[PATCH] Fix detection of gfortran compilers.

[PATCH] Fix detection of gfortran compilers.

[Bernhard Heckel]

From: "bernhard.heckel" <bernhard.heckel@intel.com>

Newer gfortran compilers changed naming convention in DW_AT_producer tag.
For example "GNU Fortran 4.6.3" vs "GNU Fortran2008 5.3.0"

2016-02-10  Bernhard Heckel  <bernhard.heckel@intel.com>

gdb/Changelog:

     * dwarf2read.c (new_symbol_full): Fix detection of gfortran compilers.

---
 gdb/dwarf2read.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
index 1020c12..dcd49e3 100644
--- a/gdb/dwarf2read.c
+++ b/gdb/dwarf2read.c
@@ -18506,7 +18506,7 @@ new_symbol_full (struct die_info *die, struct type *type, struct dwarf2_cu *cu,
 		  if (cu->language == language_fortran && die->parent
 		      && die->parent->tag == DW_TAG_module
 		      && cu->producer
-		      && startswith (cu->producer, "GNU Fortran "))
+		      && startswith (cu->producer, "GNU Fortran"))
 		    SYMBOL_ACLASS_INDEX (sym) = LOC_UNRESOLVED;
 
 		  /* A variable with DW_AT_external is never static,
-- 
2.7.1.339.g0233b80

[PATCH] testsuite: Nullify pointers before first usage.

[Bernhard Heckel]

From: "bernhard.heckel" <bernhard.heckel@intel.com>

Nullify pointers to avoid an undefined association status.

2016-02-18  Bernhard Heckel  <bernhard.heckel@intel.com>

gdb/Changelog:

     * vla.f90: Nullify pointer after declaration.

---
 gdb/testsuite/gdb.mi/vla.f90 | 1 +
 1 file changed, 1 insertion(+)

diff --git a/gdb/testsuite/gdb.mi/vla.f90 b/gdb/testsuite/gdb.mi/vla.f90
index 10c3dcc..c3e48dc 100644
--- a/gdb/testsuite/gdb.mi/vla.f90
+++ b/gdb/testsuite/gdb.mi/vla.f90
@@ -18,6 +18,7 @@ program vla
   real, target, allocatable :: vla2(:, :)
   real, pointer :: pvla2 (:, :)
   logical :: l
+  nullify (pvla2)
 
   allocate (vla1 (5))         ! vla1-not-allocated
   l = allocated(vla1)         ! vla1-allocated
-- 
2.7.1.339.g0233b80

[PATCH] testsuite: Fix timeout issues during print of vla-arrays.

[Bernhard Heckel]

From: "bernhard.heckel" <bernhard.heckel@intel.com>

Printing and resolving of dynamic array's causes sporadic timeout issues on loaded systems.

2016-02-12  Bernhard Heckel  <bernhard.heckel@intel.com>

gdb/Changelog:

     * vla-history.exp: Lookup array elements and printing exceeds timeout.


Conflicts:

	gdb/testsuite/gdb.fortran/vla-history.exp
	gdb/testsuite/gdb.fortran/vla-value.exp
---
 gdb/testsuite/gdb.fortran/vla-history.exp | 17 +++++++++--------
 1 file changed, 9 insertions(+), 8 deletions(-)

diff --git a/gdb/testsuite/gdb.fortran/vla-history.exp b/gdb/testsuite/gdb.fortran/vla-history.exp
index 1478e6b..2012961 100644
--- a/gdb/testsuite/gdb.fortran/vla-history.exp
+++ b/gdb/testsuite/gdb.fortran/vla-history.exp
@@ -30,19 +30,20 @@ gdb_breakpoint [gdb_get_line_number "vla1-init"]
 gdb_continue_to_breakpoint "vla1-init"
 gdb_test "print vla1" " = <not allocated>" "print non-allocated vla1"
 
-gdb_breakpoint [gdb_get_line_number "vla2-allocated"]
-gdb_continue_to_breakpoint "vla2-allocated"
-
 gdb_breakpoint [gdb_get_line_number "vla1-filled"]
 gdb_continue_to_breakpoint "vla1-filled"
-gdb_test "print vla1" \
-  " = \\( *\\( *\\( *1311, *1311, *1311,\[()1311, .\]*\\)" \
-  "print vla1 filled"
+with_timeout_factor 2 {
+    gdb_test "print vla1" \
+      " = \\( *\\( *\\( *1311, *1311, *1311,\[()1311, .\]*\\)" \
+      "print vla1 filled"
+}
 
