Org: Improve inline src block fontification

This fixes the significant issue where a unpaired bracket would cause
the rest of the buffer to be parsed, ignoring `limit'. Not nice.

I also now use inline src blocks a lot more in my config.

config.org

@@ -5,9 +5,10 @@
 #+macro: timezone (eval (substring (shell-command-to-string "date +%Z") 0 -1))
 #+macro: git-rev (eval (format "@@html:<a href=\"https://github.com/tecosaur/emacs-config/commit/%1$s\" style=\"text-decoration: none\"><code style=\"padding: 0; color: var(--text-light); font-size: inherit; opacity: 0.7\">%1$s</code></a>@@@@latex:\\href{https://github.com/tecosaur/emacs-config/commit/%1$s}{\\normalsize\\texttt{%1$s}}@@" (substring (shell-command-to-string "git rev-parse --short HEAD") 0 -1)))
 #+html_head: <link rel='shortcut icon' type='image/png' href='https://www.gnu.org/software/emacs/favicon.png'>
-#+property: header-args:emacs-lisp :tangle yes :cache yes :results silent :comments link
+#+property: header-args:emacs-lisp :tangle yes :comments link
+#+property: header-args:elisp :exports code
 #+property: header-args:shell :tangle "setup.sh"
-#+property: header-args :tangle no :results silent
+#+property: header-args :tangle no :results silent :eval no-export
 #+options: coverpage:yes
 #+startup: fold
 
@@ -136,7 +137,7 @@ digraph {
 
 #+caption: Some sample workflow integrations that can be used within Emacs
 #+attr_html: :class invertible :alt Graph of possible Emacs task integrations :style max-width:min(24em,100%)
-#+attr_latex: :width 0.5\linewidth
+#+attr_latex: :width 0.55\linewidth
 [[file:misc/emacs-platform.svg]]
 
 *** Some notably unique features
@@ -243,10 +244,11 @@ wonder why things aren't working), here's a list of sections which rely on
 external setup (i.e. outside of this config).
 
 + dictionary :: I've downloaded a custom [[http://app.aspell.com/create][SCOWL]] dictionary, which I use in [[*Ispell][ispell]].
-  If this causes issues, just delete the ~(setq ispell-dictionary ...)~ bit.
+  If this causes issues, just delete the src_elisp{(setq ispell-dictionary ...)}
+  bit.
 + uni-units file :: I've got a file in =~/.org/.uni-units= which I use in ~org-capture~
   If this causes issues, just remove the reference to that file in [[*Capture][Capture]] and
-  instances of ~unit-prompt~ used in ~(doct ...)~
+  instances of ~unit-prompt~ used in src_elisp{(doct ...)}.
 
 Oh, did I mention that I started this config when I didn't know any =elisp=, and
 this whole thing is a hack job? If you can suggest any improvements, please do
@@ -260,7 +262,8 @@ few extras.
 + A [[https://www.tug.org/texlive/][LaTeX Compiler]] is required for the mathematics rendering performed in [[#org][Org]],
   and by [[*CalcTeX][CalcTeX]].
 + I use the [[https://overpassfont.org/][Overpass]] font as a go-to sans serif.
-  It's used as my ~doom-variable-pitch-font~ and in the graph generated by [[*Roam][Roam]].
+  It's used as my ~doom-variable-pitch-font~ and in the graph generated
+  by [[*Roam][Roam]].
   I have chosen it because it possesses a few characteristics I consider
   desirable, namely:
   - A clean, and legible style. Highway-style fonts tend to be designed to be
@@ -321,8 +324,8 @@ Apparently this is used by ~GPG~, and all sorts of other things.
 Speaking of ~GPG~, I want to use =~/.authsource.gpg= instead of the default in
 =~/.emacs.d=. Why? Because my home directory is already cluttered, so this won't
 make a difference, and I don't want to accidentaly purge this file (I have done
-~rm -rf~/.emac.d~ before). I also want to cache as much as possible, as my home
+src_shell{rm -rf~/.emac.d~ before}. I also want to cache as much as possible, as
-machine is pretty safe, and my laptop is shutdown a lot.
+my home machine is pretty safe, and my laptop is shutdown a lot.
 #+begin_src emacs-lisp
 (setq auth-sources '("~/.authinfo.gpg")
       auth-source-cache-expiry nil) ; default is 7200 (2h)
@@ -330,7 +333,9 @@ machine is pretty safe, and my laptop is shutdown a lot.
 ** Better defaults
 *** Simple settings
 Browsing the web and seeing [[https://github.com/angrybacon/dotemacs/blob/master/dotemacs.org#use-better-defaults][angrybacon/dotemacs]] and comparing with the values
-shown by =SPC h v= and selecting what I thought looks good, I've ended up adding the following:
+shown by =SPC h v= and selecting what I thought looks good, I've ended up adding
+the following:
+
 #+begin_src emacs-lisp
 (setq-default
  delete-by-moving-to-trash t                      ; Delete files to trash
@@ -353,8 +358,8 @@ shown by =SPC h v= and selecting what I thought looks good, I've ended up adding
 #+end_src
 *** Frame sizing
 It's nice to control the size of new frames, when launching Emacs that can be
-done with ~emacs -geometry 160x48~. After the font size adjustment during
+done with src_shell{emacs -geometry 160x48}. After the font size adjustment
-initialisation this works out to be ~102x31~.
+during initialisation this works out to be ~102x31~.
 
