bug#8036: [PATCH] Prevent hide-show mode from getting confused by braces in comments

bug#8036: [PATCH] Prevent hide-show mode from getting confused by braces in comments

[Dima Kogan]

[-- Attachment #1: Type: text/plain, Size: 519 bytes --]

Emacs bug #700 (http://debbugs.gnu.org/cgi/bugreport.cgi?bug=700)
contained a patch by Dmitry Bolshakov addressing 2 issues in hide-show
minor mode. The fix to one of the issues got integrated into the code
base, but the other one fell through the cracks, I think. The issue it
fixes is to prevent hide-show mode from parsing braces inside comments.
These braces can confuse the parser, since they don't necessarily match
up.

I'm attaching Dmitry's test case that demonstrates the issue, and his patch to
fix it.

dima

[-- Attachment #2: unbalanced-braces-test.pl --]
[-- Type: application/x-perl, Size: 155 bytes --]

[-- Attachment #3: hs.patch --]
[-- Type: text/x-patch, Size: 1258 bytes --]

--- hideshow.el.original_emacs_22.2	2008-01-10 12:15:42.000000000 -0800
+++ hideshow_my.el	2008-08-12 15:39:12.000000000 -0700
@@ -541,7 +541,10 @@
              (q
               ;; `q' is the point at the end of the block
               (progn (hs-forward-sexp mdata 1)
-                     (end-of-line)
+					 ;; see http://www.emacswiki.org/cgi-bin/wiki/HideShow and search page for "hs-block-includes-last-line"
+					 ;(forward-line -1)
+                     ;(end-of-line)
+					 (backward-char 1) ; better
                      (point)))
              ov)
         (when (and (< p (point)) (> (count-lines p q) 1))
@@ -672,10 +675,13 @@
            (forward-comment (buffer-size))
            (and (< (point) maxp)
                 (re-search-forward hs-block-start-regexp maxp t)))
-    (if (> arg 1)
-        (hs-hide-level-recursive (1- arg) minp maxp)
-      (goto-char (match-beginning hs-block-start-mdata-select))
-      (hs-hide-block-at-point t)))
+	(when (not (nth 4 (syntax-ppss))) ; not inside comments
+	  (if (> arg 1)
+		  (hs-hide-level-recursive (1- arg) minp maxp)
+		(goto-char (match-beginning hs-block-start-mdata-select))
+		(hs-hide-block-at-point t))
+	  )
+	)
   (goto-char maxp))
 
 (defmacro hs-life-goes-on (&rest body)

bug#8036: Acknowledgement ([PATCH] Prevent hide-show mode from getting confused by braces in comments)

[Dima Kogan]

[-- Attachment #1: Type: text/plain, Size: 449 bytes --]

I accidentally attached the Dmitry's original patch file, that
contains both the merged and unmerged fixes. Further, it appears that
Dmitry's original patch handles the case where hs-hide-all is called
in the attached sample perl script, BUT it does not handle the case
where hs-hide-level is called with the point at the end of the
buffer. I'm attaching a new patch that ONLY contains the pertinent
changes and that addresses the second issue also.

[-- Attachment #2: hideshow-comments.patch --]
[-- Type: text/x-patch, Size: 1188 bytes --]

--- hideshow.orig.el	2011-02-13 23:01:07.000000000 -0800
+++ hideshow.el	2011-02-14 20:11:17.000000000 -0800
@@ -690,6 +690,7 @@
         (point)
       ;; look backward for the start of a block that contains the cursor
       (while (and (re-search-backward hs-block-start-regexp nil t)
+                  (save-match-data (not (nth 4 (syntax-ppss)))) ; not inside comments
                   (not (setq done
                              (< here (save-excursion
                                        (hs-forward-sexp (match-data t) 1)
@@ -712,10 +713,11 @@
            (forward-comment (buffer-size))
            (and (< (point) maxp)
                 (re-search-forward hs-block-start-regexp maxp t)))
-    (if (> arg 1)
-        (hs-hide-level-recursive (1- arg) minp maxp)
-      (goto-char (match-beginning hs-block-start-mdata-select))
-      (hs-hide-block-at-point t)))
+    (when (not (nth 4 (syntax-ppss))) ; not inside comments
+      (if (> arg 1)
+          (hs-hide-level-recursive (1- arg) minp maxp)
+        (goto-char (match-beginning hs-block-start-mdata-select))
+        (hs-hide-block-at-point t))))
   (goto-char maxp))
 
 (defmacro hs-life-goes-on (&rest body)

bug#8036: [PATCH] Prevent hide-show mode from getting confused by braces in comments

[Dima Kogan]

[-- Attachment #1: Type: text/plain, Size: 122 bytes --]

I have discovered a case where the previous patch causes an infinite
loop. The new patch I'm attaching resolves that issue

[-- Attachment #2: hideshow-comments.patch --]
[-- Type: text/x-patch, Size: 1206 bytes --]

--- hideshow.orig.el	2011-02-13 23:01:07.000000000 -0800
+++ hideshow.el	2011-02-15 20:58:14.000000000 -0800
@@ -690,6 +690,7 @@
         (point)
       ;; look backward for the start of a block that contains the cursor
       (while (and (re-search-backward hs-block-start-regexp nil t)
+                  (save-match-data (not (nth 4 (syntax-ppss)))) ; not inside comments
                   (not (setq done
                              (< here (save-excursion
                                        (hs-forward-sexp (match-data t) 1)
@@ -712,10 +713,11 @@
            (forward-comment (buffer-size))
            (and (< (point) maxp)
                 (re-search-forward hs-block-start-regexp maxp t)))
-    (if (> arg 1)
-        (hs-hide-level-recursive (1- arg) minp maxp)
-      (goto-char (match-beginning hs-block-start-mdata-select))
-      (hs-hide-block-at-point t)))
+    (when (save-match-data (not (nth 4 (syntax-ppss)))) ; not inside comments
+      (if (> arg 1)
+          (hs-hide-level-recursive (1- arg) minp maxp)
+        (goto-char (match-beginning hs-block-start-mdata-select))
+        (hs-hide-block-at-point t))))
   (goto-char maxp))
 
 (defmacro hs-life-goes-on (&rest body)

bug#8036: [PATCH] Prevent hide-show mode from getting confused by braces in comments

[Chong Yidong]

Dima Kogan <dkogan@cds.caltech.edu> writes:

> I have discovered a case where the previous patch causes an infinite
> loop. The new patch I'm attaching resolves that issue

Looks reasonable.  Committed, and thanks.