Re: Fallout for VAX (was: --enable-checking for gas)

[Alan Modra <amodra at bigpond dot net dot au>]

On Sat, Apr 21, 2007 at 02:42:36PM +0200, Jan-Benedict Glaw wrote:
> 	* config/atof-vax.c: Update copyright year.
> 	(atof_vax_sizeof): Return unsigned int instead of int.
> 	(md_atof): Make number_of_chars unsigned int, too.
> 	(MAXIMUM_NUMBER_OF_LITTLENUMS): Remove single space.

I just committed a very similar patch, reverting Nick's fix (to avoid
a cast).  I was waiting for testsuite runs to finish..

The only fallout I have now is for mmix, which triggers the assertion
at read.c:709 due to the hack at tc-mmix.c:2991.  I'm inclined to
make the following change, but I'm waiting for another testsuite
run.

	* read.c (read_a_source_file): Skip multiple spaces to
	cover hack in mmix md_start_line_hook which overwrites a
	colon with a space.  Delete sermon and needless assertion.

Index: gas/read.c
===================================================================
RCS file: /cvs/src/src/gas/read.c,v
retrieving revision 1.125
diff -u -p -r1.125 read.c
--- gas/read.c	21 Apr 2007 06:54:56 -0000	1.125
+++ gas/read.c	21 Apr 2007 13:26:01 -0000
@@ -695,18 +695,11 @@ read_a_source_file (char *name)
 
 	     Depending on what compiler is used, the order of these tests
 	     may vary to catch most common case 1st.
-	     Each test is independent of all other tests at the (top) level.
-	     PLEASE make a compiler that doesn't use this assembler.
-	     It is crufty to waste a compiler's time encoding things for this
-	     assembler, which then wastes more time decoding it.
-	     (And communicating via (linear) files is silly!
-	     If you must pass stuff, please pass a tree!)  */
-	  if ((c = *input_line_pointer++) == '\t'
-	      || c == ' '
-	      || c == '\f')
+	     Each test is independent of all other tests at the (top)
+	     level.  */
+	  do
 	    c = *input_line_pointer++;
-
-	  know (c != ' ');	/* No further leading whitespace.  */
+	  while (c == '\t' || c == ' ' || c == '\f');
 
 #ifndef NO_LISTING
 	  /* If listing is on, and we are expanding a macro, then give

-- 
Alan Modra
IBM OzLabs - Linux Technology Centre