 Thanks to hotkeys, it's easy for me to expand a frame to half/full-screen, so it
 makes sense to be conservative with the sizing of new frames.
@@ -378,17 +383,20 @@ setting, and load it if it exists.
 *** Windows
 I find it rather handy to be asked which buffer I want to see after splitting
 the window. Let's make that happen.
+
 First, we'll enter the new window
 #+begin_src emacs-lisp
 (setq evil-vsplit-window-right t
       evil-split-window-below t)
 #+end_src
+
 Then, we'll pull up ~ivy~
 #+begin_src emacs-lisp
 (defadvice! prompt-for-buffer (&rest _)
   :after '(evil-window-split evil-window-vsplit)
   (+ivy/switch-buffer))
 #+end_src
+
 Oh, and previews are nice
 #+begin_src emacs-lisp
 (setq +ivy-buffer-preview t)
@@ -481,8 +489,9 @@ have flags applied to tweak their behaviour.
 As you may have noticed by this point, this is a [[https://en.wikipedia.org/wiki/Literate_programming][literate]] configuration. Doom
 has good support for this which we access though the ~literate~ module.
 
-While we're in the ~:config~ section, we'll use Dooms nicer defaults, along with
+While we're in the src_elisp{:config} section, we'll use Dooms nicer defaults,
-the bindings and smartparens behaviour (the flags aren't documented, but they exist).
+along with the bindings and smartparens behaviour (the flags aren't documented,
+but they exist).
 #+name: doom-config
 #+begin_src emacs-lisp
 literate
@@ -829,8 +838,8 @@ that Org doesn't try to confirm that I want to allow evaluation (I do!).
 Thankfully Doom supports =$DOOMDIR/cli.el= file which is sourced every time a CLI
 command is run, so we can just enable evaluation by setting
 ~org-confirm-babel-evaluate~ to ~nil~ there.
-While we're at it, we should silence ~org-babel-execute-src-block~ to avoid
+While we're at it, we should silence ~org-babel-execute-src-block~ to
-polluting the output.
+avoid polluting the output.
 
 #+begin_src emacs-lisp :tangle cli.el :comments none
 ;;; cli.el -*- lexical-binding: t; -*-
@@ -1267,7 +1276,7 @@ This is where you install packages, by declaring them with the ~package!~
 macro, then running ~doom refresh~ on the command line. You'll need to
 restart Emacs for your changes to take effect! Or at least, run =M-x doom/reload=.
 
-WARNING: Don't disable core packages listed in ~~/.emacs.d/core/packages.el~.
+*Warning*: Don't disable core packages listed in =~/.emacs.d/core/packages.el=.
 Doom requires these, and disabling them may have terrible side effects.
 
