[PATCH] gdb.mi/mi-var-display.exp: fix racy tests (PR testsuite/12649)

[PATCH] gdb.mi/mi-var-display.exp: fix racy tests (PR testsuite/12649)

[Marek Polacek]

Another one where we need to replace gdb_test_multiple with mi_gdb_test
to prevent failures with read1().  Note that when setting `fp', we need
to s/1/3/ since the index 1 and 2 is used by mi_gdb_test.  Tested with 
both read{,1}.  OK to apply?

2011-06-21  Marek Polacek  <mpolacek@redhat.com>

        * gdb.mi/mi-var-display.exp: Replace gdb_test_multiple with mi_gdb_test.

--- gdb/gdb/testsuite/gdb.mi/mi-var-display.exp.mp      2011-06-21 13:05:15.679575309 +0200
+++ gdb/gdb/testsuite/gdb.mi/mi-var-display.exp 2011-06-21 13:34:06.561961094 +0200
@@ -605,12 +605,8 @@ mi_gdb_test "-var-list-children anone" \
 
 
 # Record fp
-
-gdb_test_multiple "p/x \$fp" "print FP register" {
-    -re ".*($hex).*\\^done\r\n$mi_gdb_prompt$" { 
-       pass "print FP register"
-       set fp $expect_out(1,string) 
-    }
+if ![mi_gdb_test "p/x \$fp" ".*($hex).*\\^done" "Xprint FP register"] {
+       set fp $expect_out(3,string) 
 }
 
 mi_continue_to "incr_a"

Re: [PATCH] gdb.mi/mi-var-display.exp: fix racy tests (PR testsuite/12649)

[Marek Polacek]

On 06/21/2011 03:56 PM, Marek Polacek wrote:
> +if ![mi_gdb_test "p/x \$fp" ".*($hex).*\\^done" "Xprint FP register"] {

Duh, redundant `X'.  Fixed.

Re: [PATCH] gdb.mi/mi-var-display.exp: fix racy tests (PR testsuite/12649)

[Tom Tromey]

>>>>> "Marek" == Marek Polacek <mpolacek@redhat.com> writes:

Marek> 2011-06-21  Marek Polacek  <mpolacek@redhat.com>
Marek>         * gdb.mi/mi-var-display.exp: Replace gdb_test_multiple with mi_gdb_test.

One nit.

Marek> +if ![mi_gdb_test "p/x \$fp" ".*($hex).*\\^done" "Xprint FP register"] {
Marek> +       set fp $expect_out(3,string) 

Too much indentation here.

Ok with that fix; plus the fix from your follow-up email.

Tom

Re: [PATCH] gdb.mi/mi-var-display.exp: fix racy tests (PR testsuite/12649)

[Marek Polacek]

On 06/22/2011 06:06 PM, Tom Tromey wrote:
> Ok with that fix; plus the fix from your follow-up email.

Thanks, checked in:
http://sourceware.org/ml/gdb-cvs/2011-06/msg00121.html

	Marek