I've run into this strange issue where if I write a formula in org with laas-mode enabled, and then end a sentance after the math fragment the snippet is still activated.

For example in an org-mode before enter: hello \( some-math \).. The entire latex math fragment is wrapped in resulting in \\dot{( nm = -3 \)}.

I'm trying to bind a snippet to use in org-mode, while also using this package laas-mode.

(use-package aas
  :straight t)

(use-package laas
  :straight t
  :hook ((org-mode . laas-mode)
         (LaTeX-mode . laas-mode))
  :config
  (put 'aas-set-snippets 'lisp-indent-function 'defun)
  (aas-set-snippets 'laas-mode
    :cond #'laas-mathp
    "On"    "\\mathcal{O}(n)"
    "O1"    "\\mathcal{O}(1)"
    "Olog"  "\\mathcal{O}(\\log n)"
    "Olon"  "\\mathcal{O}(n \\log n)"
    "<<" nil)

  (aas-set-snippets 'org-mode
    "<<" (lambda! ()
           (yas-expand-snippet
            (concat ",#+BEGIN_SRC $1\n"
                    "$0\n"
                    ",#+END_SRC")))))

But nothing I do seems to work :?. I know laas-mode has it's own binding for << so I disabled it, but the org-mode snippet still checks laas-mathp despite me not specifiying it as a condition for aas-set-snippets.

@ymarco

Works like a charm. Thank you.

@ymarco

I personally already have a yasnippet with / as a key that expands to \frac{$1}{$2}$0. This is a parallel to that, requiring // instead of /<TAB> to expand the snippet.

Interesting... although I think // will be more intuitive than /<TAB> once you get used to it.

Anyway, I'd have no problem accepting this if you refactor laas-identify-adjacent-tex-object for us.

No problem. I'll look into this a little later if that's all right. Thnx.

@ymarco

But I'd like snippets (this specific snippet) to expand at the start of the line? Is there a reason laas--no-backslash-before-point? was put into aas-global-condition-hook instead of included in laas-mathp? I don't imagine it's good for laas to interfere with org-mode snippets defined outside of laas.

Before if you tried to do x5 -> x_5 and then x_51 -> x_{51} your point would be moved outside of the closing brace so if you wanted to add more to the body of the subscript you couldn't. Now you're kept in there and have to exit the closing brace manually.

@ymarco

Okay, I see the intention here. Truth be told I was just experimenting with laas and wondered why 2 sub digits is OK but more seem to require movement beforehand. Thinking on it this does make more sense... in so-far as I doubt I'll be subscripting into the 100s.

This can be closed. Thnx.

Actually if I apply the above patch the snippet doesn't expand at all at the start of the line, even if there's no longer any error triggered.

@ymarco

Yes, you're correct. Thank you.

I'm still getting an error with my binding at the start of the buffer though:

Debugger entered--Lisp error: (wrong-type-argument number-or-marker-p nil)
  laas--no-backslash-before-point\?()
  aas-expand-snippet-maybe("<<" (closure (t) nil (interactive) nil (yas-expand-snippet (concat "#+BEGIN_SRC $1\n" "$0\n" "#+END_SRC"))) nil)
  #f(compiled-function () #<bytecode 0x207588126261>)()
  aas-post-self-insert-hook()
  org-self-insert-command(1)
  funcall-interactively(org-self-insert-command 1)
  call-interactively(org-self-insert-command nil nil)
  command-execute(org-self-insert-command)

I think it's because (char-before) is nil in laas--no-backslash-before-point?.

--- /home/mohkale/.config/dotfiles/prog/editors/emacs/lisp/straight/repos/LaTeX-auto-activating-snippets/laas.el
+++ #<buffer laas.el>
@@ -384,7 +384,8 @@
 
 (defun laas--no-backslash-before-point? ()
   "Check that the char before `point' is not a backslash."
-  (/= (char-before) ?\\))
+  (when-let ((ch (char-before)))
+    (/= ch ?\\)))
 
 
 (apply #'aas-set-snippets 'laas-mode laas-basic-snippets)

I'm trying to bind a snippet to use in org-mode, while also using this package laas-mode.

(use-package aas
  :straight t)

(use-package laas
  :straight t
  :hook ((org-mode . laas-mode)
         (LaTeX-mode . laas-mode))
  :config
  (put 'aas-set-snippets 'lisp-indent-function 'defun)
  (aas-set-snippets 'laas-mode
    :cond #'laas-mathp
    "On"    "\\mathcal{O}(n)"
    "O1"    "\\mathcal{O}(1)"
    "Olog"  "\\mathcal{O}(\\log n)"
    "Olon"  "\\mathcal{O}(n \\log n)"
    "<<" nil)

  (aas-set-snippets 'org-mode
    "<<" (lambda! ()
           (yas-expand-snippet
            (concat ",#+BEGIN_SRC $1\n"
                    "$0\n"
                    ",#+END_SRC")))))

But nothing I do seems to work :?. I know laas-mode has it's own binding for << so I disabled it, but the org-mode snippet still checks laas-mathp despite me not specifiying it as a condition for aas-set-snippets.

Before if you tried to do x5 -> x_5 and then x_51 -> x_{51} your point would be moved outside of the closing brace so if you wanted to add more to the body of the subscript you couldn't. Now you're kept in there and have to exit the closing brace manually.

Closes #12

The README for this package suggests binding // to an empty fraction snippet but that won't work with the current implementation because it interferes with the / snippet. I'd suggest allowing the / snippet to support // as well. Thoughts?

This is really minor, but every function starts with "aas", except for "ass-activate-for-major-mode". Was this intentional?

ok, thank you for your quick reply; I try to get familiar with it

Hmmm. That's would be nice. I can't say I'll work on this any time soon but feel free to propose something.

Hi, very interesting snippets approach... is there some thinking about handling siunitx with the convenient approach as well? It would be great if the units get converted like m gets converted to \m and e.g. kg to \kg... but only if it is located in the second part of the SI-environment $\SI{10}{\m}... maybe this already possible!?

Thank you in advance!

Fab

Never mind, I think I misunderstood what you meant by fall back. I can put a conditional in the evil-tex-go-{back,forward}-section for org-mode if that's what you mean.

So the problem with that I think is that evil-tex-mode-map is a minor mode map which overrides the bindings in the org-mode major mode map. Minor mode maps appear to be "global" and the way to have different minor mode mappings per major-mode is to use minor-mode-overriding-map-alist.

Fixes #8.

We could have a check like:

(condition-case nil (evil-tex-an-env)
  (error (when (and (eq major-mode 'org-mode) (fboundp 'evil-org-an-object))
           (evil-org-an-object)))

This would avoid placing an explicit dependency on evil-org.