profile
viewpoint
If you are wondering where the data of this site comes from, please visit https://api.github.com/users/oantolin/events. GitMemory does not store any data, but only uses NGINX to cache data for a period of time. The idea behind GitMemory is simply to give users a better reading experience.

minad/marginalia 231

:scroll: marginalia.el - Marginalia in the minibuffer

oantolin/embark 221

Emacs Mini-Buffer Actions Rooted in Keymaps

oantolin/orderless 190

Emacs completion style that matches multiple regexps in any order

oantolin/icomplete-vertical 79

Global Emacs minor mode to display icomplete candidates vertically

oantolin/emacs-config 28

My personal Emacs configuration

oantolin/live-completions 21

Live updating of the *Completions* buffer

oantolin/epithet 18

Emacs package to rename buffers with descriptive names

oantolin/async-completing-read 15

An async-capable completing read function

oantolin/math-delimiters 14

Insert math delimiters in TeX, LaTeX and Org buffers

oantolin/restricto 11

Restrict to current completion candidates

pull request commentoantolin/embark

Correction after embark--become-command got fixed!

Yup, works fine!

minad

comment created time in 15 hours

push eventoantolin/embark

Daniel Mendler

commit sha 252c7ef77a97162c04d219c0aabb44e526785de6

Correction after embark--become-command got fixed!

view details

Omar Antolín Camarena

commit sha fb2c85c67df7f86acdb04a5d3b286e2012b793b5

Merge pull request #327 from minad/fix Correction after embark--become-command got fixed!

view details

push time in 15 hours

PR merged oantolin/embark

Correction after embark--become-command got fixed!

There is no problem with let-binding/setq-local, the problem was the parenthesis mistake! goto fail!

+0 -5

0 comment

1 changed file

minad

pr closed time in 15 hours

push eventoantolin/embark

Omar Antolín Camarena

commit sha 14be7fafeea899a1374e113f2327b9e82fc0fff5

Fix typo in embark--become-command that kept command from running!

view details

push time in 16 hours

pull request commentoantolin/embark

Add embark--restart (Fix #117)

I deserve most of that, but will point out that the way I use embark-collect-completions, which is only by demand and usually after typing several characters, it is instant not at all sluggish.

I think most of all I like the peace of not having things popup or move around automatically. It's why I prefer VC to magit for simple operations, why I prefer embark-prefix-help-command to which-key, why I prefer the minimal indicator over the verbose or mixed indicators, and why I prefer completions to only appear when summoned.

minad

comment created time in 16 hours

issue commentoantolin/embark

Adding Embark to GNU ELPA

Omar, given Prot's analysis what do you think about sending the package to ELPA?

I guess it's bound to happen sooner or later, might as well get it over with.

oantolin

comment created time in 16 hours

pull request commentoantolin/embark

Vertico/Selectrum: Force candidate computation

I also think it's not necessary for normal interactive use, but I like that it's a small change, clearly more correct and simplifies automation.

minad

comment created time in 16 hours

push eventoantolin/embark

Daniel Mendler

commit sha 35cc82079a7b953bd1c67096ac669c19a7618baf

Vertico/Selectrum: Force candidate computation

view details

Omar Antolín Camarena

commit sha ba55927bcbeb8ee0b894ad8f0c7dcc453cc978ff

Merge pull request #325 from minad/vertico-force Vertico/Selectrum: Force candidate computation

view details

push time in 16 hours

PR merged oantolin/embark

Vertico/Selectrum: Force candidate computation

As mentioned on reddit. Do you want to have this? (In my opinion it is not strictly necessary to add this. One can always use run-at-time 0).

+15 -0

0 comment

1 changed file

minad

pr closed time in 16 hours

pull request commentoantolin/embark

Add embark--restart (Fix #117)

Thanks for insisting on this, and I'm sorry my reluctance made it necessary to insist.

minad

comment created time in 16 hours

push eventoantolin/embark

Daniel Mendler

commit sha 57c97b8e001926954bfa8437affee565655d875d

Add embark--restart (Fix #117)

view details

Daniel Mendler

commit sha 36a6a9ae2d21245662a1b5150e696b35e0b98106

Fix executing multiple actions in sequence

view details

Omar Antolín Camarena

commit sha 8fd93959447cd8bc816d639d2a09e5842321ddc4

Merge pull request #326 from minad/restart Add embark--restart (Fix #117)

view details

push time in 16 hours

PR merged oantolin/embark

Add embark--restart (Fix #117)

Now that we have the flexible embark-post-action-hooks, this addition is cheap. Furthermore we can reuse the code of embark-become.

+39 -11

5 comments

1 changed file

minad

pr closed time in 16 hours

pull request commentoantolin/embark

Add embark--restart (Fix #117)

It is a pragmatic solution, which is actually satisfactory for me.

It's good enough for me too, which is why I already said we should do this. 😉

If not restoring the selected candidate is good enough for you, it's good enough for me too and we can merge now.

minad

comment created time in 17 hours

pull request commentoantolin/embark

Add embark--restart (Fix #117)

This looks good as it is, but we still need to worry about restoring the selected candidate for Vertico, Selectrum and Icomplete, right? Or do you think it's fine to not restore the selected candidate?

minad

comment created time in 17 hours

pull request commentoantolin/embark

Add embark--restart (Fix #117)

Fine, I guess this is a net win for users, so let's do it.

