RE: How well does CVS handle other types of data?

[Paul Sander]

>--- Forwarded mail from address@hidden

>[ On Friday, July 13, 2001 at 14:57:41 (-0700), Paul Sander wrote: ]
>> Subject: RE: How well does CVS handle other types of data?
>>
>> How so?  By giving the hint as stated in the proposal, the user has
>> specified the desired result.  Whether the desired result is the same
>> as the correct result is another question, also answerable only by the
>> user.  But at least CVS didn't mangle his sources as it would today.

>Because you gain nothing and yet still leave ample room for error.  The
>user must still manually resolve the implied remaining conflict.  It
>might be nice to have some simple tools that can resolve these conflicts
>by taking one of the actions you specified, but if you leave that to a
>manual step afterwards then you can much more easily integrate that
>action into the necessary action of telling CVS that the conflict has
>been resolved.  I.e. your proposal still always requires at least two
>steps, even if the "hint" got things right.  If you instead just flag
>the conflict as I suggested then the user can do whatever is in fact
>correct for that file, and tell CVS that the conflict is now resolved,
>all in one step.

Yes, two steps are needed according to my proposal.  This includes bringing
all of the relevant versions into the working directory and performing the
merge.

According to your proposal, the user must take at least five steps
to produce the same result.

The gain here is convenience, and comfort in the knowledge that CVS won't
produce corrupt files by applying the wrong merge algorithm to non-ASCII
files.

>--- End of forwarded message from address@hidden