Fix some SGML syntax edge cases (Bug#33887)

* lisp/textmodes/sgml-mode.el (sgml-syntax-propertize-rules): Handle
single and double quotes symmetrically.  Don't skip quoted comment
enders.
* test/lisp/textmodes/sgml-mode-tests.el (sgml-tests--quotes-syntax):
Add more test cases.
(sgml-mode-quote-in-long-text): New test.

diff --git a/lisp/textmodes/sgml-mode.el b/lisp/textmodes/sgml-mode.el
index 75f20722b05fc406ecff2effe813b81b627a6bbd..1df7e78afc5a0cdfb48655f02e35adc66dc7641a 100644 (file)
--- a/lisp/textmodes/sgml-mode.el
+++ b/lisp/textmodes/sgml-mode.el
@@ -363,9 +363,12 @@ Any terminating `>' or `/' is not matched.")
      ;; the resulting number of calls to syntax-ppss made it too slow
      ;; (bug#33887), so we're now careful to leave alone any pair
      ;; of quotes that doesn't hold a < or > char, which is the vast majority.
-     ("\\(?:\\(?1:\"\\)[^\"<>]*\\|\\(?1:'\\)[^'\"<>]*\\)"
+     ("\\([\"']\\)[^\"'<>]*"
       (1 (if (eq (char-after) (char-after (match-beginning 0)))
              (forward-char 1)
+           ;; Avoid skipping comment ender.
+           (when (eq (char-after) ?>)
+             (skip-chars-backward "-"))
            ;; Be careful to call `syntax-ppss' on a position before the one
            ;; we're going to change, so as not to need to flush the data we
            ;; just computed.

diff --git a/test/lisp/textmodes/sgml-mode-tests.el b/test/lisp/textmodes/sgml-mode-tests.el
index 1b8965e344057aec0eb6e2f446319fb7fe97a1e3..34d26480a45e341db1b84cf702a6a335313de7b3 100644 (file)
--- a/test/lisp/textmodes/sgml-mode-tests.el
+++ b/test/lisp/textmodes/sgml-mode-tests.el
@@ -161,15 +161,46 @@ The point is set to the beginning of the buffer."
       (should (string= "&&" (buffer-string))))))
 
 (ert-deftest sgml-tests--quotes-syntax ()
+  (dolist (str '("a\"b <t>c'd</t>"
+                 "a'b <t>c\"d</t>"
+                 "<t>\"a'</t>"
+                 "<t>'a\"</t>"
+                 "<t>\"a'\"</t>"
+                 "<t>'a\"'</t>"
+                 "a\"b <tag>c'd</tag>"
+                 "<tag>c>'d</tag>"
+                 "<t><!-- \" --></t>"
+                 "<t><!-- ' --></t>"
+                 ))
+   (with-temp-buffer
+     (sgml-mode)
+     (insert str)
+     (ert-info ((format "%S" str) :prefix "Test case: ")
+       ;; Check that last tag is parsed as a tag.
+       (should (= 1 (car (syntax-ppss (1- (point-max))))))
+       (should (= 0 (car (syntax-ppss (point-max)))))))))
+
+(ert-deftest sgml-mode-quote-in-long-text ()
   (with-temp-buffer
     (sgml-mode)
-    (insert "a\"b <tag>c'd</tag>")
-    (should (= 1 (car (syntax-ppss (1- (point-max))))))
-    (should (= 0 (car (syntax-ppss (point-max)))))
-    (erase-buffer)
-    (insert "<tag>c>d</tag>")
-    (should (= 1 (car (syntax-ppss (1- (point-max))))))
-    (should (= 0 (car (syntax-ppss (point-max)))))))
+    (insert "<t>"
+            ;; `syntax-propertize-wholelines' extends chunk size based
+            ;; on line length, so newlines are significant!
+            (make-string syntax-propertize-chunk-size ?a) "\n"
+            "'"
+            (make-string syntax-propertize-chunk-size ?a) "\n"
+            "</t>")
+    ;; If we just check (syntax-ppss (point-max)) immediately, then
+    ;; we'll end up propertizing the whole buffer in one chunk (so the
+    ;; test is useless).  Simulate something more like what happens
+    ;; when the buffer is viewed normally.
+    (cl-loop for pos from (point-min) to (point-max)
+             by syntax-propertize-chunk-size
+             do (syntax-ppss pos))
+    (syntax-ppss (point-max))
+    ;; Check that last tag is parsed as a tag.
+    (should (= 1 (- (car (syntax-ppss (1- (point-max))))
+                    (car (syntax-ppss (point-max))))))))
 
 (provide 'sgml-mode-tests)
 ;;; sgml-mode-tests.el ends here