Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Oct 07 16:02
    DarwinAwardWinner opened #174
  • Jun 15 12:12
    DarwinAwardWinner opened #173
  • May 26 15:49
    DarwinAwardWinner closed #171
  • May 26 15:49

    DarwinAwardWinner on master

    Enable tests in emacs-snapshot … Eliminate explicit version chec… Fix recursive autoload issue and 3 more (compare)

  • May 20 15:35

    DarwinAwardWinner on bleeding-edge

    Fix a typo in a docstring (compare)

  • May 20 14:35
    DarwinAwardWinner closed #172
  • May 20 14:35
    DarwinAwardWinner commented #172
  • May 19 02:47
    wyuenho commented #172
  • May 18 15:15
    DarwinAwardWinner commented #172
  • May 18 13:06
    wyuenho commented #172
  • May 16 23:48
    DarwinAwardWinner commented #172
  • May 16 23:18
    DarwinAwardWinner commented #172
  • May 16 21:29
    wyuenho opened #172
  • Mar 11 18:38
    DarwinAwardWinner closed #165
  • Mar 11 18:38
    DarwinAwardWinner commented #165
  • Mar 11 18:36
    DarwinAwardWinner commented #165
  • Mar 11 13:42

    DarwinAwardWinner on bleeding-edge

    Switch to MELPA unstable for te… Improve code for detecting bug … (compare)

  • Mar 11 01:30
    DarwinAwardWinner commented #171
  • Mar 11 01:30

    DarwinAwardWinner on bleeding-edge

    Eliminate explicit version chec… Fix recursive autoload issue (compare)

  • Mar 11 01:15

    DarwinAwardWinner on bleeding-edge

    Fix recursive autoload issue (compare)

Jonas Bernoulli
@tarsius
here i am
Ryan C. Thompson
@DarwinAwardWinner
Hi!
Jonas Bernoulli
@tarsius
If I am not mistaken magit currently only has one command which requires that "" is returned for emtpy input. The second prompt of magit-branch-set-upstream: "Change upstream to branch (empty to unset)". So I am just gonna wrap that with (let ((ido-ubiquitous-next-override 'enable-old)) ...) and not use that anywhere else.
Ryan C. Thompson
@DarwinAwardWinner
Sounds fair enough.
Jonas Bernoulli
@tarsius
However I am currently having problems doing "ido-completion-as-intended" using completing-read-ido-ubiquitous. It always falls back to non-ido-completion for me.
Well, not always this works: (completing-read-ido-ubiquitous ": " '("a" "b" "c")). Let's try to narrow it down a bit.
Ryan C. Thompson
@DarwinAwardWinner
Hmm
Keep in mind that the magit override is still in there by default
you might have to fix that
Obviously I will fix the default override list before releasing.
Do M-x customize-group ido-ubiquitous and then delete any references to magit from the two override variables.
Jonas Bernoulli
@tarsius
Okay removing that helped (I was a bit overwhelmed by the custom interface, which is why this took me a while :-)
Ryan C. Thompson
@DarwinAwardWinner
Ok, give me a littlewhile to set up magit and dependencies in a clean emacs.
Jonas Bernoulli
@tarsius
Looks good now. magit-branch-set-upstream is the ultimate command for testing, because it needs both variants.
Okay.
Ryan C. Thompson
@DarwinAwardWinner
In the redesign, I’ve set things up so that the fallback mode is triggered by throwing a signal. Which means I can now implement throwing a reason for falling back along with the signal and then add a debug mode to echo it to aid with debugging.
Jonas Bernoulli
@tarsius
This sounds all very complicates :-)
Doesn't have to be a bad thing, but I don't really understand much.
Ryan C. Thompson
@DarwinAwardWinner
Most of the complexity is to make sure we never ever break completing-read because then emacs becomes unusable
Which I managed to do several times while testing upgrading ido-ubiquitous to this new version
Ditto for call-interactively and called-interactively-p, which also tend to break everything if you break them
Jonas Bernoulli
@tarsius
It's always fun having to kill Emacs from the outside :-)
I am just wondering if you couldn't save a lot of work if you just patched ido.el instead.
Ryan C. Thompson
@DarwinAwardWinner
Somehow I got Emacs into a state where even pressing C-g oiverflowed the stack
Oh wow, magit next is prettier
I should upgrade :)
Jonas Bernoulli
@tarsius
Yup
What theme are you using?
Ryan C. Thompson
@DarwinAwardWinner
It’s a default Emacs with just magit and ido-ubiquitous loaded
On OSX
White BG, black text
Jonas Bernoulli
@tarsius
In case you usually use zenburn or solarized-light, then you should try my forks of these themes. These changes will be merged eventually, but for now you have to get the magit-next branches from my fork. Solarized-dark should also look okay, but I haven't actually tried making it look good yet. The default themes both in the dark and light variants I have already themed too (that's what you are currently using), but that has fallen behind a bit by now.
Anyway how should we proceed? Looks like I will have to depend on ido-ubiquitious for ido support, not just ido-completing-read+. But I won't actually add that as a dependency of the package, to avoid forcing a dependency on users who don't actually want to use ido with magit. Instead I will show a warning when these packages are not installed and the user attempts to use ido completion.
Ryan C. Thompson
@DarwinAwardWinner
That sounds reasonable. I would maybe set up magit-ido-completing-read so it tries to require idi-ubiquitous, and if it fails, issue a warning and either use ido-completing-read or stanard emacs completion instead.
Jonas Bernoulli
@tarsius
Ryan C. Thompson
@DarwinAwardWinner
Yes, it looks reasonable. I’m in the process of testing it now
Ryan C. Thompson
@DarwinAwardWinner
I think my testing is being stymyied by stale elc files
Giveev me a few minutes to reset
Jonas Bernoulli
@tarsius
You might want to (setq load-prefer-newer t) and give my auto-compile package a try.
Ryan C. Thompson
@DarwinAwardWinner
Yeah, I’ll check those out. I’m testing in a clean Emacs process right now, though
Jonas Bernoulli
@tarsius
Well, not now :-)