 *** Packages in MELPA/ELPA/emacsmirror
@@ -1874,7 +1883,7 @@ let's just switch to an ~o~, which still looks decent but doesn't cause any
 issues.
 A 'active-bar' is nice, so let's have one of those. If we have it ~under~ needs us to
 turn on ~x-underline-at-decent~ though. For some reason this didn't seem to work
-inside the ~(after! ... )~ block ¯\_(ツ)_/¯.
+inside the src_elisp{(after! ... )} block ¯\_(ツ)_/¯.
 Then let's change the font to a sans serif, but the default one doesn't fit too
 well somehow, so let's switch to 'P22 Underground Book'; it looks much nicer.
 #+begin_src emacs-lisp
@@ -2100,8 +2109,8 @@ nearby (also means that if I change the 'main' dictionary I keep my addition).
 #+end_src
 ** Ivy
 While in an ivy mini-buffer =C-o= shows a list of all possible actions one may take.
-By default this is ~#'ivy-read-action-by-key~ however a better interface to this
+By default this is src_elisp{#'ivy-read-action-by-key} however a better
-is using Hydra.
+interface to this is using Hydra.
 #+begin_src emacs-lisp
 (setq ivy-read-action-function #'ivy-hydra-read-action)
 #+end_src
@@ -2287,7 +2296,8 @@ Let's make this popup a bit faster
 #+begin_src emacs-lisp
 (setq which-key-idle-delay 0.5) ;; I need the help, I really do
 #+end_src
-I also think that having =evil-= appear in so many popups is a bit too verbose, let's change that, and do a few other similar tweaks while we're at it.
+I also think that having =evil-= appear in so many popups is a bit too verbose,
+let's change that, and do a few other similar tweaks while we're at it.
 #+begin_src emacs-lisp
 (setq which-key-allow-multiple-replacements t)
 (after! which-key
@@ -2867,7 +2877,8 @@ omitting the =#= in channel names we can have a list of channels comma-separated
                       accounts))))
 #+end_src
 
-We'll just call ~(register-irc-auths)~ on a hook when we start Circe up.
+We'll just call src_elisp{(register-irc-auths)} on a hook when we start Circe
+up.
 
 Now we're ready to go, let's actually wire-up Circe, with one or two
 configuration tweaks.
@@ -2907,7 +2918,7 @@ configuration tweaks.
 #+end_src
 
 *** Org-style emphasis
-Let's do our *bold*, /italic/, and _underline_ in org-syntax, using IRC control characters
+Let's do our *bold*, /italic/, and _underline_ in org-syntax, using IRC control characters.
 #+name: org-emph-to-irc
 #+begin_src emacs-lisp
 (defun lui-org-to-irc ()
@@ -3469,8 +3480,8 @@ When run without flags this will perform the following actions
   - ~Patterns~
 + Call ~mbsync --list ACCOUNT~, and filter results according to ~Patterns~
 + Construct a imapnotify config for each account, with the following hooks
-  - onNewMail :: ~mbsync --pull ACCOUNT:MAILBOX~
+  - onNewMail :: src_shell{mbsync --pull ACCOUNT:MAILBOX}
-  - onNewMailPost :: ~if mu index --lazy-check; then test -f /tmp/mu_reindex_now && rm /tmp/mu_reindex_now; else touch /tmp/mu_reindex_now; fi~
+  - onNewMailPost :: src_shell{if mu index --lazy-check; then test -f /tmp/mu_reindex_now && rm /tmp/mu_reindex_now; else touch /tmp/mu_reindex_now; fi}
 + Compare accounts list to previous accounts, enable/disable the relevant
   systemd services, called with the ~--now~ flag (start/stop services as well)
 
@@ -4214,7 +4225,7 @@ figure.png╶─╧─▶ PROJECT.ORG ▶───╴filters╶───╧─
 :end:
 
 Finally, because this section is fairly expensive to initialise, we'll wrap it
-in an ~(after! ...)~ block.
+in an src_elisp{(after! ...)} block.
 #+begin_src emacs-lisp :noweb no-export :tangle yes :noweb-ref nil
 (after! org
   <<org-conf>>
@@ -4340,7 +4351,8 @@ Entries of the form (subject . id)."
       org-export-with-sub-superscripts '{}        ; don't treat lone _ / ^ as sub/superscripts, require _{} / ^{}
       org-re-reveal-root "https://cdn.jsdelivr.net/npm/reveal.js")
 #+end_src
-I also like the ~:comments~ header-argument, so let's make that a default.
+I also like the src_elisp{:comments} header-argument, so let's make that a
+default.
 #+begin_src emacs-lisp
 (setq org-babel-default-header-args
       '((:session . "none")
@@ -5895,11 +5907,19 @@ we can apply language-appropriate syntax highlighting. Then, continuing on to
 the value-surrounding constructs hidden by mimicking the behaviour of
 ~prettify-symbols-mode~.
 
