[Rcml-group1] Re: hi david..

[John B. <jx...@ga...>]

> 
> may be what i was saying is not relevent to the
> getrevision..what i said that..
> if u are using getrevision in checkout then i found
> that as we can give a revision number higher than
> head id like ..if the headid=1.1 and the revision to
> be
> checked out is 1.3 then it checkout 1.1..
> but if the revision to be checked out is specified as
> 2.3 or something like that..it says that the revision
> is absent..

  Oh!  I understand now.  I'm slow, sometimes. :)

  Yes, we want it to work just like RCS.  To RCS, if you checkout
something on a non-existent branch, it isn't there.  The branch
numbers are any odd-count number, such as 1, 2, or 1.3.1 .  So, the
increment 3 system should refuse to perform a checkout for anything
with a branch other than 2.

  You're right; this should be added into getRevision().  I believe it
should have a fatal error (error message and exit) if the user does
this.

> anyway i guess the above thing should be included in
> the pseudocode of checkout..
> i have an exam tommorrow..that's why i can't do much
> today..i'll be coding..the appdiff..but i guess 
> i didn't have the total idea about the deque and 
> pop and push things..so please gothrough that code
> after cvs updating it..
> i got quite hectic schedule ..i got to submit my db
> project on next monday..n/w's project due this
> wednesday..it's 7:45 p.m..i'll be here upto 9 p.m

  I understand.  I've been there. :)  Just do your best, and that will
be all right.  Check the docs for the STL at
http://www.sgi.com/Technology/STL/ for information on the methods
available on a deque; it can be accessed just like an array, too.

  Please do your best to make sure that everything at least compiles
before doing a cvs commit.  (I haven't checked yet to see if you did
one yet, but discovered that problem Saturday; commits can break what
other people are working on at times.)

jdb