feature/tree-sitter 5cd3db73be 4/4: Improve treesit-node-at

[Yuan Fu]

branch: feature/tree-sitter
commit 5cd3db73bed06e394ea8e7b0e332b1b1e5bd9571
Author: Yuan Fu <casouri@gmail.com>
Commit: Yuan Fu <casouri@gmail.com>

    Improve treesit-node-at
    
    * doc/lispref/parsing.texi (Retrieving Node): Update manual.
    * lisp/treesit.el (treesit-node-at): Change semantic.  It tries to
    return the node that a user would expect in various circumstances.
    * test/src/treesit-tests.el (treesit-node-at): New test.
---
 doc/lispref/parsing.texi  | 28 +++++++++++++++++--------
 lisp/treesit.el           | 53 +++++++++++++++++++++++++++++++----------------
 test/src/treesit-tests.el | 31 +++++++++++++++++++++++++++
 3 files changed, 85 insertions(+), 27 deletions(-)

diff --git a/doc/lispref/parsing.texi b/doc/lispref/parsing.texi
index bc2f0dda91..f690e6a604 100644
--- a/doc/lispref/parsing.texi
+++ b/doc/lispref/parsing.texi
@@ -492,9 +492,22 @@ Using an outdated node signals the 
@code{treesit-node-outdated} error.
 @cindex syntax tree, retrieving nodes
 
 @defun treesit-node-at pos &optional parser-or-lang named
-This function returns the @emph{smallest} node that starts at or after
-the buffer position @var{pos}.  In other words, the start of the node
-is greater or equal to @var{pos}.
+This function returns a @emph{leaf} node at buffer position @var{pos}.
+A leaf node is a node that doesn't have any child nodes.
+
+This function tries to return a node whose span covers @var{pos}: the
+node's beginning is less or equal to @var{pos}, and the node's end is
+greater or equal to @var{pos}.
+
+But if no leaf node's span covers @var{pos} (e.g., @var{pos} is on the
+whitespace between two leaf nodes), this function returns the first
+leaf node after @var{pos}.
+
+Finally, if there is no leaf node after @var{pos}, return the first
+leaf node before @var{pos}.
+
+If @var{pos} is in between two adjacent nodes, this function returns
+the one after @var{pos}.
 
 When @var{parser-or-lang} is @code{nil} or omitted, this function uses
 the first parser in @code{(treesit-parser-list)} of the current
@@ -503,15 +516,12 @@ parser; if @var{parser-or-lang} is a language, it finds 
the first
 parser using that language in @code{(treesit-parser-list)}, and uses
 that.
 
+If this function cannot find a suitable node to return, it returns
+nil.
+
 If @var{named} is non-@code{nil}, this function looks for a named node
 only (@pxref{tree-sitter named node, named node}).
 
-When @var{pos} is after all the text in the buffer, technically there
-is no node after @var{pos}.  But for convenience, this function will
-return the last leaf node in the parse tree.  If @var{strict} is
-non-@code{nil}, this function will strictly comply to the semantics and
-return @var{nil}.
-
 Example:
 
 @example
diff --git a/lisp/treesit.el b/lisp/treesit.el
index 1c61b1efeb..796b85ef74 100644
--- a/lisp/treesit.el
+++ b/lisp/treesit.el
@@ -168,44 +168,61 @@ parser in `treesit-parser-list', or nil if there is no 
parser."
   (treesit-parser-language
    (treesit-node-parser node)))
 
-(defun treesit-node-at (pos &optional parser-or-lang named strict)
-  "Return the smallest node that starts at or after buffer position POS.
+(defun treesit-node-at (pos &optional parser-or-lang named)
+  "Return the leaf node at position POS.
 
-\"Starts at or after POS\" means the start of the node is greater
-than or equal to POS.
+A leaf node is a node that doesn't have any child nodes.
 
