[darcs-users] Re: Frustrations diffing against the last change to a file

Michael G Schwern schwern at pobox.com
Thu Mar 31 21:40:27 UTC 2005

On Thu, Mar 31, 2005 at 05:00:17PM +0000, Mark Stosberg wrote:
> On 2005-03-31, Michael G Schwern <schwern at pobox.com> wrote:
> >
> > I mention it because, well... its the truth.  Its a lot easier to remember 
> > and type in "3214" than "a TODO test which confirms that RT#266 is a 
> > regression" or "Sat Mar 12 14:08:46 PST 2005".  
> If you already have to remember that you working ticket #266, why do you
> want to remember yet another number?
> My sense is that it's difficult to "just know" 3214, in either case it's
> probably a two step process: 
>     1. Look up the revision
>     2. Use it.

Yes, I outlined that two step process.  This is what I do with conventional
version control and this is what I suspect I will be doing a lot with darcs
as well (see below). 

You're assumption is that you can get away with 
"darcs diff --from-match='...some text...'" directly without having to grep 
the change log for a sepecific change and get a unique identifier (currently 
its date).  And when there's enough unambiguous information to do that its

> I'd much prefer to try to remember an existing ticket number, which I
> would use several times over, or a some human text. 

Yes, sometimes you can just use a number or other convient unique identifier
in the change description and simply search for that.  Its pretty neat that
darcs lets you do this in one step.  But there are many times you cannot. 
Such as...

What if the identifier is not unique?  For example, what if there were
several patches to resolve RT#266?

What if there is no convenient identifier?  Skimming through the darcs
change log most of the changes do not have an RT ticket number or anything
else to identify them.

In those cases you're back to the two step process.  Grep the change log
for your change, enter the identifier into diff (or whatever).  In these
cases its nice to have a simple identifier in addition to darcs' patch
searching capabilities.

More information about the darcs-users mailing list