 # Try to access history values for full vla prints.
 gdb_test "print \$1" " = <not allocated>" "print \$1"
-gdb_test "print \$2" \
-  " = \\( *\\( *\\( *1311, *1311, *1311,\[()1311, .\]*\\)" "print \$2"
+with_timeout_factor 2 {
+    gdb_test "print \$2" \
+      " = \\( *\\( *\\( *1311, *1311, *1311,\[()1311, .\]*\\)" "print \$4"
+}
 
 gdb_breakpoint [gdb_get_line_number "vla2-filled"]
 gdb_continue_to_breakpoint "vla2-filled"
-- 
2.7.1.339.g0233b80

[PATCH] testsuite: Fix run to main issue introduced by GCC 5.x.

[Bernhard Heckel]

From: "bernhard.heckel" <bernhard.heckel@intel.com>

Adding an dummy assignment as an new breakpoint anchor because
breakpoint on return statement doesn't work for GCC 5.x.

2016-02-16  Bernhard Heckel  <bernhard.heckel@intel.com>

gdb/Changelog:

     * vla-cxx.cc: Insert dummy assignment as anchor for an breakpoint.

---
 gdb/testsuite/gdb.cp/vla-cxx.cc | 1 +
 1 file changed, 1 insertion(+)

diff --git a/gdb/testsuite/gdb.cp/vla-cxx.cc b/gdb/testsuite/gdb.cp/vla-cxx.cc
index 1eb80f2..a1fd510 100644
--- a/gdb/testsuite/gdb.cp/vla-cxx.cc
+++ b/gdb/testsuite/gdb.cp/vla-cxx.cc
@@ -45,5 +45,6 @@ int main(int argc, char **argv)
     vla[i] = 5 + 2 * i;
 
   // vlas_filled
+  vla[0] = 2 * vla[0];
   return vla[2];
 }
-- 
2.7.1.339.g0233b80

Re: [PATCH] testsuite: Fix timeout issues during print of vla-arrays.

[Luis Machado]

On 02/22/2016 11:49 AM, Bernhard Heckel wrote:
> From: "bernhard.heckel" <bernhard.heckel@intel.com>
>
> Printing and resolving of dynamic array's causes sporadic timeout issues on loaded systems.
>
> 2016-02-12  Bernhard Heckel  <bernhard.heckel@intel.com>
>
> gdb/Changelog:
>
>       * vla-history.exp: Lookup array elements and printing exceeds timeout.
>
>
> Conflicts:
>
> 	gdb/testsuite/gdb.fortran/vla-history.exp
> 	gdb/testsuite/gdb.fortran/vla-value.exp
> ---
>   gdb/testsuite/gdb.fortran/vla-history.exp | 17 +++++++++--------
>   1 file changed, 9 insertions(+), 8 deletions(-)
>
> diff --git a/gdb/testsuite/gdb.fortran/vla-history.exp b/gdb/testsuite/gdb.fortran/vla-history.exp
> index 1478e6b..2012961 100644
> --- a/gdb/testsuite/gdb.fortran/vla-history.exp
> +++ b/gdb/testsuite/gdb.fortran/vla-history.exp
> @@ -30,19 +30,20 @@ gdb_breakpoint [gdb_get_line_number "vla1-init"]
>   gdb_continue_to_breakpoint "vla1-init"
>   gdb_test "print vla1" " = <not allocated>" "print non-allocated vla1"
>
> -gdb_breakpoint [gdb_get_line_number "vla2-allocated"]
> -gdb_continue_to_breakpoint "vla2-allocated"
> -

Is the change above spurious? I see code being moved to higher timeout 
blocks, but the above doesn't seem to reappear anywhere. It was just 
deleted.

Did it cause problems?

Re: [PATCH] testsuite: Fix timeout issues during print of vla-arrays.