-Return nil if none was found.  If NAMED is non-nil, only look for
-named node.
+The returned node's span covers POS: the node's beginning is less
+or equal to POS, and the node's end is greater or equal to POS.
+
+If no leaf node's span covers POS (e.g., POS is on whitespace
+between two leaf nodes), return the first leaf node after POS.
+
+If there is no leaf node after POS, return the first leaf node
+before POS.
+
+If POS is in between two adjacent nodes, return the one after
+POS.
+
+Return nil if no leaf node can be returned.  If NAMED is non-nil,
+only look for named nodes.
 
 If PARSER-OR-LANG is nil, use the first parser in
 `treesit-parser-list'; if PARSER-OR-LANG is a parser, use
 that parser; if PARSER-OR-LANG is a language, find a parser using
-that language in the current buffer, and use that.
-
-If POS is after all the text in the buffer, i.e., there is no
-node after POS, return the last leaf node in the parse tree, even
-though that node is before POS.  If STRICT is non-nil, return nil
-in this case."
+that language in the current buffer, and use that."
   (let* ((root (if (treesit-parser-p parser-or-lang)
                    (treesit-parser-root-node parser-or-lang)
                  (treesit-buffer-root-node parser-or-lang)))
          (node root)
+         (node-before root)
+         (pos-1 (max (1- pos) (point-min)))
          next)
     (when node
       ;; This is very fast so no need for C implementation.
       (while (setq next (treesit-node-first-child-for-pos
                          node pos named))
         (setq node next))
-      ;; If we are at the end of buffer or after all the text, we will
-      ;; end up with NODE = root node.  For convenience, return the last
-      ;; leaf node in the tree.
+      ;; If POS is at the end of buffer, after all the text, we will
+      ;; end up with NODE = root node.  Instead of returning nil,
+      ;; return the last leaf node in the tree for convenience.
       (if (treesit-node-eq node root)
-          (if strict
-              nil
+          (progn
             (while (setq next (treesit-node-child node -1 named))
               (setq node next))
             node)
-        node))))
+        ;; Normal case, where we found a node.
+        (if (<= (treesit-node-start node) pos)
+            node
+          ;; So the node we found is completely after POS, try to find
+          ;; a node whose end equals to POS.
+          (while (setq next (treesit-node-first-child-for-pos
+                             node-before pos-1 named))
+            (setq node-before next))
+          (if (eq (treesit-node-end node-before) pos)
+              node-before
+            node))))))
 
 (defun treesit-node-on (beg end &optional parser-or-lang named)
   "Return the smallest node covering BEG to END.
diff --git a/test/src/treesit-tests.el b/test/src/treesit-tests.el
index 5e4aea3ad4..7fc810492b 100644
--- a/test/src/treesit-tests.el
+++ b/test/src/treesit-tests.el
@@ -474,6 +474,37 @@ visible_end.)"
     ;; `treesit-search-forward-goto'
     ))
 
+(ert-deftest treesit-node-at ()
+  "Test `treesit-node-at'."
+  (skip-unless (treesit-language-available-p 'json))
+  (let (parser root-node)
+    (progn
+      (insert "[1,  2, 3,4]  ")
+      (setq parser (treesit-parser-create 'json))
+      (setq root-node (treesit-parser-root-node
+                       parser)))
+    ;; Point at ",", should return ",".
+    (goto-char (point-min))
+    (search-forward "1")
+    (should (equal (treesit-node-text
+                    (treesit-node-at (point)))
+                   ","))
+    ;; Point behind ",", should still return the ",".
+    (search-forward ",")
+    (should (equal (treesit-node-text
+                    (treesit-node-at (point)))
+                   ","))
+    ;; Point between "," and "2", should return 2.
+    (forward-char)
+    (should (equal (treesit-node-text
+                    (treesit-node-at (point)))
+                   "2"))
+    ;; EOF, should return the last leaf node "]".
+    (goto-char (point-max))
+    (should (equal (treesit-node-text
+                    (treesit-node-at (point)))
+                   "]"))))
+
 (ert-deftest treesit-misc ()
   "Misc helper functions."
   (let ((settings '((t 0 t)