Eric Kow wrote:
...
hmm, stuff to think about.
I went and played around with it a little bit in the beta.
Not enough to get used to the whole workflow, but enough to suggest that
my intuition is not entirely unreasonable:
It could start out by looking basically like:
=== REFERENCE OLD STATE, do not edit --v ========
some old lines
some old lines
=== PATCH here between this state --^ and that state --v ==
(something... see my paragraph below)
=== PATCH here between this state --^ and that state --v ==
some new lines
some new lines
=== (REFERENCE NEW STATE ABOVE, do not edit --^
and you edit the stuff in the middle. As per Florent who usually only
edits the second set of lines in the current darcs-beta, it should
probably start out being equal to the "new" lines, for convenience.
This is the operationally the same as current-beta in which you can't
edit the first set of lines. The interface as above makes sense to me
in its idea that you naturally split *one* patch into *two* patches (of
course!), and it's quite obvious where those two patches are... hopefully...
(incidentally, with syntax like this we could, if we wanted, allow the
user to insert more PATCH lines and it would work naturally -- not sure
if that workflow would be any faster/slower/easier-to-see-what-I'm-doing
than repeatedly splitting into two pieces)
if we like this, we'll still want to work on wording.
(er, since we include lines from files literally, what happens when they
happen to contain lines like "=== PATCH here ..."? maybe some trick like
we outnumber their equals-signs or something?)
-Isaac
_______________________________________________
darcs-users mailing list
[email protected]
http://lists.osuosl.org/mailman/listinfo/darcs-users
