master 145aab0: cc-mode: add support for Doxygen documentation style

[Micha? Nazarewicz]

branch: master
commit 145aab0672ae259736ee9230f8e0ff4effa5f4fd
Author: Michal Nazarewicz <address@hidden>
Commit: Michal Nazarewicz <address@hidden>

    cc-mode: add support for Doxygen documentation style
    
    * lisp/progmodes/cc-fonts.el (doxygen-font-lock-doc-comments,
      doxygen-font-lock-keywords): New constants defining Doxygen
      comment style support.
    * lisp/progmodes/cc-vars.el (c-doc-comment-style): Updated docstring
      to mention now-supported Doxygen mode.
---
 etc/NEWS                   | 16 +++++++++-
 lisp/progmodes/cc-fonts.el | 78 ++++++++++++++++++++++++++++++++++++++++++++++
 lisp/progmodes/cc-vars.el  |  1 +
 3 files changed, 94 insertions(+), 1 deletion(-)

diff --git a/etc/NEWS b/etc/NEWS
index 8fc2311..753b7a7 100644
--- a/etc/NEWS
+++ b/etc/NEWS
@@ -198,7 +198,6 @@ like cell phones, tablets or cameras.
 Previously, assigning a new template to an already defined tag had no
 effect.
 
-
 ** map.el
 
 *** Pcase 'map' pattern added keyword symbols abbreviation.
@@ -288,6 +287,21 @@ prefix on the Subject line in various languages.
 These new navigation commands are bound to 'n' and 'p' in
 'apropos-mode'.
 
+** cc-mode
+
+*** Added support for Doxygen documentation style.
+‘doxygen’ is now valid ‘c-doc-comment-style’ which recognises all
+comment styles supported by Doxygen (namely ‘///’, ‘//!’, ‘/** … */’
+and ‘/*! … */’.  ‘gtkdoc’ remains the default for C and C++ modes; to
+use ‘doxygen’ by default one might evaluate:
+
+    (setq-default c-doc-comment-style
+                  '((java-mode . javadoc)
+                    (pike-mode . autodoc)
+                    (c-mode    . doxygen)
+                    (c++-mode  . doxygen)))
+
+or use it in a custom ‘c-style’.
 
 * New Modes and Packages in Emacs 28.1
 
diff --git a/lisp/progmodes/cc-fonts.el b/lisp/progmodes/cc-fonts.el
index 2cbbc66..9ea08a4 100644
--- a/lisp/progmodes/cc-fonts.el
+++ b/lisp/progmodes/cc-fonts.el
@@ -3016,6 +3016,84 @@ need for `pike-font-lock-extra-types'.")
        (c-font-lock-doc-comments "/[*/]!" limit
          autodoc-font-lock-doc-comments)))))
 
