for a single, simple, vendor branch, and even that has
been implemented in such a way that you don't even need to do
a subsequent cvs update.

It is not pleasant doing things differently from other people, but
unfortunately that is the best way to work from a productivity
point of view.  It just exposes me to integrity problems with
CVS as a side-effect.  It's not the end of the world though, assuming
that conflict on added files gets committed, there will be no
KNOWN integrity problems involving source code lost, which is
a damn good start!!!  Oh, and the fact that there are now separate
release and development branches also makes a hell of a difference
too.  At the end of the day, it's pretty good.  :-)

BFN.  Paul.