What happended to sepdebug.exp?

What happended to sepdebug.exp?

[Mark Kettenis]

On OpenBSD I'm seeing the following testsuite regression/freakout:

 PASS: gdb.base/sepdebug.exp: next over recursive call
 PASS: gdb.base/sepdebug.exp: backtrace from factorial(5.1)
 PASS: gdb.base/sepdebug.exp: continue until exit at recursive next test
-PASS: gdb.base/sepdebug.exp: set separate debug location
-PASS: gdb.base/sepdebug.exp: breakpoint function, optimized file
-PASS: gdb.base/sepdebug.exp: breakpoint small function, optimized file
-PASS: gdb.base/sepdebug.exp: run until function breakpoint, optimized file
-PASS: gdb.base/sepdebug.exp: run until breakpoint set at small function, optimized file
+PASS: debuglink: set separate debug location
+PASS: debuglink: breakpoint function, optimized file
+PASS: debuglink: breakpoint small function, optimized file
+PASS: debuglink: run until function breakpoint, optimized file
+PASS: debuglink: run until breakpoint set at small function, optimized file
+ERROR: tcl error sourcing ../../../../src/gdb/gdb/testsuite/gdb.base/sepdebug.exp.
+ERROR: objcopy: there are no sections to be copied!
+    while executing
+"exec objcopy -j .note.gnu.build-id -O binary $exec $tmp"
+    (procedure "build_id_debug_filename_get" line 3)
+    invoked from within
+"build_id_debug_filename_get $binfile"
+    invoked from within
+"set build_id_debug_filename [build_id_debug_filename_get $binfile]"
+    (file "../../../../src/gdb/gdb/testsuite/gdb.base/sepdebug.exp" line 956)
+    invoked from within
+"source ../../../../src/gdb/gdb/testsuite/gdb.base/sepdebug.exp"
+"exec objcopy -j .note.gnu.build-id -O binary $exec $tmp"
+    (procedure "build_id_debug_filename_get" line 3)
+    invoked from within
+"build_id_debug_filename_get $binfile"
+    invoked from within
+"set build_id_debug_filename [build_id_debug_filename_get $binfile]"
+    (file "../../../../src/gdb/gdb/testsuite/gdb.base/sepdebug.exp" line 956)
+    invoked from within
+"source ../../../../src/gdb/gdb/testsuite/gdb.base/sepdebug.exp"
+    ("uplevel" body line 1)
+    invoked from within
+"uplevel #0 source ../../../../src/gdb/gdb/testsuite/gdb.base/sepdebug.exp"
+    invoked from within
+"catch "uplevel #0 source $test_file_name""

Now probably the OpenBSD toolchain doesn't support the build ID stuff
completely, but.

1. The "PASS: debuglink: XXXX" look wrong to me, do people see those
   on other systems too?

2. This test should fail a bit more graceful if the necessary
   toolchain support is missing.

The diff below addresses the second issue.  Does it look reasonable?

Mark


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

	* lib/gdb.exp (build_id_debug_filename_get): Improve check for
	build-id.

Index: lib/gdb.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/lib/gdb.exp,v
retrieving revision 1.87
diff -u -p -r1.87 gdb.exp
--- lib/gdb.exp 1 Sep 2007 08:16:16 -0000 1.87
+++ lib/gdb.exp 9 Sep 2007 21:46:57 -0000
@@ -382,7 +382,7 @@ proc runto { function args } {
     
     # the "at foo.c:36" output we get with -g.
     # the "in func" output we get without -g.
-    gdb_expect 30 {
+    gdb_expect 60 {
 	-re "Break.* at .*:$decimal.*$gdb_prompt $" {
 	    return 1
 	}
@@ -2487,7 +2487,14 @@ proc separate_debug_filename { exec } {
 # Return "" if no build-id found.
 proc build_id_debug_filename_get { exec } {
     set tmp "${exec}-tmp"
-    exec objcopy -j .note.gnu.build-id -O binary $exec $tmp
+    set objcopy_program [transform objcopy]
+
+    set result [catch "exec $objcopy_program -j .note.gnu.build-id -O binary $exec $tmp" output]
+    verbose "result is $result"
+    verbose "output is $output"
+    if {$result == 1} {
+	return ""
+    }
     set fi [open $tmp]
     # Skip the NOTE header.
     read $fi 16

Re: What happended to sepdebug.exp?

[Daniel Jacobowitz]

On Sun, Sep 09, 2007 at 11:49:04PM +0200, Mark Kettenis wrote:
> On OpenBSD I'm seeing the following testsuite regression/freakout:

I noticed something similar today while testing ia64-linux.

> 1. The "PASS: debuglink: XXXX" look wrong to me, do people see those
>    on other systems too?

I've been meaning to fix this.  See the pf_prefix setting.  It should
go in each individual test name, instead, like we do in other testcases.

> 2. This test should fail a bit more graceful if the necessary
>    toolchain support is missing.
> 
> The diff below addresses the second issue.  Does it look reasonable?

It does to me!

-- 
Daniel Jacobowitz
CodeSourcery

Re: What happended to sepdebug.exp?

[Mark Kettenis]

> Date: Sun, 9 Sep 2007 17:55:27 -0400
> From: Daniel Jacobowitz <drow@false.org>
> 
> On Sun, Sep 09, 2007 at 11:49:04PM +0200, Mark Kettenis wrote:
> > On OpenBSD I'm seeing the following testsuite regression/freakout:
> 
> I noticed something similar today while testing ia64-linux.
> 
> > 1. The "PASS: debuglink: XXXX" look wrong to me, do people see those
> >    on other systems too?
> 
> I've been meaning to fix this.

Please do ;).

> > 2. This test should fail a bit more graceful if the necessary
> >    toolchain support is missing.
> > 
> > The diff below addresses the second issue.  Does it look reasonable?
> 
> It does to me!

Committed!