Re: [Gnu-arch-users] Re: Linus

[Tom Lord]

    > From: Robert Collins <address@hidden>

    > On Mon, 2003-10-13 at 07:47, Tom Lord wrote:

    > > I have some questions: do CVS server-side hooks get to vote on whether
    > > or not a transaction completes?  =20

    > Yes.

    > > Do they get to postpone a second
    > > transaction until after the hooks from a first transaction completes?

    > They could, [...]

The answer is "yes" in the sense that I meant the question.
Schematically, I asked "Is it A;  is it B"  where A implies B but B
does not imply A.   No matter....

Yes, to do a server-side prevent of a transaction would need more
hacking than I think counts as "write a hook".  So, yes, arch doesn't
have that CVS functionality -- I estimate about 4K LOC to provide it
in a good way (fewer LOC to provide it in a mediocre way).  The next
question is (assuming your information is accurate): does anybody
actually use that property of hooks?   

The only examples of CVS hooks I've seen would be very broken if they
prevented transactions from completing.


    > > If the answers are both `no', then arch has server-side hooks in the
    > > CVS sense too.   If either or both answers are `yes', then, yes, it's
    > > reasonable to say there's a functional difference here.

    > Any word on whether you plan to apply Eric's patch, the one that gives
    > hook a pre-commit vote ?

I long ago "approved" the idea.  I don't see Eric's patch in the bug
tracker....

-t