+#+begin_warning
+This currently only highlights a single inline src block per line.
+I have no idea why it stops, but I'd rather it didn't.
+If you have any idea what's going on or how to fix this /please/ get in touch.
+#+end_warning
 
 #+begin_src emacs-lisp
 (defvar org-prettify-inline-results t
   "Whether to use (ab)use prettify-symbols-mode on {{{results(...)}}}.")
 
+(defvar org-fontify-inline-src-blocks-max-length 200
+  "Maximum content length of an inline src block that will be fontified.")
+
 (defun org-fontify-inline-src-blocks (limit)
   "Try to apply `org-fontify-inline-src-blocks-1'."
   (condition-case nil
@@ -5909,7 +5929,7 @@ the value-surrounding constructs hidden by mimicking the behaviour of
                     (line-number-at-pos)))))
 
 (defun org-fontify-inline-src-blocks-1 (limit)
-  "Fontify inline src_LANG blocks, from  `point' up to LIMIT."
+  "Fontify inline src_LANG blocks, from `point' up to LIMIT."
   (let ((case-fold-search t))
     (when (re-search-forward "\\_<src_\\([^ \t\n[{]+\\)[{[]?" limit t) ; stolen from `org-element-inline-src-block-parser'
       (let ((beg (match-beginning 0))
@@ -5921,19 +5941,24 @@ the value-surrounding constructs hidden by mimicking the behaviour of
         (font-lock-append-text-property beg lang-beg 'face 'shadow)
         (font-lock-append-text-property beg lang-end 'face 'org-block)
         (setq pt (goto-char lang-end))
-        (when (org-element--parse-paired-brackets ?\[)
+        ;; `org-element--parse-paired-brackets' doesn't take a limit, so to
-          (remove-text-properties pt (point) '(face nil))
+        ;; prevent it searching the entire rest of the buffer we temporarily
-          (font-lock-append-text-property pt (point) 'face 'org-block)
+        ;; narrow the active region.
-          (setq pt (point)))
+        (save-restriction
-        (when (org-element--parse-paired-brackets ?\{)
+          (narrow-to-region beg (min (point-max) limit (+ lang-end org-fontify-inline-src-blocks-max-length)))
-          (remove-text-properties pt (point) '(face nil))
+          (when (ignore-errors (org-element--parse-paired-brackets ?\[))
-          (font-lock-append-text-property pt (1+ pt) 'face '(org-block shadow))
+            (remove-text-properties pt (point) '(face nil))
-          (unless (= (1+ pt) (1- (point)))
+            (font-lock-append-text-property pt (point) 'face 'org-block)
-            (if org-src-fontify-natively
+            (setq pt (point)))
-                (org-src-font-lock-fontify-block (buffer-substring-no-properties lang-beg lang-end) (1+ pt) (1- (point)))
+          (when (ignore-errors (org-element--parse-paired-brackets ?\{))
-              (font-lock-append-text-property (1+ pt) (1- (point)) 'face 'org-block)))
+            (remove-text-properties pt (point) '(face nil))
-          (font-lock-append-text-property (1- (point)) (point) 'face '(org-block shadow))
+            (font-lock-append-text-property pt (1+ pt) 'face '(org-block shadow))
-          (setq pt (point)))
+            (unless (= (1+ pt) (1- (point)))
+              (if org-src-fontify-natively
+                  (org-src-font-lock-fontify-block (buffer-substring-no-properties lang-beg lang-end) (1+ pt) (1- (point)))
+                (font-lock-append-text-property (1+ pt) (1- (point)) 'face 'org-block)))
+            (font-lock-append-text-property (1- (point)) (point) 'face '(org-block shadow))
+            (setq pt (point))))
         (when (and org-prettify-inline-results (re-search-forward "\\= {{{results(" limit t))
           (font-lock-append-text-property pt (1+ pt) 'face 'org-block)
           (goto-char pt))))
@@ -6488,8 +6513,8 @@ the vast majority of the change in behaviour comes from switch statements in:
 #+end_src
 
 There are quite a few instances where I want to modify variables defined in
-=ox-html=, so we'll wrap the contents of this section in an ~(after! ox-html ...)~
+=ox-html=, so we'll wrap the contents of this section in a
-block.
+src_elisp{(after! ox-html ...)} block.
 #+begin_src emacs-lisp :noweb no-export :noweb-ref org-conf
 (after! ox-html
   <<ox-html-conf>>
@@ -6959,8 +6984,8 @@ org provides.
 **** Make verbatim different to code
 Since we have =verbatim= and ~code~, let's make use of the difference.
 
