emacs-29 ebef8905b0d: Make indirect buffers use tree-sitter parsers of their base buffer

[Yuan Fu]

branch: emacs-29
commit ebef8905b0df9572e80e20fdc8da7829b9270e3f
Author: Yuan Fu <casouri@gmail.com>
Commit: Yuan Fu <casouri@gmail.com>

    Make indirect buffers use tree-sitter parsers of their base buffer
    
    Fix the problem described in bug#59693.
    
    * src/treesit.c (treesit_record_change): Always use the base buffer.
    (Ftreesit_parser_create): Always use the base buffer.  Also change the
    for loop into FOR_EACH_TAIL (stylistic change).
    (Ftreesit_parser_list): Always use the base buffer.
    
    * doc/lispref/parsing.texi (Using Parser): Update manual.
    * test/src/treesit-tests.el (treesit-indirect-buffer): New test.
---
 doc/lispref/parsing.texi  | 10 +++++++-
 src/treesit.c             | 61 ++++++++++++++++++++++++++++++++++-------------
 test/src/treesit-tests.el | 34 ++++++++++++++++++++++++++
 3 files changed, 87 insertions(+), 18 deletions(-)

diff --git a/doc/lispref/parsing.texi b/doc/lispref/parsing.texi
index 3223875320a..af7be2ebf36 100644
--- a/doc/lispref/parsing.texi
+++ b/doc/lispref/parsing.texi
@@ -409,6 +409,13 @@ Create a parser for the specified @var{buffer} and 
@var{language}
 By default, this function reuses a parser if one already exists for
 @var{language} in @var{buffer}, but if @var{no-reuse} is
 non-@code{nil}, this function always creates a new parser.
+
+If that buffer is an indirect buffer, its base buffer is used instead.
+That is, indirect buffers use their base buffer's parsers.  If the
+base buffer is narrowed, an indirect buffer might not be able to
+retrieve information of the portion of the buffer text that are
+invisible in the base buffer.  Lisp programs should widen as necessary
+should they want to use a parser in an indirect buffer.
 @end defun
 
 Given a parser, we can query information about it.
@@ -447,7 +454,8 @@ tree incrementally.
 @defun treesit-parser-list &optional buffer
 This function returns the parser list of @var{buffer}.  If
 @var{buffer} is @code{nil} or omitted, it defaults to the current
-buffer.
+buffer.  If that buffer is an indirect buffer, its base buffer is used
+instead.  That is, indirect buffers use their base buffer's parsers.
 @end defun
 
 @defun treesit-parser-delete parser
diff --git a/src/treesit.c b/src/treesit.c
index 8b485ca4ece..d361a3da932 100644
--- a/src/treesit.c
+++ b/src/treesit.c
@@ -384,7 +384,18 @@ init_treesit_functions (void)
    mysteriously drops.  3) what if a user uses so many stuff that the
    default cache size (20) is not enough and we end up thrashing?
    These are all imaginary scenarios but they are not impossible
-   :-) */
+   :-)
+
+   Parsers in indirect buffers: We make indirect buffers to share the
+   parser of its base buffer.  Indirect buffers and their base buffer
+   share the same buffer content but not other buffer attributes.  If
+   they have separate parser lists, changes made in an indirect buffer
+   will only update parsers of that indirect buffer, and not parsers
+   in the base buffer or other indirect buffers, and vice versa.  We
+   could keep track of all the base and indirect buffers, and update
+   all of their parsers, but ultimately decide to take a simpler
+   approach, which is to make indirect buffers share their base
+   buffer's parser list.  The discussion can be found in bug#59693.  */
 
 
 /*** Initialization */
@@ -697,9 +708,10 @@ void
 treesit_record_change (ptrdiff_t start_byte, ptrdiff_t old_end_byte,
                       ptrdiff_t new_end_byte)
 {
-  Lisp_Object parser_list;
-
-  parser_list = BVAR (current_buffer, ts_parser_list);
+  struct buffer *base_buffer = current_buffer;
+  if (current_buffer->base_buffer)
+    base_buffer = current_buffer->base_buffer;
+  Lisp_Object parser_list = BVAR (base_buffer, ts_parser_list);
 
   FOR_EACH_TAIL_SAFE (parser_list)
     {
@@ -1252,12 +1264,16 @@ DEFUN ("treesit-parser-create",
        1, 3, 0,
        doc: /* Create and return a parser in BUFFER for LANGUAGE.
 
