emacs-29 31b58161bb5: Fix FOR_EACH_TAIL in c-ts-mode (bug#62951)

[Yuan Fu]

branch: emacs-29
commit 31b58161bb5e2afd30ade92854241976ee71bae0
Author: Yuan Fu <casouri@gmail.com>
Commit: Yuan Fu <casouri@gmail.com>

    Fix FOR_EACH_TAIL in c-ts-mode (bug#62951)
    
    * lisp/progmodes/c-ts-mode.el
    (c-ts-mode--indent-styles): New indent rule.
    
    (c-ts-mode--for-each-tail-regexp)
    (c-ts-mode--for-each-tail-body-matcher)
    (c-ts-mode--emacs-c-range-query)
    (c-ts-mode--for-each-tail-ranges)
    (c-ts-mode--reverse-ranges)
    (c-ts-mode--emacs-set-ranges): New functions and variables.
    
    (c-ts-mode): Create a emacs-c parser.  More setup for Emacs source
    support.
    
    * lisp/treesit.el (treesit-query-range): Ignore underscore-prefixed
    capture names.
---
 lisp/progmodes/c-ts-mode.el | 99 ++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 97 insertions(+), 2 deletions(-)

diff --git a/lisp/progmodes/c-ts-mode.el b/lisp/progmodes/c-ts-mode.el
index 6100f00e3ba..8f7a0f26b44 100644
--- a/lisp/progmodes/c-ts-mode.el
+++ b/lisp/progmodes/c-ts-mode.el
@@ -357,7 +357,9 @@ PARENT, BOL, ARGS are the same as other anchor functions."
   "Indent rules supported by `c-ts-mode'.
 MODE is either `c' or `cpp'."
   (let ((common
-         `(((parent-is "translation_unit") column-0 0)
+         `((c-ts-mode--for-each-tail-body-matcher prev-line 
c-ts-mode-indent-offset)
+
+           ((parent-is "translation_unit") column-0 0)
            ((query "(ERROR (ERROR)) @indent") column-0 0)
            ((node-is ")") parent 1)
            ((node-is "]") parent-bol 0)
@@ -969,6 +971,81 @@ if `c-ts-mode-emacs-sources-support' is non-nil."
   (or (treesit-add-log-current-defun)
       (c-ts-mode--defun-name (c-ts-mode--emacs-defun-at-point))))
 
