[PATCH] Fix gdb.fortran "failed to extract expected results" errors

[PATCH] Fix gdb.fortran "failed to extract expected results" errors

[Keith Seitz]

When running the gdb.fortran tests array-slices.exp and lbound-ubound.exp,
the test suite throws several ERRORs on native-gdbserver/-m{32,64},
and native-extended-gdbsever/-m{32,64}:

[on native-extended-gdbserver/-m64]
Running /home/keiths/work/gdb/branches/testsuite-errors/linux/gdb/testsuite/../../../src/gdb/testsuite/gdb.fortran/array-slices.exp ...
ERROR: failed to extract expected results
ERROR: failed to extract expected results
Running /home/keiths/work/gdb/branches/testsuite-errors/linux/gdb/testsuite/../../../src/gdb/testsuite/gdb.fortran/lbound-ubound.exp ...
ERROR: failed to extract expected results for lbound

This occurs because the tests require inferior I/O which we do not have
access to while using these targets.

This patch skips these tests when running on non-native targets.
---
 gdb/testsuite/gdb.fortran/array-slices.exp  | 6 ++++++
 gdb/testsuite/gdb.fortran/lbound-ubound.exp | 6 ++++++
 2 files changed, 12 insertions(+)

diff --git a/gdb/testsuite/gdb.fortran/array-slices.exp b/gdb/testsuite/gdb.fortran/array-slices.exp
index 5a909812fe3..8f9c012e399 100644
--- a/gdb/testsuite/gdb.fortran/array-slices.exp
+++ b/gdb/testsuite/gdb.fortran/array-slices.exp
@@ -62,6 +62,12 @@ proc run_test { repack } {
 	return -1
     }
 
+    # This test relies on output from the inferior and is not supported on
+    # remote targets.
+    if {![gdb_is_target_native]} {
+	return 0
+    }
+
     # Avoid libc symbols, in particular the 'array' type.
     gdb_test_no_output "nosharedlibrary"
 
diff --git a/gdb/testsuite/gdb.fortran/lbound-ubound.exp b/gdb/testsuite/gdb.fortran/lbound-ubound.exp
index 51dd3818499..671b251c799 100644
--- a/gdb/testsuite/gdb.fortran/lbound-ubound.exp
+++ b/gdb/testsuite/gdb.fortran/lbound-ubound.exp
@@ -32,6 +32,12 @@ if ![fortran_runto_main] {
     return -1
 }
 
+# This test relies on output from the inferior and is not supported on
+# remote targets.
+if {![gdb_is_target_native]} {
+   return 0
+}
+
 # Avoid libc symbols, in particular the 'array' type.
 gdb_test_no_output "nosharedlibrary"
 
-- 
2.34.1

Re: [PATCH] Fix gdb.fortran "failed to extract expected results" errors

[Tom Tromey]

>>>>> "Keith" == Keith Seitz via Gdb-patches <gdb-patches@sourceware.org> writes:

Keith> When running the gdb.fortran tests array-slices.exp and lbound-ubound.exp,
Keith> the test suite throws several ERRORs on native-gdbserver/-m{32,64},
Keith> and native-extended-gdbsever/-m{32,64}:

Keith> [on native-extended-gdbserver/-m64]
Keith> Running /home/keiths/work/gdb/branches/testsuite-errors/linux/gdb/testsuite/../../../src/gdb/testsuite/gdb.fortran/array-slices.exp ...
Keith> ERROR: failed to extract expected results
Keith> ERROR: failed to extract expected results
Keith> Running /home/keiths/work/gdb/branches/testsuite-errors/linux/gdb/testsuite/../../../src/gdb/testsuite/gdb.fortran/lbound-ubound.exp ...
Keith> ERROR: failed to extract expected results for lbound

Keith> This occurs because the tests require inferior I/O which we do not have
Keith> access to while using these targets.

Keith> This patch skips these tests when running on non-native targets.

Looks good to me.

