Testsuite failures

Testsuite failures

[Mark Kettenis]

Hi Nick,

I'm seeing some new testsuite failures on several OpenBSD machines:

Running ../../../../src/gdb/gdb/testsuite/gdb.base/annota3.exp ...
FAIL: gdb.base/annota3.exp: run until main breakpoint (pattern 6)
FAIL: gdb.base/annota3.exp: send SIGUSR1 (pattern 5)
FAIL: gdb.base/annota3.exp: annotate ignore count change (pattern 3)

I think these are caused by your last commit:

2006-07-13  Nick Roberts  <nickrob@snap.net.nz>

	* annotate.c (annotate_frame_begin): Re-instate frame-begin
	annotation for level 3 annotations.

Here's a patch that adjusts the testsuite to deal with the re-instated
frame-begin annotation.  Can you verify that this is indeed the
intended effect of your change?


Index: ChangeLog
from  Mark Kettenis  <kettenis@gnu.org>

	* gdb.base/annota3.exp: Adjust patterns for re-instated
	frame-begin annotation.

Index: gdb.base/annota3.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.base/annota3.exp,v
retrieving revision 1.7
diff -u -p -r1.7 annota3.exp
--- gdb.base/annota3.exp 7 Mar 2006 15:23:32 -0000 1.7
+++ gdb.base/annota3.exp 16 Jul 2006 11:41:02 -0000
@@ -112,6 +112,7 @@ gdb_expect_list "breakpoint info" "$gdb_
 
 #
 # run to a break point will test:
+# annotate-frame-begin
 #
 #exp_internal 1
 send_gdb "run\n"
@@ -121,7 +122,9 @@ gdb_expect_list "run until main breakpoi
     "\r\n\032\032starting\r\n"
     "\r\n\032\032breakpoint 1\r\n"
     "\r\n"
-    "Breakpoint 1, main \\(\\) at .*annota3.c:32\r\n"
+    "Breakpoint 1, "
+    "\r\n\032\032frame-begin 0 0x\[0-9a-z\]+\r\n"
+    "main \\(\\) at .*annota3.c:32\r\n"
     "\r\n\032\032source.*annota3.c:32:.*:beg:0x\[0-9a-z\]+\r\n"
     "\r\n\032\032stopped\r\n"
 }
@@ -222,7 +225,10 @@ if [target_info exists gdb,nosignals] {
 	"Continuing with signal SIGUSR1.\r\n"
 	"\r\n\032\032starting\r\n"
 	"\r\n\032\032breakpoint 2\r\n"
-	"Breakpoint 2, handle_USR1 \\(sig=\[0-9\]+\\) at .*annota3.c:\[0-9\]+\r\n"
+	"\r\n"
+	"Breakpoint 2, "
+	"\r\n\032\032frame-begin 0 0x\[0-9a-z\]+\r\n"
+	"handle_USR1 \\(sig=\[0-9\]+\\) at .*annota3.c:\[0-9\]+\r\n"
 	"\r\n\032\032source .*annota3.c:\[0-9\]+:\[0-9\]+:beg:0x\[0-9a-z\]+\r\n"
 	"\r\n\032\032stopped\r\n"
     }
@@ -324,7 +330,10 @@ send_gdb "continue\n"
 gdb_expect_list "annotate ignore count change" "$gdb_prompt$" {
     "\r\n\032\032post-prompt\r\n"
     "\r\n\032\032breakpoint 5\r\n"
-    "Breakpoint 5, main \\(\\) at .*annota3.c:46\r\n"
+    "\r\n"
+    "Breakpoint 5, "
+    "\r\n\032\032frame-begin 0 0x\[0-9a-z\]+\r\n"
+    "main \\(\\) at .*annota3.c:46\r\n"
     "\r\n\032\032source .*annota3.c:46:\[0-9\]+:beg:0x\[0-9a-z\]+\r\n"
     "1: value = 11\r\n"
     "\r\n\032\032stopped\r\n"

Re: Testsuite failures

[Nick Roberts]

 > I'm seeing some new testsuite failures on several OpenBSD machines:
 > 
 > Running ../../../../src/gdb/gdb/testsuite/gdb.base/annota3.exp ...
 > FAIL: gdb.base/annota3.exp: run until main breakpoint (pattern 6)
 > FAIL: gdb.base/annota3.exp: send SIGUSR1 (pattern 5)
 > FAIL: gdb.base/annota3.exp: annotate ignore count change (pattern 3)
 > 
 > I think these are caused by your last commit:
 > 
 > 2006-07-13  Nick Roberts  <nickrob@snap.net.nz>
 > 
 > 	* annotate.c (annotate_frame_begin): Re-instate frame-begin
 > 	annotation for level 3 annotations.

Yes.  I forgot that there were tests for level 3 annotations.  Sorry about
that.

 > Here's a patch that adjusts the testsuite to deal with the re-instated
 > frame-begin annotation.  Can you verify that this is indeed the
 > intended effect of your change?

