Re: import inconsistency

[Derek Robert Price]

Paul Edwards wrote:

> First of all, the example test case I gave was just the minimum
> required to invoke the problem, the real situation is when working
> on a branch.
>
> On a branch I add a file.
>
> On an import the file also gets added.
>
> Now I attempt to join all imports onto the branch.  It says "conflict -
> has been added" and doesn't know what to do.  I have two versions
> of the same file, and it doesn't know which one is correct.  But it
> doesn't actually mark it as a conflict, so if I do a "cvs commit", it
> doesn't produce an error.  Now it does.
>  

Okay, that makes more sense, now.  I'll test out your patch and see what 
I think.

You are right that a 3-way diff against /dev/null is probably the full 
solution for the merge of two added files (on update w/o -j as well). 
Your solution allows something which simply touches the file to cause 
CVS to lose track of the conflict.  A three-way diff would leave merge 
markers for CVS to check for as well.

The test case you submitted leaves a little to be desired since if the 
three-way diff is every implemented, CVS will hopefully not flag a 
conflict since the content of both files was the same.

Derek

--
               *8^)

Email: derek@ximbiot.com

Get CVS support at <http://ximbiot.com>!
--
In my day, we didn't have virtual reality. If a one-eyed razorback
barbarian warrior was chasing you with an ax, you just had to hope
you could outrun him.

        - Sarah M. Wolford, Hanover in a Washington Post contest