As you know, I'm a little reluctant to add here a workaround for this limitation of some completion tables that are too static and some completion UIs. For example, Vertico does things right! And Vertico works just fine with the built-in buffer and file completion tables that are appropriately dynamic.

This thing is not Embark's fault and Embark can't actually fix the problem! This PR only deals with the case when the problem is triggered by using Embark actions. The issue really should be fixed at the level of completion tables and completion UIs, so that the fix doesn't just apply to Embark actions, but also to recursive minibuffers, recursive editing, and just plain external changes like removing files from an external file manager.

But I think there is little or no will to correctly address the problem, and certainly having this here will help solve the part of the problem that Embark exposes. And there are many Selectum users of Embark, who would greatly benefit from this. So I guess we should do it. I do however wish people would just switch to Vertico and that we'd fix the completion tables in Consult, instead.

minad

comment created time in 17 hours

PullRequestReviewEvent

Pull request review commentoantolin/embark

Add embark--restart (Fix #117)

 respects symbol boundaries."  ;;; Setup and pre-action hooks +(defun embark--restart (&rest _)+  "Restart current command with current input.+Use this to refresh the list of candidates for commands that do+not handle that themselves."+  (when (minibufferp)+    ;; XXX: This setter is needed in order to allow restarting+    ;; again after the restart. I don't understand how+    ;; embark--command is treated in Embark, sometimes it is+    ;; let-bound, sometimes setq'ed.

Sometimes using let and sometimes setq is not the result of careful thought and is not on purpose. It's just me being sloppy. Probably they can all be let.

minad

comment created time in 17 hours

issue commentoantolin/embark

Correct or remove out of date information from the wiki

Yes! Great idea, thank you! Let's just filter out the genteel actions from all the other keymaps.

oantolin

comment created time in a day

issue commentminad/consult

Feature like `isearch-yank-word-or-char`

depending on the completion UI

It was nice of you to throw that in their, but I don't mind it when people talk as if default completion didn't exist. 😛

Lenbok

comment created time in a day

issue commentminad/consult

Feature like `isearch-yank-word-or-char`

Me too! Even before consult-line, back when I used Ivy, I tried using swiper instead of isearch and didn't get used to it. I think isearch is best for getting the point to some location, while swiper and consult-line are better at browsing and getting an overview.

Lenbok

comment created time in a day

issue commentminad/consult

Feature like `isearch-yank-word-or-char`

To be fair, none of those techniques substitutes for isearch's C-w in full generality. You can press C-w more than once, and you can even move to a different match between C-w's.

For example, if point is right before foo bar, a little later in the buffer there is a foo baz and you want to search for foo baz you can go C-s C-w C-s C-w. If in the same example you want to search for foo bar (i.e, C-s C-w C-w), that you can also do with embark-act: just mark foo bar and then use the C l action.

Lenbok

comment created time in a day

push eventoantolin/embark

Omar Antolín

commit sha 61a301dfdad74adfff2bd9b20764c944f9880250

Typo in README

view details

push time in 2 days

push eventoantolin/embark

Omar Antolín

commit sha 67d50b9681690f194a9cf3ebf0f72f0707b6bb2b

Mention change to hooks in README

view details

push time in 2 days

push eventoantolin/embark

Omar Antolín

commit sha bb15b8632517de64e2947c2c33ecb5ac60282868

Add some obsolescence warnings for old hook variables

view details

push time in 2 days

issue commentoantolin/embark

Can I include a command in a keymap, that does not pass an argument?

On versions before this commit, this will just effectively be ignored?

Ah, no. You'll get an error message that says:

Symbol’s value as variable is void: embark-pre-action-hooks

Because the variable used to not have an s at the end. You can wrap the add-to-list in (when (boundp 'embark-pre-action-hooks) ...) if you want it to be ignore when the variable does not exist.

BTW, what's your strategy on versioning?

I don't really have one.

bdarcus

comment created time in 2 days

pull request commentoantolin/embark

Use alists for the hooks, pass target and bounds to the hooks

I think I'm going to remove the special casing of the minibuffer. I do sometimes want other actions in there, mostly when using eval-expression, but not only then.

minad

comment created time in 2 days

issue commentoantolin/embark

Can I include a command in a keymap, that does not pass an argument?

OK, fair enough. I just meant you could always have written a pre-action hook to ignore a target, but I agree there is a qualitative difference having to write a hook yourself and just setting up an existing function to be used as a hook.

bdarcus

comment created time in 2 days

issue commentoantolin/embark

Can I include a command in a keymap, that does not pass an argument?

You can do something like:

(with-eval-after-load 'embark
  (add-to-list 'embark-pre-action-hooks '(org-cite-insert embark--ignore-target)))
bdarcus

comment created time in 2 days

issue commentoantolin/embark

Can I include a command in a keymap, that does not pass an argument?

We are currently adding support for this. See embark--ignore-target in #323.

@minad, the support was already there, you mean you changed the interface. 😛

bdarcus

comment created time in 2 days

issue commentoantolin/embark

Can I include a command in a keymap, that does not pass an argument?

Does the function prompt you for anything at the minibuffer? If so, when used as an action, Embark inserts the target at the first minibuffer prompt and simulates pressing RET afterwards. It doesn't matter if the thing prompted for is not an argument to the function: if the command opens a minibuffer at least once, the first time it does so, the target gets injected. If you want to avoid this injection you need to ignore the target somehow: either with a wrapper or with a pre-action hook (which already existed before this change @minad mentioned, but got easier to use now).

bdarcus

comment created time in 2 days