mi-break.exp intermitent failure

mi-break.exp intermitent failure

[Pedro Alves]

[-- Attachment #1: Type: text/plain, Size: 1810 bytes --]

Hi,

I'm getting these intermitent FAILures in mi-break.exp (and a couple
more in chain reaction) in sync mode, due to previous pattern
considering the gdb prompt optional, hence sometimes consuming the
*stopped line, and leaving the (gdb) unconsumed.

Running ../../../src/gdb/testsuite/gdb.mi/mi-break.exp ...
FAIL: gdb.mi/mi-break.exp: breakpoint at nonexistent function
FAIL: gdb.mi/mi-break.exp: create varobj for function call
FAIL: gdb.mi/mi-break.exp: update varobj for function call


220-exec-run 
220^running
(gdb) 
=thread-created,id="1"
*running,thread-id="1"
Hello, World!callme
*stopped,reason="breakpoint-hit",disp="keep",bkptno="5",thread-id="1",frame={addr="0x08048441",func="callme",args=[{name="i",value="2"}],file="../../../src/gdb/testsuite/gdb.mi/basics.c",fullname="/home/pedro/gdb/nonstop_head/src/gdb/testsuite/gdb.mi/basics.c",line="51"}
PASS: gdb.mi/mi-break.exp: run to breakpoint with ignore count
-break-insert function_that_does_not_exist
(gdb) 
FAIL: gdb.mi/mi-break.exp: breakpoint at nonexistent function
-var-create V * return_1()
^error,msg="Function \"function_that_does_not_exist\" not defined."

vs

220-exec-run 
220^running
(gdb) 
=thread-created,id="1"
*running,thread-id="1"
Hello, World!callme
*stopped,reason="breakpoint-hit",disp="keep",bkptno="5",thread-id="1",frame={addr="0x08048441",func="callme",args=[{name="i",value="2"}],file="../../../src/gdb/testsuite/gdb.mi/basics.c",fullname="/home/pedro/gdb/nonstop_head/src/gdb/testsuite/gdb.mi/basics.c",line="51"}
(gdb) 
PASS: gdb.mi/mi-break.exp: run to breakpoint with ignore count
-break-insert function_that_does_not_exist
^error,msg="Function \"function_that_does_not_exist\" not defined."
(gdb) 
PASS: gdb.mi/mi-break.exp: breakpoint at nonexistent function


The attached fixes it for me.

-- 
Pedro Alves

[-- Attachment #2: mi-break.exp.diff --]
[-- Type: text/x-diff, Size: 1262 bytes --]

2008-06-11  Pedro Alves  <pedro@codesourcery.com>

	* gdb.mi/mi-break.exp (test_ignore_count): Condition expecting the
	prompt based on async mode.

---
 gdb/testsuite/gdb.mi/mi-break.exp |    9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

Index: src/gdb/testsuite/gdb.mi/mi-break.exp
===================================================================
--- src.orig/gdb/testsuite/gdb.mi/mi-break.exp	2008-06-11 12:27:48.000000000 +0100
+++ src/gdb/testsuite/gdb.mi/mi-break.exp	2008-06-11 12:55:37.000000000 +0100
@@ -151,6 +151,13 @@ proc test_rbreak_creation_and_listing {}
 
 proc test_ignore_count {} {
     global mi_gdb_prompt
+    global async
+
+    if {$async} {
+        set prompt_re ""
+    } else {
+        set prompt_re "$mi_gdb_prompt"
+    }
 
     mi_gdb_test "-break-insert -i 1 callme" \
         "\\^done.*ignore=\"1\".*" \
@@ -159,7 +166,7 @@ proc test_ignore_count {} {
     mi_run_cmd
 
     gdb_expect {
-        -re ".*\\*stopped.*func=\"callme\".*args=\\\[\{name=\"i\",value=\"2\"\}\\\].*\r\n($mi_gdb_prompt)?$" {
+        -re ".*\\*stopped.*func=\"callme\".*args=\\\[\{name=\"i\",value=\"2\"\}\\\].*\r\n$prompt_re$" {
             pass "run to breakpoint with ignore count"
         }
         -re ".*$mi_gdb_prompt$" {

