[applies mips testsuite patch] init mips arch array.

[applies mips testsuite patch] init mips arch array.

[cgd]

Applied as obvious.  Ran a test on compilers which included multilibs
("multiple variations" if you will 8-), and happened to also check
gas...  and noticed that the test ERRORed out!  Made me feel lame!

I've also applied it to the branch.

2003-05-01  Chris Demetriou  <cgd@broadcom.com>

	* gas/mips/mips.exp (mips_arch_init): New function.
	Call it before setting up architecture array entries.

Index: gas/mips/mips.exp
===================================================================
RCS file: /cvs/src/src/gas/testsuite/gas/mips/mips.exp,v
retrieving revision 1.67
diff -u -p -r1.67 mips.exp
--- gas/mips/mips.exp	11 Apr 2003 01:56:51 -0000	1.67
+++ gas/mips/mips.exp	2 May 2003 05:11:25 -0000
@@ -78,6 +78,15 @@
 # based on the target of the assembler being built.  If always has
 # empty "as_flags" and "objdump_flags."
 
+# mips_arch_init
+#
+# This function initializes the architecture data array ("mips_arches")
+# to be empty.
+proc mips_arch_init {} {
+    global mips_arches
+    array unset mips_arches
+}
+
 # mips_arch_create ARCH GPRSIZE EXTENDS PROPS AS_FLAGS OBJDUMP_FLAGS \
 #		   (optional:) DEFAULT_FOR_TARGETS
 #
@@ -334,6 +343,7 @@ proc run_list_test_arches { name opts ar
 # Note that several targets pick default CPU based on ABI.  We
 # can't easily handle that; do NOT list those targets as defaulting
 # to any architecture.
+mips_arch_init
 mips_arch_create mips1 	32	{}	{} \
 			{ -march=mips1 -mtune=mips1 } { -mmips:3000 }
 mips_arch_create mips2 	32	mips1	{ gpr_ilocks } \

Re: [applies mips testsuite patch] init mips arch array.

[cgd]

At Fri, 2 May 2003 05:32:49 +0000 (UTC), cgd@broadcom.com wrote:
> Applied as obvious.  Ran a test on compilers which included multilibs
> ("multiple variations" if you will 8-), and happened to also check
> gas...  and noticed that the test ERRORed out!  Made me feel lame!
> 
> I've also applied it to the branch.

This worked with the current 'src' tcl/expect, but upon further
testing with a random old version that i happened to have in my path
when doing a non-unified build, i found that it wasn't OK with old
versions of tcl.

looking at the tcl manuals, looks like 'array unset' is new.  whee!

the following has been verified on old and new, and applied.


cgd
--
[ gas/testsuite/ChangeLog ]
2003-05-02  Chris Demetriou  <cgd@broadcom.com>

	* gas/mips/mips.exp (mips_arch_init): Reimplement in
	a way that's safe with ancient versions of tcl/expect.

Index: gas/mips/mips.exp
===================================================================
RCS file: /cvs/src/src/gas/testsuite/gas/mips/mips.exp,v
retrieving revision 1.68
diff -u -p -r1.68 mips.exp
--- gas/mips/mips.exp	2 May 2003 05:29:13 -0000	1.68
+++ gas/mips/mips.exp	3 May 2003 06:49:30 -0000
@@ -84,7 +84,9 @@
 # to be empty.
 proc mips_arch_init {} {
     global mips_arches
-    array unset mips_arches
+
+    # Catch becuase the variable won't be set the first time through.
+    catch {unset mips_arches}
 }
 
 # mips_arch_create ARCH GPRSIZE EXTENDS PROPS AS_FLAGS OBJDUMP_FLAGS \