Tom

Re: [PATCH] Fix gdb.fortran "failed to extract expected results" errors

[Simon Marchi]

On 2022-02-25 11:40, Tom Tromey wrote:
>>>>>> "Keith" == Keith Seitz via Gdb-patches <gdb-patches@sourceware.org> writes:
> 
> Keith> When running the gdb.fortran tests array-slices.exp and lbound-ubound.exp,
> Keith> the test suite throws several ERRORs on native-gdbserver/-m{32,64},
> Keith> and native-extended-gdbsever/-m{32,64}:
> 
> Keith> [on native-extended-gdbserver/-m64]
> Keith> Running /home/keiths/work/gdb/branches/testsuite-errors/linux/gdb/testsuite/../../../src/gdb/testsuite/gdb.fortran/array-slices.exp ...
> Keith> ERROR: failed to extract expected results
> Keith> ERROR: failed to extract expected results
> Keith> Running /home/keiths/work/gdb/branches/testsuite-errors/linux/gdb/testsuite/../../../src/gdb/testsuite/gdb.fortran/lbound-ubound.exp ...
> Keith> ERROR: failed to extract expected results for lbound
> 
> Keith> This occurs because the tests require inferior I/O which we do not have
> Keith> access to while using these targets.
> 
> Keith> This patch skips these tests when running on non-native targets.
> 
> Looks good to me.
> 
> Tom

An alternative would be to change those tests to not rely on inferior IO.

They could perhaps save the expected outcome to a string, which GDB would
read, rather than printing it to stdout.

But, that takes some time, and most importantly, someone who knows
Fortran :).  So I also think your patch is ok.

Simon

Re: [PATCH] Fix gdb.fortran "failed to extract expected results" errors

[Keith Seitz]

On 2/25/22 08:48, Simon Marchi via Gdb-patches wrote:
> 
> 
> On 2022-02-25 11:40, Tom Tromey wrote:
>>>>>>> "Keith" == Keith Seitz via Gdb-patches <gdb-patches@sourceware.org> writes:
>>
>>
>> Keith> This patch skips these tests when running on non-native targets.
>>
>> Looks good to me.
> 
> An alternative would be to change those tests to not rely on inferior IO.
> 
> They could perhaps save the expected outcome to a string, which GDB would
> read, rather than printing it to stdout.
> 
> But, that takes some time, and most importantly, someone who knows
> Fortran :).  So I also think your patch is ok.

Oh, you got me! :-)

That's a good point, though. I will see if I can update this test, but in the
meantime, I've pushed this change. All these errors are interfering with our
Fedora CI efforts.

Thank you, Tom and Simon!

Keith

Re: [PATCH] Fix gdb.fortran "failed to extract expected results" errors

[Pedro Alves]

On 2022-02-28 15:35, Keith Seitz via Gdb-patches wrote:
> On 2/25/22 08:48, Simon Marchi via Gdb-patches wrote:
>>
>>
>> On 2022-02-25 11:40, Tom Tromey wrote:
>>>>>>>> "Keith" == Keith Seitz via Gdb-patches <gdb-patches@sourceware.org> writes:
>>>
>>>
>>> Keith> This patch skips these tests when running on non-native targets.
>>>
>>> Looks good to me.
>>
>> An alternative would be to change those tests to not rely on inferior IO.
>>
>> They could perhaps save the expected outcome to a string, which GDB would
>> read, rather than printing it to stdout.
>>
>> But, that takes some time, and most importantly, someone who knows
>> Fortran :).  So I also think your patch is ok.
> 
> Oh, you got me! :-)
> 
> That's a good point, though. I will see if I can update this test, but in the
> meantime, I've pushed this change. All these errors are interfering with our
> Fedora CI efforts.
> 
> Thank you, Tom and Simon!

Well, the change is incorrect, because when testing with gdbserver we do have
inferior output.  :-P  

We're not supposed to check whether the target is native.

Instead, we're supposed to use gdb_test_stdio & gdb_skip_stdio_test.

[PATCH] Make gdb.fortran/{array-slices,lbound-ubound} work against gdbserver

[Pedro Alves]

On 2022-03-03 20:02, Pedro Alves wrote:

> Well, the change is incorrect, because when testing with gdbserver we do have
> inferior output.  :-P  
> 
> We're not supposed to check whether the target is native.
> 
> Instead, we're supposed to use gdb_test_stdio & gdb_skip_stdio_test.