+;;; FOR_EACH_TAIL fix
+;;
+;; FOR_EACH_TAIL (and FOR_EACH_TAIL_SAFE) followed by a unbracketed
+;; body will mess up the parser, which parses the thing as a function
+;; declaration.  We "fix" it by adding a shadow parser, emacs-c (which
+;; is just c but under a different name).  We use emacs-c to find each
+;; FOR_EACH_TAIL with a unbracketed body, and set the ranges of the C
+;; parser so that it skips those FOR_EACH_TAIL's.  Note that we only
+;; ignore FOR_EACH_TAIL's with a unbracketed body.  Those with a
+;; bracketed body parses more or less fine.
+
+(defvar c-ts-mode--for-each-tail-regexp
+  (rx "FOR_EACH_" (or "TAIL" "TAIL_SAFE" "ALIST_VALUE"
+                      "LIVE_BUFFER" "FRAME"))
+  "A regexp matching all the FOR_EACH_TAIL variants.")
+
+(defun c-ts-mode--for-each-tail-body-matcher (_n _p bol &rest _)
+  "A matcher that matches the first line after a FOR_EACH_TAIL.
+For BOL see `treesit-simple-indent-rules'."
+  (when c-ts-mode-emacs-sources-support
+    (save-excursion
+      (goto-char bol)
+      (forward-line -1)
+      (skip-chars-forward " \t")
+      (looking-at c-ts-mode--for-each-tail-regexp))))
+
+(defvar c-ts-mode--emacs-c-range-query
+  (treesit-query-compile
+   'emacs-c `(((declaration
+                type: (macro_type_specifier
+                       name: (identifier) @_name)
+                @for-each-tail)
+               (:match ,c-ts-mode--for-each-tail-regexp
+                       @_name))))
+  "Query that finds the FOR_EACH_TAIL with a unbracketed body.")
+
+(defvar-local c-ts-mode--for-each-tail-ranges nil
+  "Ranges covering all the FOR_EACH_TAIL's in the buffer.")
+
+(defun c-ts-mode--reverse-ranges (ranges beg end)
+  "Reverse RANGES and return the new ranges between BEG and END.
+Positions that were included RANGES are not in the returned
+ranges, and vice versa.
+
+Return nil if RANGES is nil.  This way, passing the returned
+ranges to `treesit-parser-set-included-ranges' will make the
+parser parse the whole buffer."
+  (if (null ranges)
+      nil
+    (let ((new-ranges nil)
+          (prev-end beg))
+      (dolist (range ranges)
+        (when (< prev-end (car range))
+          (push (cons prev-end (car range)) new-ranges))
+        (setq prev-end (cdr range)))
+      (when (< prev-end end)
+        (push (cons prev-end end) new-ranges))
+      (nreverse new-ranges))))
+
+(defun c-ts-mode--emacs-set-ranges (beg end)
+  "Set ranges for the C parser to skip some FOR_EACH_TAIL's.
+BEG and END are described in `treesit-range-rules'."
+  (let* ((c-parser (treesit-parser-create 'c))
+         (old-ranges c-ts-mode--for-each-tail-ranges)
+         (new-ranges (treesit-query-range
+                      'emacs-c c-ts-mode--emacs-c-range-query beg end))
+         (set-ranges (treesit--clip-ranges
+                      (treesit--merge-ranges
+                       old-ranges new-ranges beg end)
+                      (point-min) (point-max)))
+         (reversed-ranges (c-ts-mode--reverse-ranges
+                           set-ranges (point-min) (point-max))))
+    (setq-local c-ts-mode--for-each-tail-ranges set-ranges)
+    (treesit-parser-set-included-ranges c-parser reversed-ranges)))
+
 ;;; Modes
 
 (defvar-keymap c-ts-base-mode-map
@@ -1072,6 +1149,17 @@ in your configuration."
   :after-hook (c-ts-mode-set-modeline)
 
   (when (treesit-ready-p 'c)
+    ;; Add a fake "emacs-c" language which is just C.  Used for
+    ;; skipping FOR_EACH_TAIL, see `c-ts-mode--emacs-set-ranges'.
+    (setf (alist-get 'emacs-c treesit-load-name-override-list)
+          '("libtree-sitter-c" "tree_sitter_c"))
+    ;; If Emacs source support is enabled, make sure emacs-c parser is
+    ;; after c parser in the parser list. This way various tree-sitter
+    ;; functions will automatically use the c parser rather than the
+    ;; emacs-c parser.
+    (when c-ts-mode-emacs-sources-support
+      (treesit-parser-create 'emacs-c))
+
     (treesit-parser-create 'c)
     ;; Comments.
     (setq-local comment-start "/* ")
@@ -1085,9 +1173,16 @@ in your configuration."
     (setq-local treesit-defun-tactic 'top-level)
     (treesit-major-mode-setup)
 
+    ;; Emacs source support: handle DEFUN and FOR_EACH_TAIL gracefully.
     (when c-ts-mode-emacs-sources-support
       (setq-local add-log-current-defun-function
-                  #'c-ts-mode--emacs-current-defun-name))))
+                  #'c-ts-mode--emacs-current-defun-name)
+
+      (setq-local treesit-range-settings
+                  (treesit-range-rules 'c-ts-mode--emacs-set-ranges))
+
+      (setq-local treesit-language-at-point-function
+                  (lambda (_pos) 'c)))))
 
 ;;;###autoload
 (define-derived-mode c++-ts-mode c-ts-base-mode "C++"