-The parser is automatically added to BUFFER's parser list, as
-returned by `treesit-parser-list'.
-LANGUAGE is a language symbol.  If BUFFER is nil or omitted, it
-defaults to the current buffer.  If BUFFER already has a parser for
-LANGUAGE, return that parser, but if NO-REUSE is non-nil, always
-create a new parser.  */)
+The parser is automatically added to BUFFER's parser list, as returned
+by `treesit-parser-list'.  LANGUAGE is a language symbol.  If BUFFER
+is nil or omitted, it defaults to the current buffer.  If BUFFER
+already has a parser for LANGUAGE, return that parser, but if NO-REUSE
+is non-nil, always create a new parser.
+
+If that buffer is an indirect buffer, its base buffer is used instead.
+That is, indirect buffers use their base buffer's parsers.  Lisp
+programs should widen as necessary should they want to use a parser in
+an indirect buffer.  */)
   (Lisp_Object language, Lisp_Object buffer, Lisp_Object no_reuse)
 {
   treesit_initialize ();
@@ -1271,16 +1287,21 @@ create a new parser.  */)
       CHECK_BUFFER (buffer);
       buf = XBUFFER (buffer);
     }
+  if (buf->base_buffer)
+    buf = buf->base_buffer;
+
   treesit_check_buffer_size (buf);
 
   /* See if we can reuse a parser.  */
-  for (Lisp_Object tail = BVAR (buf, ts_parser_list);
-       NILP (no_reuse) && !NILP (tail);
-       tail = XCDR (tail))
+  if (NILP (no_reuse))
     {
-      struct Lisp_TS_Parser *parser = XTS_PARSER (XCAR (tail));
-      if (EQ (parser->language_symbol, language))
-       return XCAR (tail);
+      Lisp_Object tail = BVAR (buf, ts_parser_list);
+      FOR_EACH_TAIL (tail)
+      {
+       struct Lisp_TS_Parser *parser = XTS_PARSER (XCAR (tail));
+       if (EQ (parser->language_symbol, language))
+         return XCAR (tail);
+      }
     }
 
   /* Load language.  */
@@ -1329,7 +1350,10 @@ DEFUN ("treesit-parser-list",
        Ftreesit_parser_list, Streesit_parser_list,
        0, 1, 0,
        doc: /* Return BUFFER's parser list.
-BUFFER defaults to the current buffer.  */)
+
+BUFFER defaults to the current buffer.  If that buffer is an indirect
+buffer, its base buffer is used instead.  That is, indirect buffers
+use their base buffer's parsers.  */)
   (Lisp_Object buffer)
 {
   struct buffer *buf;
@@ -1340,6 +1364,9 @@ BUFFER defaults to the current buffer.  */)
       CHECK_BUFFER (buffer);
       buf = XBUFFER (buffer);
     }
+  if (buf->base_buffer)
+    buf = buf->base_buffer;
+
   /* Return a fresh list so messing with that list doesn't affect our
      internal data.  */
   Lisp_Object return_list = Qnil;
diff --git a/test/src/treesit-tests.el b/test/src/treesit-tests.el
index aba12759c34..1cc2217bd3b 100644
--- a/test/src/treesit-tests.el
+++ b/test/src/treesit-tests.el
@@ -161,6 +161,40 @@
       (should (treesit-node-eq root-node root-node))
       (should (not (treesit-node-eq root-node doc-node))))))
 
+(ert-deftest treesit-indirect-buffer ()
+  "Tests for indirect buffers."
+  (skip-unless (treesit-language-available-p 'json))
+  (let ((base (get-buffer-create "*treesit test*"))
+        parser indirect)
+    (unwind-protect
+        (progn
+          (with-current-buffer base
+            (setq indirect (clone-indirect-buffer "*treesit test 1*" nil)))
+          (with-current-buffer indirect
+            (setq parser (treesit-parser-create 'json)))
+          ;; 1. Parser created in the indirect buffer should be
+          ;; actually be created in the base buffer.
+          (with-current-buffer base
+            (should (equal (list parser)
+                           (treesit-parser-list)))
+            (insert "[1,2,3]"))
+          ;; Change in the base buffer should be reflected in the
+          ;; indirect buffer.
+          (with-current-buffer indirect
+            (should (eq (treesit-node-end
+                         (treesit-buffer-root-node))
+                        8))
+            (erase-buffer))
+          ;; Change in the indirect buffer should be reflected in the
+          ;; base buffer.
+          (with-current-buffer base
+            (should (eq (treesit-node-end
+                         (treesit-buffer-root-node))
+                        1))
+            (erase-buffer)))
+      (kill-buffer base)
+      (kill-buffer indirect))))
+
 (ert-deftest treesit-query-api ()
   "Tests for query API."
   (skip-unless (treesit-language-available-p 'json))