Keith mentioned he might be away this week, so I went ahead and played with this.  We
can't use gdb_test_stdio here, because that maps to gdb_test, and we want gdb_test_multiple
instead.  So we use inferior_spawn_id with gdb_test_multiple instead.

It took me some head scratching and staring at logs to figure out why the patch
wasn't working -- I was hitting a buglet in gdb.fortran/array-slices.f90, argh.
With that out of the way, both testcases now pass cleanly with gdbserver too.

Here's the proposed patch.  Any comments?

From cb3fadb145a5ac6a2ddd5714ec17374ef5254cba Mon Sep 17 00:00:00 2001
From: Pedro Alves <pedro@palves.net>
Date: Fri, 4 Mar 2022 17:34:18 +0000
Subject: [PATCH] Make gdb.fortran/{array-slices,lbound-ubound} work against
 gdbserver

gdb.fortran/array-slices.exp and gdb.fortran/lbound-ubound.exp were
recently disabled unless testing with the native target, because they
rely on inferior I/O.  However, when testing against gdbserver using
the native-gdbserver/native-extended-gdbserver boards, we do have
access to inferior I/O.

The right way to check whether the board can do I/O, is via checking
the gdb,noinferiorio board variable.  Switch to using that.

And then, tweak the testcases to expect output to appear in
inferior_spawn_id, instead of gdb_spawn_id.  When testing against the
native target, inferior_spawn_id is the same as gdb_spawn_id.  When
testing against gdbserver, it maps to gdbserver_spawn_id.

This exposed a buglet in gdb.fortran/array-slices.f90's show_1d
subroutine -- it was missing printing newline at the end of the
"Expected GDB Output" text, leading to a test timeout.  All other
subroutines end with advance=yes, except this one.  Fix it by using
advance=yes here too.

Change-Id: I4640729f334431cfc24b0917e7d3977b677c6ca5
---
 gdb/testsuite/gdb.fortran/array-slices.exp  | 30 +++++++++++++++------
 gdb/testsuite/gdb.fortran/array-slices.f90  |  2 +-
 gdb/testsuite/gdb.fortran/lbound-ubound.exp | 28 ++++++++++++++-----
 3 files changed, 45 insertions(+), 15 deletions(-)

diff --git a/gdb/testsuite/gdb.fortran/array-slices.exp b/gdb/testsuite/gdb.fortran/array-slices.exp
index 8f9c012e399..a38276c67e4 100644
--- a/gdb/testsuite/gdb.fortran/array-slices.exp
+++ b/gdb/testsuite/gdb.fortran/array-slices.exp
@@ -35,6 +35,11 @@
 
 if {[skip_fortran_tests]} { return -1 }
 
+# This test relies on output from the inferior.
+if [target_info exists gdb,noinferiorio] {
+    return -1
+}
+
 standard_testfile ".f90"
 load_lib fortran.exp
 
@@ -62,12 +67,6 @@ proc run_test { repack } {
 	return -1
     }
 
-    # This test relies on output from the inferior and is not supported on
-    # remote targets.
-    if {![gdb_is_target_native]} {
-	return 0
-    }
-
     # Avoid libc symbols, in particular the 'array' type.
     gdb_test_no_output "nosharedlibrary"
 
@@ -97,11 +96,19 @@ proc run_test { repack } {
 	    set found_final_breakpoint false
 	    set expected_result ""
 	    set func_name ""
+	    set found_prompt false
 	    gdb_test_multiple "continue" "continue" {
+		-i $::inferior_spawn_id
+
 		-re ".*GDB = (\[^\r\n\]+)\r\n" {
 		    set expected_result $expect_out(1,string)
-		    exp_continue
+		    if {!$found_prompt} {
+			exp_continue
+		    }
 		}
+
+		-i $::gdb_spawn_id
+
 		-re "! Display Element" {
 		    set func_name "show_elem"
 		    exp_continue
@@ -119,7 +126,14 @@ proc run_test { repack } {
 		    exp_continue
 		}
 		-re "$gdb_prompt $" {
-		    # We're done.
+		    set found_prompt true
+
+		    if {$found_final_breakpoint
+			|| ($expected_result != "" && $func_name != "")} {
+			# We're done.
+		    } else {
+			exp_continue
+		    }
 		}
 	    }
 
