[COMMIT] x 2; Re: Still a few bugs in the system...

 Ar an séú lá déag de mí Márta, scríobh Vin Shelton: 

 > After a few days away, I had some problems with my latest 21.5 builds:

Thank you Vin, these both now work for me with corresponding systems and
configure lines, please get back to me if there are still issues.

APPROVE COMMIT

Note: these two patches have been committed.

# HG changeset patch
# User Aidan Kehoe <[email protected]>
# Date 1426552179 0
#      Tue Mar 17 00:29:39 2015 +0000
# Node ID df50aaeddca5679fba85dc8b1ebc8fd708c02028
# Parent  04b8549344077926229ea11d7f5d96967d03441c
Don't tickle an unreasonable but standard C lib limitation, thank you Vin!

src/ChangeLog addition:

2015-03-17  Aidan Kehoe  <[email protected]>

	* data.c (fill_ichar_array):
	Be a bit more careful never to supply the C library's isupper()
	with a value outside of that range specified by the C
	standard. Thank you Vin Shelton for reporting the Linux
	segmentation violation this provokes, and happy St. Patrick's Day
	to everyone reading the XEmacs lists!

diff -r 04b854934407 -r df50aaeddca5 src/ChangeLog
--- a/src/ChangeLog	Mon Mar 16 09:03:22 2015 +0000
+++ b/src/ChangeLog	Tue Mar 17 00:29:39 2015 +0000
@@ -1,3 +1,12 @@
+2015-03-17  Aidan Kehoe  <[email protected]>
+
+	* data.c (fill_ichar_array):
+	Be a bit more careful never to supply the C library's isupper()
+	with a value outside of that range specified by the C
+	standard. Thank you Vin Shelton for reporting the Linux
+	segmentation violation this provokes, and happy St. Patrick's Day
+	to everyone reading the XEmacs lists!
+
 2015-03-08  Aidan Kehoe  <[email protected]>
   
 	* data.c (Fparse_integer):
diff -r 04b854934407 -r df50aaeddca5 src/data.c
--- a/src/data.c	Mon Mar 16 09:03:22 2015 +0000
+++ b/src/data.c	Tue Mar 17 00:29:39 2015 +0000
@@ -1484,9 +1484,15 @@
 	}
       /* Maybe our own case infrastructure is not available yet. Use the C
          library's. */
-      else if (isupper (range->ch) && !isupper (cctable[valint]))
+      else if (current_buffer == NULL)
         {
-	  cctable[valint] = range->ch;          
+	  /* The C library can't necessarily handle values outside of
+	     the range EOF to CHAR_MAX, inclusive. */
+	  assert (range->ch == EOF || range->ch <= CHAR_MAX);
+	  if (isupper (range->ch) && !isupper (cctable[valint]))
+	    {
+	      cctable[valint] = range->ch;
+	    }
         }
       /* Otherwise, save it if this character has a numerically lower value
          (preferring ASCII over fullwidth Chinese and so on). */

# HG changeset patch
# User Aidan Kehoe <[email protected]>
# Date 1426553268 0
#      Tue Mar 17 00:47:48 2015 +0000
# Node ID 58e72e27fb816c54211fcdaeec7729b97bcdae9c
# Parent  df50aaeddca5679fba85dc8b1ebc8fd708c02028
Don't rely on zero termination for Vfixnum_to_char_map, thank you Vin.

lisp/ChangeLog addition:

	(vars_of_data): Don't rely on zero termination here for
	Vfixnum_to_char_map, give an explicit bytecount. Again, thank you
	Vin!

diff -r df50aaeddca5 -r 58e72e27fb81 src/ChangeLog
--- a/src/ChangeLog	Tue Mar 17 00:29:39 2015 +0000
+++ b/src/ChangeLog	Tue Mar 17 00:47:48 2015 +0000
@@ -6,6 +6,9 @@
 	standard. Thank you Vin Shelton for reporting the Linux
 	segmentation violation this provokes, and happy St. Patrick's Day
 	to everyone reading the XEmacs lists!
+	(vars_of_data): Don't rely on zero termination here for
+	Vfixnum_to_char_map, give an explicit bytecount. Again, thank you
+	Vin!
 
 2015-03-08  Aidan Kehoe  <[email protected]>
   
diff -r df50aaeddca5 -r 58e72e27fb81 src/data.c
--- a/src/data.c	Tue Mar 17 00:29:39 2015 +0000
+++ b/src/data.c	Tue Mar 17 00:47:48 2015 +0000
@@ -4302,7 +4302,9 @@
     /* Sigh, we can't call build_fixnum_to_char_map() on Vdigit_fixnum_map,
        this is too early in the boot sequence to map across a char table. Do
        it by hand. */
-    Vfixnum_to_char_map = build_ascstring (fixnum_tab);
+    ASSERT_ASCTEXT_ASCII_LEN (fixnum_tab, 36 * MAX_ICHAR_LEN);
+    Vfixnum_to_char_map
+	    = make_string ((const Ibyte*) fixnum_tab, 36 * MAX_ICHAR_LEN);
     staticpro (&Vfixnum_to_char_map);
   }
 

-- 
‘Tramadol is further fed to cattle […] when working them […] (as draft
animals) so that the animals do not get tired quickly. …’
	— Angewandte Chemie, Sept 2014, describing the social context of
	(synthetic) tramadol having been found in Cameroon tree roots.

_______________________________________________
XEmacs-Patches mailing list
[email protected]
http://lists.xemacs.org/mailman/listinfo/xemacs-patches