[obv] testsuite: gdb.base/foll-exec.exp: Fix racy FAIL
Jan Kratochvil
jan.kratochvil@redhat.com
Thu Dec 1 22:10:00 GMT 2011
Hi,
I was getting:
(gdb) PASS: gdb.base/foll-exec.exp: continue to first exec catchpoint
kill
Kill the program being debugged? (y or n) y
+file gdb/testsuite.unix.-m32/gdb.base/(gdb) FAIL: gdb.base/foll-exec.exp: loading symbols (timeout)
+ERROR: info breakpoints (timeout)
read1/writew did not help in this case. strace of expect revealed:
read(6, "Kill the program being debugged? (y or n) ", 4096) = 42
write(6, "y\n", 2) = 2
write(6, "file /home/jkratoch/redhat/gdb-clean/gdb/testsuite/gdb.base/foll-exec\n", 70) = 70
read(6, "y\r\nfile /home/jkratoch/redhat/gdb-clean/gdb/testsuite(gdb) ", 4096) = 59
select(7, [6], [], [6], {9, 999999}) = 0 (Timeout)
write(4, "FAIL: gdb.base/foll-exec.exp: loading symbols (timeout)\n", 56) = 56
Therefore checked in as obvious.
The problem was 90% reproducible for me and now 0% reproducible for 30 passes.
Tested on x86_64-fedora16-linux-gnu.
Thanks,
Jan
http://sourceware.org/ml/gdb-cvs/2011-12/msg00004.html
--- src/gdb/testsuite/ChangeLog 2011/11/30 08:25:15 1.2952
+++ src/gdb/testsuite/ChangeLog 2011/12/01 22:06:45 1.2953
@@ -1,3 +1,9 @@
+2011-12-01 Jan Kratochvil <jan.kratochvil@redhat.com>
+
+ Fix racy FAILs.
+ * gdb.base/foll-exec.exp (zap_session): Use gdb_test_no_output instead
+ of send_gdb.
+
2011-11-30 Ulrich Weigand <uweigand@de.ibm.com>
* gdb.base/testenv.exp: Skip on remote targets. Unset global
--- src/gdb/testsuite/gdb.base/foll-exec.exp 2011/01/11 19:16:23 1.15
+++ src/gdb/testsuite/gdb.base/foll-exec.exp 2011/12/01 22:06:45 1.16
@@ -51,7 +51,7 @@
send_gdb "kill\n"
gdb_expect {
-re ".*Kill the program being debugged.*y or n. $" {
- send_gdb "y\n"
+ gdb_test_no_output "y" ""
send_gdb "file $binfile\n"
gdb_expect {
-re ".*Load new symbol table from.*y or n. $" {
More information about the Gdb-patches
mailing list