Re: mi-break.exp intermitent failure

[Vladimir Prus]

[-- Attachment #1: Type: text/plain, Size: 961 bytes --]

Pedro Alves wrote:

> Hi,
> 
> I'm getting these intermitent FAILures in mi-break.exp (and a couple
> more in chain reaction) in sync mode, due to previous pattern
> considering the gdb prompt optional, hence sometimes consuming the
> *stopped line, and leaving the (gdb) unconsumed.
> 
> Running ../../../src/gdb/testsuite/gdb.mi/mi-break.exp ...
> FAIL: gdb.mi/mi-break.exp: breakpoint at nonexistent function
> FAIL: gdb.mi/mi-break.exp: create varobj for function call
> FAIL: gdb.mi/mi-break.exp: update varobj for function call

Thanks for reporting; somehow I never run into this.

As I've told on IRC, the test should be using mi_expect_stop; I've no idea
why I did not covert it when I've introduced mi_expect_stop but I've
checked in the following patch. Can you check that no unexpected
failures happen for you now.

Incidentally, mi-async.exp is the only file that has direct checks for *stopped,
but those will die very soon now.

Thanks,
Volodya


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: commit.diff --]
[-- Type: text/x-diff; name="commit.diff", Size: 1674 bytes --]

Index: gdb/testsuite/ChangeLog
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/ChangeLog,v
retrieving revision 1.1661
diff -u -p -r1.1661 ChangeLog
--- gdb/testsuite/ChangeLog	10 Jun 2008 10:23:54 -0000	1.1661
+++ gdb/testsuite/ChangeLog	11 Jun 2008 13:34:48 -0000
@@ -1,5 +1,9 @@
 2008-06-10  Vladimir Prus  <vladimir@codesourcery.com>
 
+	* gdb.mi/mi-break.exp (test_ignore_count): Use mi_expect_stop.
+
+2008-06-10  Vladimir Prus  <vladimir@codesourcery.com>
+
         * gdb.mi/mi-console.exp: Adjust.
         * gdb.mi/mi-syn-frame.exp: Adjust.
         * gdb.mi/mi2-console.exp: Adjust.
Index: gdb/testsuite/gdb.mi/mi-break.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.mi/mi-break.exp,v
retrieving revision 1.18
diff -u -p -r1.18 mi-break.exp
--- gdb/testsuite/gdb.mi/mi-break.exp	10 Jun 2008 09:35:08 -0000	1.18
+++ gdb/testsuite/gdb.mi/mi-break.exp	11 Jun 2008 13:34:48 -0000
@@ -158,17 +158,8 @@ proc test_ignore_count {} {
 
     mi_run_cmd
 
-    gdb_expect {
-        -re ".*\\*stopped.*func=\"callme\".*args=\\\[\{name=\"i\",value=\"2\"\}\\\].*\r\n($mi_gdb_prompt)?$" {
-            pass "run to breakpoint with ignore count"
-        }
-        -re ".*$mi_gdb_prompt$" {
-            fail "run to breakpoint with ignore count"
-        }
-        timeout {
-            fail "run to breakpoint with ignore count (timeout)"
-        }
-    }  
+    mi_expect_stop "breakpoint-hit" "callme" "\{name=\"i\",value=\"2\"\}" ".*basics.c" "51" \
+        {"" "disp=\"keep\"" } "run to breakpoint with ignore count"
 }
 
 proc test_error {} {

Re: mi-break.exp intermitent failure

[Pedro Alves]

A Wednesday 11 June 2008 14:38:56, Vladimir Prus wrote:

> As I've told on IRC, the test should be using mi_expect_stop; I've no idea
> why I did not covert it when I've introduced mi_expect_stop but I've
> checked in the following patch. Can you check that no unexpected
> failures happen for you now.

All perfect.  Thanks!

-- 
Pedro Alves