[darcs-users] darcs patch: some cabal fixes for 2.2

[Trent W. Buck]

"Judah Jacobson" <judah.jacobson at gmail.com> writes:

> On Wed, Dec 17, 2008 at 6:04 AM, Salvatore Insalaco <kirby81 at gmail.com> wrote:
>> I had to disable haskeline on windows for two reasons:
>
> First, thanks for the haskeline bug reports!  I should mention that
> you can log any additional issues at http://trac.haskell.org/haskeline
> .
>
>> 1) On windows, even without haskeline, the "usual" prompt keys (home, end) work.
>
> That issue is fixed in haskeline-0.5 (released yesterday :-).
> Incidentally, that version also implements custom key bindings (for
> example, to make ctrl-left/right skip words).

Can it read my ~/.inputrc yet? ;-)
I'm awfully fond of

    $if mode=emacs
      # poor-man's insert-parentheses
      (: "\C-v(\C-v)\C-b"
      [: "\C-v[\C-v]\C-b"
      {: "\C-v{\C-v}\C-b"
      }: forward-char
      ]: forward-char
      ): forward-char
    $endif

> It might be worth bumping up the dependency to haskeline-0.5 for
> darcs-2.2; it could use a little more testing, but no code changes
> would be necessary since the part of the API that darcs uses has
> stayed the same.  I'll send a patch/proposal once the testing/bugfixes
> are completed.

Is there any reason it won't work with 0.4?  Why not just increase the
ceiling (>= 0.4 && <= 0.5), rather than the floor as well (== 0.5)?
That's what I get told to do for Debian package dependencies, because it
makes it easier to "backport" (i.e. cherry pick) new upstream of a
critical package into an otherwise stable OS.

(Easier because the backporter doesn't have to override the dependency,
nor backport extra library packages.)