diff --git a/gdb/testsuite/gdb.fortran/array-slices.f90 b/gdb/testsuite/gdb.fortran/array-slices.f90
index ff3964ffb6b..1eaee82b7b7 100644
--- a/gdb/testsuite/gdb.fortran/array-slices.f90
+++ b/gdb/testsuite/gdb.fortran/array-slices.f90
@@ -61,7 +61,7 @@ subroutine show_1d (array)
      end if
      write(*, fmt="(I0)", advance="no") array (i)
   end do
-  write(*, fmt="(A)", advance="no") ")"
+  write(*, fmt="(A)", advance="yes") ")"
 
   print *, ""	! Display Array Slice 1D
 end subroutine show_1d
diff --git a/gdb/testsuite/gdb.fortran/lbound-ubound.exp b/gdb/testsuite/gdb.fortran/lbound-ubound.exp
index 671b251c799..709b74a69e8 100644
--- a/gdb/testsuite/gdb.fortran/lbound-ubound.exp
+++ b/gdb/testsuite/gdb.fortran/lbound-ubound.exp
@@ -32,9 +32,8 @@ if ![fortran_runto_main] {
     return -1
 }
 
-# This test relies on output from the inferior and is not supported on
-# remote targets.
-if {![gdb_is_target_native]} {
+# This test relies on output from the inferior.
+if [target_info exists gdb,noinferiorio] {
    return 0
 }
 
@@ -55,15 +54,25 @@ while { $test_count < 500 } {
 
 	set expected_lbound ""
 	set expected_ubound ""
+	set found_prompt false
 	gdb_test_multiple "continue" "continue" {
+	    -i $::inferior_spawn_id
+
 	    -re ".*LBOUND = (\[^\r\n\]+)\r\n" {
 		set expected_lbound $expect_out(1,string)
-		exp_continue
+		if {!$found_prompt} {
+		    exp_continue
+		}
 	    }
 	    -re ".*UBOUND = (\[^\r\n\]+)\r\n" {
 		set expected_ubound $expect_out(1,string)
-		exp_continue
+		if {!$found_prompt} {
+		    exp_continue
+		}
 	    }
+
+	    -i $::gdb_spawn_id
+
 	    -re "! Test Breakpoint" {
 		set func_name "show_elem"
 		exp_continue
@@ -73,7 +82,14 @@ while { $test_count < 500 } {
 		exp_continue
 	    }
 	    -re "$gdb_prompt $" {
-		# We're done.
+		set found_prompt true
+
+		if {$found_final_breakpoint
+		    || ($expected_lbound != "" && $expected_ubound != "")} {
+		    # We're done.
+		} else {
+		    exp_continue
+		}
 	    }
 	}
 

base-commit: 0daa5af85a9b5548a1bd7e8085a326162b41fd6f
-- 
2.26.2

Re: [PATCH] Make gdb.fortran/{array-slices,lbound-ubound} work against gdbserver

[Tom Tromey]

>>>>> "Pedro" == Pedro Alves <pedro@palves.net> writes:

Pedro> Keith mentioned he might be away this week, so I went ahead and played with this.  We
Pedro> can't use gdb_test_stdio here, because that maps to gdb_test, and we want gdb_test_multiple
Pedro> instead.  So we use inferior_spawn_id with gdb_test_multiple instead.

Thanks.

Pedro> It took me some head scratching and staring at logs to figure out why the patch
Pedro> wasn't working -- I was hitting a buglet in gdb.fortran/array-slices.f90, argh.
Pedro> With that out of the way, both testcases now pass cleanly with gdbserver too.

Pedro> Here's the proposed patch.  Any comments?

It looks fine to me.

Tom