-We can use ~code~ exclusively for code snippets and commands like: "calling ~(message
+We can use ~code~ exclusively for code snippets and commands like: "calling
-"Hello")~ in batch-mode Emacs prints to stdout like ~echo~".
+src_elisp{(message "Hello")} in batch-mode Emacs prints to stdout like ~echo~".
 Then we can use =verbatim= for miscellaneous 'other monospace' like keyboard
 shortcuts: "either =C-c C-c= or =C-g= is likely the most useful keybinding in Emacs",
 or file names: "I keep my configuration in =~/.config/doom/=", among other things.
@@ -7552,12 +7577,12 @@ First we want to process our fancy keywords in ~org-latex-feature-implementation
 to produce an /expanded/ list of features. We'll do that by performing the
 following steps.
 + The dependencies for each listed feature are added to feature list
-  (=:requires=).
+  (src_elisp{:requires}).
-+ The =:when= conditions of each feature, and available features with =:eager t=,
++ The src_elisp{:when} conditions of each feature, and available features with
-  are evaluated, and added/removed accordingly
+  src_elisp{:eager t}, are evaluated, and added/removed accordingly
-+ Any features present in a =:prevents= value are removed
++ Any features present in a src_elisp{:prevents} value are removed
 + The feature list is scrubbed of duplicates
-+ The feature list is sorted by =:order= (ascending)
++ The feature list is sorted by src_elisp{:order} (ascending)
 
 #+begin_src emacs-lisp
 (defun org-latex-expand-features (features)
@@ -7665,10 +7690,11 @@ If \"nil\" no custom fonts will ever be used.")
 Then a function is needed to generate a LaTeX snippet which applies the fontset. It
 would be nice if this could be done for individual styles and use different
 styles as the main document font. If the individual typefaces for a fontset are
-defined individually as (=:serif=, =:sans=, =:mono=, and =:maths=). I can use those to
+defined individually as
-generate LaTeX for subsets of the full fontset. Then, if I don't let any fontset
+src_elisp{:serif}, src_elisp{:sans}, src_elisp{:mono}, and src_elisp{:maths}.
-names have =-= in them, I can use =-sans= and =-mono= as suffixes that specify the
+I can use those to generate LaTeX for subsets of the full fontset. Then, if I
-document font to use.
+don't let any fontset names have =-= in them, I can use =-sans= and =-mono= as
+suffixes that specify the document font to use.
 
 #+begin_src emacs-lisp
 (defun org-latex-fontset-entry ()
@@ -8113,13 +8139,13 @@ better) we add bit to the preamble:
 
 At some point it would be nice to make the box colours easily customisable. At
 the moment it's fairly easy to change the syntax highlighting colours with
-src_emacs-lisp[:eval no :exports code]{(setq engrave-faces-preset-styles
+src_elisp{(setq engrave-faces-preset-styles (engrave-faces-generate-preset))},
-(engrave-faces-generate-preset))}, but perhaps a toggle which specifies whether
+but perhaps a toggle which specifies whether to use the default values, the
-to use the default values, the current theme, or any named theme could be a good
+current theme, or any named theme could be a good idea. It should also possible
-idea. It should also possible to set the box background dynamically to match.
+to set the box background dynamically to match. The named theme could work by
-The named theme could work by looking for a style definition with a certain name
+looking for a style definition with a certain name in a cache dir, and then
-in a cache dir, and then switching to that theme and producing (and saving) the
+switching to that theme and producing (and saving) the style definition if it
-style definition if it doesn't exist.
+doesn't exist.
 
 Now let's have the example block be styled similarly.
 #+begin_src emacs-lisp
@@ -8253,10 +8279,10 @@ fairly easily.
 
 Now we just need to hook this handy function into Org's export.
 We can't use standard string-replacement as we rely on the buffer modifications
-enacted by =(emojify-mode)=.
+enacted by src_elisp{(emojify-mode)}.
 
 As I have not yet implemented a nice way of sharing feature detection
-information outside of =(org-latex-generate-features-preamble)=, we'll
+information outside of src_elisp{(org-latex-generate-features-preamble)}, we'll
 use the same check before attempting to LaTeXify emojis and hope that nothing
 strange happens.