It looks good to me.  Also annota3.exp in gdb.cp (patch attached).  It might
be easier to use the $hex variable but I couldn't get it to work.

 > Index: ChangeLog
 > from  Mark Kettenis  <kettenis@gnu.org>
 > 
 > 	* gdb.base/annota3.exp: Adjust patterns for re-instated
 > 	frame-begin annotation.
 > 
 > Index: gdb.base/annota3.exp
 > ===================================================================
 > RCS file: /cvs/src/src/gdb/testsuite/gdb.base/annota3.exp,v
 > retrieving revision 1.7
 > diff -u -p -r1.7 annota3.exp
 > --- gdb.base/annota3.exp 7 Mar 2006 15:23:32 -0000 1.7
 > +++ gdb.base/annota3.exp 16 Jul 2006 11:41:02 -0000
 > @@ -112,6 +112,7 @@ gdb_expect_list "breakpoint info" "$gdb_
 >  
 >  #
 >  # run to a break point will test:
 > +# annotate-frame-begin

I don't think this comment should be added (it means nothing to me).

 >  #exp_internal 1
 >...

-- 
Nick                                           http://www.inet.net.nz/~nickrob



*** annota3.exp	12 Mar 2006 19:42:40 +1300	1.5
--- annota3.exp	17 Jul 2006 00:20:54 +1200	
*************** gdb_expect_list "first run until main br
*** 96,102 ****
      "\r\n\032\032starting\r\n"
      "\r\n\032\032breakpoint 1\r\n"
      "\r\n"
!     "Breakpoint 1, main \\(\\) at .*annota3.cc:25\r\n"
      "\r\n\032\032source.*annota3.cc:25:.*:beg:0x\[0-9a-z\]+\r\n"
      "\r\n\032\032stopped\r\n"
  }
--- 96,104 ----
      "\r\n\032\032starting\r\n"
      "\r\n\032\032breakpoint 1\r\n"
      "\r\n"
!     "Breakpoint 1, "
!     "\r\n\032\032frame-begin 0 0x\[0-9a-z\]+\r\n"
!     "main \\(\\) at .*annota3.cc:25\r\n"
      "\r\n\032\032source.*annota3.cc:25:.*:beg:0x\[0-9a-z\]+\r\n"
      "\r\n\032\032stopped\r\n"
  }
*************** gdb_expect_list "second run until main b
*** 164,170 ****
      "\r\n\032\032starting\r\n"
      "\r\n\032\032breakpoint 2\r\n"
      "\r\n"
!     "Breakpoint 2, main \\(\\) at .*annota3.cc:22\r\n"
      "\r\n\032\032source.*annota3.cc:22:.*:beg:0x\[0-9a-z\]+\r\n"
      "\r\n\032\032stopped\r\n"
  }
--- 166,174 ----
      "\r\n\032\032starting\r\n"
      "\r\n\032\032breakpoint 2\r\n"
      "\r\n"
!     "Breakpoint 2, "
!     "\r\n\032\032frame-begin 0 0x\[0-9a-z\]+\r\n"
!     "main \\(\\) at .*annota3.cc:22\r\n"
      "\r\n\032\032source.*annota3.cc:22:.*:beg:0x\[0-9a-z\]+\r\n"
      "\r\n\032\032stopped\r\n"
  }

Re: Testsuite failures

[Daniel Jacobowitz]

On Mon, Jul 17, 2006 at 12:37:05AM +1200, Nick Roberts wrote:
> It looks good to me.  Also annota3.exp in gdb.cp (patch attached).  It might
> be easier to use the $hex variable but I couldn't get it to work.

You probably need a "global hex" somewhere?

-- 
Daniel Jacobowitz
CodeSourcery

Re: Testsuite failures

[Nick Roberts]

 > > It looks good to me.  Also annota3.exp in gdb.cp (patch attached).  It
 > > might be easier to use the $hex variable but I couldn't get it to work.
 > 
 > You probably need a "global hex" somewhere?

Yes, but where?  It didn't seem to work at top level in annota3.exp.  The
other files that use $hex have "global hex" but their use is always within a
procedure.  I can't see where hex is defined (I had thought it would be in
src/gdb/testsuite/lib/gdb.exp).


-- 
Nick                                           http://www.inet.net.nz/~nickrob

Re: Testsuite failures

[Daniel Jacobowitz]

On Mon, Jul 17, 2006 at 12:25:36PM +1200, Nick Roberts wrote:
>  > > It looks good to me.  Also annota3.exp in gdb.cp (patch attached).  It
>  > > might be easier to use the $hex variable but I couldn't get it to work.
>  > 
>  > You probably need a "global hex" somewhere?
> 
> Yes, but where?  It didn't seem to work at top level in annota3.exp.  The
> other files that use $hex have "global hex" but their use is always within a
> procedure.  I can't see where hex is defined (I had thought it would be in
> src/gdb/testsuite/lib/gdb.exp).

If you're not in a procedure you're normally at global scope.  Must
have been some other problem, maybe quoting.  The file already uses
$hex elsewhere.

