Re: Failure of cross-assembling for ARC cpu target

[amodra at one dot net dot au]

On Sun, May 20, 2001 at 08:17:34PM +0200, Angel Manchado wrote:
> On Fri, May 18, 2001 at 01:22:17PM +0930, amodra@one.net.au wrote:
> > On Thu, May 17, 2001 at 10:56:37PM +0200, A. Manchado wrote:
> > > /tmp/cc4ga3g6.s:46: Error: junk at end of line: `ld blink,[fp,4]'
> > 
> > Someone else will have to handle this one.  Peter?  I don't really
> > know much about ARC.
> 
> One possible solution:
> comparing gcc/config/arc/initfini.c between gcc-2.95.3 and the old egcs-1.0.2, I've seen that they differ in the trailing backslashes in the assembler function definitions.

I found some time to look at tc-arc.c.  This should fix the problem.

gas/ChangeLog
	* config/tc-arc.c (md_assemble): Use is_end_of_line instead of
	testing for NULs.

Index: gas/config/tc-arc.c
===================================================================
RCS file: /cvs/src/src/gas/config/tc-arc.c,v
retrieving revision 1.13
diff -u -p -r1.13 tc-arc.c
--- tc-arc.c	2001/05/16 01:08:20	1.13
+++ tc-arc.c	2001/05/22 10:47:42
@@ -631,7 +631,7 @@ md_assemble (str)
 	      /* Is there anything left to parse?
 		 We don't check for this at the top because we want to parse
 		 any trailing fake arguments in the syntax string.  */
-	      if (*str == '\0')
+	      if (is_end_of_line[(unsigned char) *str])
 		break;
 
 	      /* Parse the operand.  */
@@ -739,7 +739,7 @@ md_assemble (str)
 	  while (isspace (*str))
 	    ++str;
 
-	  if (*str != '\0')
+	  if (!is_end_of_line[(unsigned char) *str])
 	    as_bad ("junk at end of line: `%s'", str);
 
 	  /* Is there a limm value?  */

------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sourceware.cygnus.com