[Heckel, Bernhard]

On 23/02/2016 20:37, Luis Machado wrote:
> On 02/22/2016 11:49 AM, Bernhard Heckel wrote:
>> From: "bernhard.heckel" <bernhard.heckel@intel.com>
>>
>> Printing and resolving of dynamic array's causes sporadic timeout 
>> issues on loaded systems.
>>
>> 2016-02-12  Bernhard Heckel  <bernhard.heckel@intel.com>
>>
>> gdb/Changelog:
>>
>>       * vla-history.exp: Lookup array elements and printing exceeds 
>> timeout.
>>
>>
>> Conflicts:
>>
>>     gdb/testsuite/gdb.fortran/vla-history.exp
>>     gdb/testsuite/gdb.fortran/vla-value.exp
>> ---
>>   gdb/testsuite/gdb.fortran/vla-history.exp | 17 +++++++++--------
>>   1 file changed, 9 insertions(+), 8 deletions(-)
>>
>> diff --git a/gdb/testsuite/gdb.fortran/vla-history.exp 
>> b/gdb/testsuite/gdb.fortran/vla-history.exp
>> index 1478e6b..2012961 100644
>> --- a/gdb/testsuite/gdb.fortran/vla-history.exp
>> +++ b/gdb/testsuite/gdb.fortran/vla-history.exp
>> @@ -30,19 +30,20 @@ gdb_breakpoint [gdb_get_line_number "vla1-init"]
>>   gdb_continue_to_breakpoint "vla1-init"
>>   gdb_test "print vla1" " = <not allocated>" "print non-allocated vla1"
>>
>> -gdb_breakpoint [gdb_get_line_number "vla2-allocated"]
>> -gdb_continue_to_breakpoint "vla2-allocated"
>> -
>
> Is the change above spurious? I see code being moved to higher timeout 
> blocks, but the above doesn't seem to reappear anywhere. It was just 
> deleted.
>
> Did it cause problems?
The vla2-allocated breakpoint was not used at all therefore I removed it 
from the test script when I investigated on the timeout issue. There 
were no problems with the breakpoint.

Regards,
Bernhard

-- 

Intel Deutschland GmbH
Registered Address: Am Campeon 10-12, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Christin Eisenschmid, Christian Lamprechter
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928

Intel Deutschland GmbH
Registered Address: Am Campeon 10-12, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Christin Eisenschmid, Christian Lamprechter
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928

Re: [PATCH] testsuite: Fix timeout issues during print of vla-arrays.

[Yao Qi]

"Heckel, Bernhard" <bernhard.heckel@intel.com> writes:

> The vla2-allocated breakpoint was not used at all therefore I removed
> it from the test script when I investigated on the timeout
> issue. There were no problems with the breakpoint.

Let us remove the unused code in a separate patch then.

-- 
Yao (齐尧)

Re: [PATCH] testsuite: Fix timeout issues during print of vla-arrays.

[Heckel, Bernhard]

On 24/02/2016 15:47, Yao Qi wrote:
> "Heckel, Bernhard" <bernhard.heckel@intel.com> writes:
>
>> The vla2-allocated breakpoint was not used at all therefore I removed
>> it from the test script when I investigated on the timeout
>> issue. There were no problems with the breakpoint.
> Let us remove the unused code in a separate patch then.
>
Thanks for reviewing.
Based on your inputs I split the patch.
You will find them here:

[PATCH V2 1/2] testsuite: Fix timeout issues during print of vla-arrays.
https://sourceware.org/ml/gdb-patches/2016-02/msg00783.html

[PATCH V2 2/2] testsuite: Remove unnecessary code in fortran vla-history 
test.
https://sourceware.org/ml/gdb-patches/2016-02/msg00782.html


Regards,
Bernhard

-- 


Intel Deutschland GmbH
Registered Address: Am Campeon 10-12, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Christin Eisenschmid, Christian Lamprechter
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928

Intel Deutschland GmbH
Registered Address: Am Campeon 10-12, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Christin Eisenschmid, Christian Lamprechter
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928

Re: [PATCH] Fix detection of gfortran compilers.

[Joel Brobecker]

> 2016-02-10  Bernhard Heckel  <bernhard.heckel@intel.com>
> 
> gdb/Changelog:
> 
>      * dwarf2read.c (new_symbol_full): Fix detection of gfortran compilers.

OK to push!

-- 
Joel

Re: [PATCH] testsuite: Nullify pointers before first usage.

[Joel Brobecker]

> Nullify pointers to avoid an undefined association status.
> 
> 2016-02-18  Bernhard Heckel  <bernhard.heckel@intel.com>
> 
> gdb/Changelog:

This should be "gdb/testsuite/ChangeLog".