-- 
Daniel Jacobowitz
CodeSourcery

Re: Testsuite failures

[Nick Roberts]

 > If you're not in a procedure you're normally at global scope.  Must
 > have been some other problem, maybe quoting.  The file already uses
 > $hex elsewhere.

I see the problem now, gdb_expect_list needs "-re":


*** annota3.exp	12 Mar 2006 19:42:40 +1300	1.5
--- annota3.exp	17 Jul 2006 12:32:54 +1200	
*************** gdb_expect_list "annotation set at level
*** 91,103 ****
  
  send_gdb "run\n"
  gdb_expect_list "first run until main breakpoint" "$gdb_prompt$" {
!     "\r\n\032\032post-prompt\r\n"
      "Starting program: .*annota3 \r\n"
      "\r\n\032\032starting\r\n"
      "\r\n\032\032breakpoint 1\r\n"
      "\r\n"
!     "Breakpoint 1, main \\(\\) at .*annota3.cc:25\r\n"
!     "\r\n\032\032source.*annota3.cc:25:.*:beg:0x\[0-9a-z\]+\r\n"
      "\r\n\032\032stopped\r\n"
  }
  
--- 91,105 ----
  
  send_gdb "run\n"
  gdb_expect_list "first run until main breakpoint" "$gdb_prompt$" {
!     -re "\r\n\032\032post-prompt\r\n"
      "Starting program: .*annota3 \r\n"
      "\r\n\032\032starting\r\n"
      "\r\n\032\032breakpoint 1\r\n"
      "\r\n"
!     "Breakpoint 1, "
!     "\r\n\032\032frame-begin 0 $hex\r\n"
!     "main \\(\\) at .*annota3.cc:25\r\n"
!     "\r\n\032\032source.*annota3.cc:25:.*:beg:$hex\r\n"
      "\r\n\032\032stopped\r\n"
  }



-- 
Nick                                           http://www.inet.net.nz/~nickrob

Re: Testsuite failures

[Daniel Jacobowitz]

On Mon, Jul 17, 2006 at 12:36:06PM +1200, Nick Roberts wrote:
>  > If you're not in a procedure you're normally at global scope.  Must
>  > have been some other problem, maybe quoting.  The file already uses
>  > $hex elsewhere.
> 
> I see the problem now, gdb_expect_list needs "-re":

I'm pretty sure it doesn't.  It always treats the list items as regular
expressions.

The problem is elsewhere:

>   gdb_expect_list "first run until main breakpoint" "$gdb_prompt$" {

You're in {} quoting, so variables are not expanded.  It is probably
not worth changing that.

-- 
Daniel Jacobowitz
CodeSourcery

Re: Testsuite failures

[Nick Roberts]

 > > I see the problem now, gdb_expect_list needs "-re":
 > 
 > I'm pretty sure it doesn't.  It always treats the list items as regular
 > expressions.

You're right.  I got distracted by the annotate-quit fail (which was there
before my change).
 
 > The problem is elsewhere:
 > 
 > >   gdb_expect_list "first run until main breakpoint" "$gdb_prompt$" {
 > 
 > You're in {} quoting, so variables are not expanded.  It is probably
 > not worth changing that.

OK.  Shall I commit Mark's and my changes for gdb.base/annota3.exp and
gdb.cp/annota3.exp (without hex)?


-- 
Nick                                           http://www.inet.net.nz/~nickrob

Re: Testsuite failures

[Daniel Jacobowitz]

On Mon, Jul 17, 2006 at 12:53:45PM +1200, Nick Roberts wrote:
>  > The problem is elsewhere:
>  > 
>  > >   gdb_expect_list "first run until main breakpoint" "$gdb_prompt$" {
>  > 
>  > You're in {} quoting, so variables are not expanded.  It is probably
>  > not worth changing that.
> 
> OK.  Shall I commit Mark's and my changes for gdb.base/annota3.exp and
> gdb.cp/annota3.exp (without hex)?

Up to Mark, since he started the thread.  They look sane to me.

-- 
Daniel Jacobowitz
CodeSourcery

Re: Testsuite failures

[Mark Kettenis]

> From: Nick Roberts <nickrob@snap.net.nz>
> Date: Mon, 17 Jul 2006 12:53:45 +1200
> 
>  > > I see the problem now, gdb_expect_list needs "-re":
>  > 
>  > I'm pretty sure it doesn't.  It always treats the list items as regular
>  > expressions.
> 
> You're right.  I got distracted by the annotate-quit fail (which was there
> before my change).
>  
>  > The problem is elsewhere:
>  > 
>  > >   gdb_expect_list "first run until main breakpoint" "$gdb_prompt$" {
>  > 
>  > You're in {} quoting, so variables are not expanded.  It is probably
>  > not worth changing that.
> 
> OK.  Shall I commit Mark's and my changes for gdb.base/annota3.exp and
> gdb.cp/annota3.exp (without hex)?

Go ahead.

Mark