Re: [PATCH] ld: Use run_host_cmd_yesno in indirect.exp instead of catch exec

[Alan Modra <amodra@gmail.com>]

On Fri, Jun 30, 2023 at 02:07:57PM +0800, YunQiang Su wrote:
> Catch "exec $CC_FOR_TARGET" won't use the gas/ld that we just build,
> and in fact run_host_cmd_yesno is a better choice for it.
> 
> ld/ChangeLog:
> 
> 	* testsuite/ld-elf/indirect.exp: use run_host_cmd_yesno
> 	instead of handwrite catch exec $CC_FOR_TARGET.

OK, yes, we should be using run_host_cmd everywhere we invoke a
compiler in the ld testsuite.  One reason being that run_host_cmd
properly inserts $gcc_B_opt in cases where a user wants to test
binutils with a newly built compiler, ie. when $CC specifies -B.

I'm going to apply this one too.

	* testsuite/ld-elf/no-section-header.exp (binutils_run_test):
	Use run_host_cmd to invoke $CC.

diff --git a/ld/testsuite/ld-elf/no-section-header.exp b/ld/testsuite/ld-elf/no-section-header.exp
index c010e572a32..2d8c687dc0a 100644
--- a/ld/testsuite/ld-elf/no-section-header.exp
+++ b/ld/testsuite/ld-elf/no-section-header.exp
@@ -136,7 +136,6 @@ if { ![isnative] || ![check_shared_lib_support] } {
 
 proc binutils_run_test { prog } {
     global CC
-    global gcc_B_opt
     global READELF
     global NM
     global objcopy
@@ -354,8 +353,7 @@ proc binutils_run_test { prog } {
 	    return
 	}
 
-	send_log "$CC $gcc_B_opt -o tmpdir/pr25617-1 tmpdir/pr25617-1b.o tmpdir/pr25617-1a-now-no-sec-hdr.so\n"
-	catch "exec $CC $gcc_B_opt -o tmpdir/pr25617-1 tmpdir/pr25617-1b.o tmpdir/pr25617-1a-now-no-sec-hdr.so" got
+	set got [run_host_cmd $CC "-o tmpdir/pr25617-1 tmpdir/pr25617-1b.o tmpdir/pr25617-1a-now-no-sec-hdr.so"]
 	if ![string match "*pr25617-1a-now-no-sec-hdr.so*file in wrong format*" $got] then {
 	    send_log "$got\n"
 	    fail "Update pr25617-1a-now-no-sec-hdr.so ($prog_name)"


-- 
Alan Modra
Australia Development Lab, IBM