+;; Doxygen
+
+(defconst doxygen-font-lock-doc-comments
+  ;; TODO: Handle @code, @verbatim, @dot, @f etc. better by not highlighting
+  ;; text inside of those commands.  Something smarter than just regexes may be
+  ;; needed to do that efficiently.
+  `((,(concat
+       ;; Make sure that the special character has not been escaped.  E.g. in
+       ;; ‘\@foo’ only ‘\@’ is a command (similarly for other characters like
+       ;; ‘\\foo’, ‘\<foo’ and ‘\&foo’).  The downside now is that we don’t
+       ;; match command started just after an escaped character, e.g. in
+       ;; ‘\@\foo’ we should match ‘\@’ as well as ‘\foo’ but only the former
+       ;; is matched.
+       "\\(?:^\\|[^\\@]\\)\\("
+         ;; Doxygen commands start with backslash or an at sign.  Note that for
+         ;; brevity in the comments only ‘\’ will be mentioned.
+         "[\\@]\\(?:"
+           ;; Doxygen commands except those starting with ‘f’
+           "[a-eg-z][a-z]*"
+           ;; Doxygen command starting with ‘f’:
+           "\\|f\\(?:"
+             "[][$}]"                         ; \f$ \f} \f[ \f]
+             "\\|{\\(?:[a-zA-Z]+\\*?}{?\\)?"  ; \f{ \f{env} \f{env}{
+             "\\|[a-z]+"                      ; \foo
+           "\\)"
+           "\\|~[a-zA-Z]*"             ; \~  \~language
+           "\\|[$@&~<=>#%\".|\\\\]"    ; single-character escapes
+           "\\|::\\|---?"              ; \:: \-- \---
+         "\\)"
+         ;; HTML tags and entities:
+         "\\|</?\\sw\\(?:\\sw\\|\\s \\|[=\n\r*.:]\\|\"[^\"]*\"\\|'[^']*'\\)*>"
+         "\\|&\\(?:\\sw+\\|#[0-9]+\\|#x[0-9a-fA-F]+\\);"
+       "\\)")
+     1 ,c-doc-markup-face-name prepend nil)
+    ;; Commands inside of strings are not commands so override highlighting 
with
+    ;; string face.  This also affects HTML attribute values if they are
+    ;; surrounded with double quotes which may or may not be considered a good
+    ;; thing.
+    ("\\(?:^\\|[^\\@]\\)\\(\"[^\"[:cntrl:]]+\"\\)"
+     1 font-lock-string-face prepend nil)
+    ;; HTML comments inside of the Doxygen comments.
+    ("\\(?:^\\|[^\\@]\\)\\(<!--.*?-->\\)"
+     1 font-lock-comment-face prepend nil)
+    ;; Autolinking. Doxygen auto-links anything that is a class name but we 
have
+    ;; no hope of matching those.  We are, however, able to match functions and
+    ;; members using explicit scoped syntax.  For functions, we can also find
+    ;; them by noticing argument-list.  Note that Doxygen accepts ‘::’ as well
+    ;; as ‘#’ as scope operators.
+    (,(let* ((ref "[\\@]ref\\s-+")
+             (ref-opt (concat "\\(?:" ref "\\)?"))
+             (id "[a-zA-Z_][a-zA-Z_0-9]*")
+             (args "\\(?:()\\|([^()]*)\\)")
+             (scope "\\(?:#\\|::\\)"))
+        (concat
+         "\\(?:^\\|[^\\@/%:]\\)\\(?:"
+                 ref-opt "\\(?1:" scope "?" "\\(?:" id scope "\\)+" "~?" id 
"\\)"
+           "\\|" ref-opt "\\(?1:" scope     "~?" id "\\)"
+           "\\|" ref-opt "\\(?1:" scope "?" "~?" id "\\)" args
+           "\\|" ref     "\\(?1:" "~?" id "\\)"
+           "\\|" ref-opt "\\(?1:~[A-Z][a-zA-Z0-9_]+\\)"
+         "\\)"))
+     1 font-lock-function-name-face prepend nil)
+    ;; Match URLs and emails.  This has two purposes.  First of all, Doxygen
+    ;; autolinks URLs.  Second of all, ‘@bar’ in ‘address@hidden’ has been 
matched
+    ;; above as a command; try and overwrite it.
+    (,(let* ((host "[A-Za-z0-9]\\(?:[A-Za-z0-9-]\\{0,61\\}[A-Za-z0-9]\\)")
+             (fqdn (concat "\\(?:" host "\\.\\)+" host))
+             (comp "[!-(*--/-=?-~]+")
+             (path (concat "/\\(?:" comp "[.]+" "\\)*" comp)))
+        (concat "\\(?:mailto:\\)?[a-zA-0_.]+@" fqdn
+                "\\|https?://" fqdn "\\(?:" path "\\)?"))
+     0 font-lock-keyword-face prepend nil)))
+
+(defconst doxygen-font-lock-keywords
+  `((,(lambda (limit)
+        (c-font-lock-doc-comments "/\\(?:/[/!]\\|\\*[\\*!]\\)"
+            limit doxygen-font-lock-doc-comments)))))
+
 
 ;; 2006-07-10:  awk-font-lock-keywords has been moved back to cc-awk.el.
 (cc-provide 'cc-fonts)
diff --git a/lisp/progmodes/cc-vars.el b/lisp/progmodes/cc-vars.el
index 3995b21..b885f6a 100644
--- a/lisp/progmodes/cc-vars.el
+++ b/lisp/progmodes/cc-vars.el
@@ -576,6 +576,7 @@ comment styles:
  javadoc -- Javadoc style for \"/** ... */\" comments (default in Java mode).
  autodoc -- Pike autodoc style for \"//! ...\" comments (default in Pike mode).
  gtkdoc  -- GtkDoc style for \"/** ... **/\" comments (default in C and C++ 
modes).
+ doxygen -- Doxygen style.
 
 The value may also be a list of doc comment styles, in which case all
 of them are recognized simultaneously (presumably with markup cues