>      * vla.f90: Nullify pointer after declaration.

The name of the file should be "gdb.mi/vla.f90". Basically, the name
is relative to the directory where the ChangeLog is.

Other than that, the patch looks good to me.

Thanks!

-- 
Joel

Re: [PATCH] testsuite: Fix run to main issue introduced by GCC 5.x.

[Joel Brobecker]

Hello,

On Mon, Feb 22, 2016 at 03:49:17PM +0100, Bernhard Heckel wrote:
> From: "bernhard.heckel" <bernhard.heckel@intel.com>
> 
> Adding an dummy assignment as an new breakpoint anchor because

"an dummy" -> "a dummy"; same for "an new" -> "a new".


> breakpoint on return statement doesn't work for GCC 5.x.

"the breakpoint on the return statement".

> 2016-02-16  Bernhard Heckel  <bernhard.heckel@intel.com>
> 
> gdb/Changelog:
> 
>      * vla-cxx.cc: Insert dummy assignment as anchor for an breakpoint.

Same remark as before for the name of the ChangeLog file and the name
of the file being changed.

Ok with those changes.

-- 
Joel

Re: [PATCH] testsuite: Fix run to main issue introduced by GCC 5.x.

[Heckel, Bernhard]

On 26/02/2016 17:55, Joel Brobecker wrote:
> Hello,
>
> On Mon, Feb 22, 2016 at 03:49:17PM +0100, Bernhard Heckel wrote:
>> From: "bernhard.heckel" <bernhard.heckel@intel.com>
>>
>> Adding an dummy assignment as an new breakpoint anchor because
> "an dummy" -> "a dummy"; same for "an new" -> "a new".
>
>
>> breakpoint on return statement doesn't work for GCC 5.x.
> "the breakpoint on the return statement".
>
>> 2016-02-16  Bernhard Heckel  <bernhard.heckel@intel.com>
>>
>> gdb/Changelog:
>>
>>       * vla-cxx.cc: Insert dummy assignment as anchor for an breakpoint.
> Same remark as before for the name of the ChangeLog file and the name
> of the file being changed.
>
> Ok with those changes.
>
Done and pushed.

Regards,
Bernhard

-- 

Intel Deutschland GmbH
Registered Address: Am Campeon 10-12, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Christin Eisenschmid, Christian Lamprechter
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928

Intel Deutschland GmbH
Registered Address: Am Campeon 10-12, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Christin Eisenschmid, Christian Lamprechter
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928

Re: [PATCH] testsuite: Nullify pointers before first usage.

[Heckel, Bernhard]

On 26/02/2016 17:49, Joel Brobecker wrote:
>> Nullify pointers to avoid an undefined association status.
>>
>> 2016-02-18  Bernhard Heckel  <bernhard.heckel@intel.com>
>>
>> gdb/Changelog:
> This should be "gdb/testsuite/ChangeLog".
>
>>       * vla.f90: Nullify pointer after declaration.
> The name of the file should be "gdb.mi/vla.f90". Basically, the name
> is relative to the directory where the ChangeLog is.
>
> Other than that, the patch looks good to me.
>
> Thanks!
>
Done and pushed.

Regards,
Bernhard

-- 

Intel Deutschland GmbH
Registered Address: Am Campeon 10-12, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Christin Eisenschmid, Christian Lamprechter
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928

Intel Deutschland GmbH
Registered Address: Am Campeon 10-12, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Christin Eisenschmid, Christian Lamprechter
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928

Re: [PATCH] Fix detection of gfortran compilers.

[Heckel, Bernhard]

On 26/02/2016 17:47, Joel Brobecker wrote:
>> 2016-02-10  Bernhard Heckel  <bernhard.heckel@intel.com>
>>
>> gdb/Changelog:
>>
>>       * dwarf2read.c (new_symbol_full): Fix detection of gfortran compilers.
> OK to push!
>
Pushed.

Regards,
Bernhard

-- 

Intel Deutschland GmbH
Registered Address: Am Campeon 10-12, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Christin Eisenschmid, Christian Lamprechter
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928

Intel Deutschland GmbH
Registered Address: Am Campeon 10-12, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Christin Eisenschmid, Christian Lamprechter
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928