[darcs-users] How does --ask-deps work?

David Roundy droundy at abridgegame.org
Fri Aug 6 10:27:31 UTC 2004


On Wed, Aug 04, 2004 at 01:43:56AM +0200, Tommy Pettersson wrote:
> I seem to not understand how the --ask-deps works.  In a
> source file I added a pair of function and macro definitions
> (like assertions) for bug detection, and also many one-liner
> invocations of these functions and macros at strategic places.
> 
> I recorded the definitions as a single hunk.  Then I wanted to
> record (with darcs rec --ask) the uses of these in a couple
> of separate patches that all, one by one, depended on the
> definition patch (D).
> 
> The first patch (U1) worked.  But I could not make the second
> patch (U2) depend on just D.  When I answered No to depend
> on U1, darcs skipped the question for D.  And when I backed
> up and still said Yes to D, it automatically also filled in
> Yes for U1.  Same thing for U3: I could chose to depend on
> [U2], [U2,U1], [U2,U1,D], but not on just D.  This dependency
> hierarchy seems upside-down to me.
> 
> Either some test condition for --ask-deps has a bug, my
> intuition is from another planet, or the --aks-deps option
> is very clever and there is a section in the docs about it
> not yet written.

This is definitely a bug.  I think I may have fixed it, but haven't yet
tested my fix.  I definitely found where the bug came in (which wasn't too
hard), but whether or not I fixed it properly I haven't figured out.  Can
you check to see if it's fixed in the latest darcs (which should have the
fix in 15 minutes or so, assuming the test suite passes)?

Thanks for reporting this, by the way! I haven't used --ask-deps for a long
time, which is part of why this bug has been sitting in the repo so long.
-- 
David Roundy
http://www.abridgegame.org




More information about the darcs-users mailing list