Make sure malloc is linked into gdb.cp/oranking.cc.

On some platforms, e.g., arm-eabi-none, we need to make certain that
malloc is linked into the program because oranking.exp uses function
calls requiring it:

(gdb) p foo101("abc")
evaluation of this expression requires the program to have a function "malloc".

While not a big deal (this is /not/ a regression), I think this would be
appropriate for inclusion in 8.0.

gdb/testsuite/ChangeLog

	* gdb.cp/oranking.cc (dummy): New function to grab malloc.
	(main): Call it.
---
 gdb/testsuite/ChangeLog          |  5 +++++
 gdb/testsuite/gdb.cp/oranking.cc | 17 +++++++++++++++++
 2 files changed, 22 insertions(+)

Keith Seitz <keiths@redhat.com> writes:

> While not a big deal (this is /not/ a regression), I think this would be
> appropriate for inclusion in 8.0.

Yes.

>
> gdb/testsuite/ChangeLog
>
> 	* gdb.cp/oranking.cc (dummy): New function to grab malloc.
> 	(main): Call it.

Patch is good to me.

Hi,

On 02/05/17 19:42, Keith Seitz wrote:
> On some platforms, e.g., arm-eabi-none, we need to make certain that
> malloc is linked into the program because oranking.exp uses function
> calls requiring it:
>
> (gdb) p foo101("abc")
> evaluation of this expression requires the program to have a function "malloc".
>
> While not a big deal (this is /not/ a regression), I think this would be
> appropriate for inclusion in 8.0.
>
> gdb/testsuite/ChangeLog
>
> 	* gdb.cp/oranking.cc (dummy): New function to grab malloc.
> 	(main): Call it.
> ---
>  gdb/testsuite/ChangeLog          |  5 +++++
>  gdb/testsuite/gdb.cp/oranking.cc | 17 +++++++++++++++++
>  2 files changed, 22 insertions(+)
>
> diff --git a/gdb/testsuite/ChangeLog b/gdb/testsuite/ChangeLog
> index 9c0d68b..2e587d6 100644
> --- a/gdb/testsuite/ChangeLog
> +++ b/gdb/testsuite/ChangeLog
> @@ -1,3 +1,8 @@
> +2017-05-02  Keith Seitz  <keiths@redhat.com>
> +
> +	* gdb.cp/oranking.cc (dummy): New function to grab malloc.
> +	(main): Call it.
> +
>  2017-05-01  Tim Wiederhake  <tim.wiederhake@intel.com>
>
>  	* gdb.python/py-record-btrace.exp: Rename prev_sibling and next_sibling
> diff --git a/gdb/testsuite/gdb.cp/oranking.cc b/gdb/testsuite/gdb.cp/oranking.cc
> index bd2f51b..1357610 100644
> --- a/gdb/testsuite/gdb.cp/oranking.cc
> +++ b/gdb/testsuite/gdb.cp/oranking.cc
> @@ -1,3 +1,18 @@
> +#include <cstdlib>
> +
> +/* Make sure `malloc' is linked into the program.  If we don't, tests
> +   in the accompanying expect file may fail:
> +
> +   evaluation of this expression requires the program to have a function
> +   "malloc".  */
> +
> +void
> +dummy ()
> +{
> +  void *p = malloc (16);
> +
> +  free (p);
> +}
>
>  /* 1. A standard covnersion sequence is better than a user-defined sequence
>        which is better than an elipses conversion sequence.  */
> @@ -165,6 +180,8 @@ test15 ()
>  }
>
>  int main() {
> +  dummy ();
> +
>    B b;
>    foo0(b);
>    foo1(b);
>

Thanks Keith. I've tested the patch on arm-none-eabi and this solve the issue at 
end. I can also confirm that this is not a regression, the KFAIL->FAIL is only 
due to the the removal of setup_kfail in 
e15c3eb45bdc8bd5717fd5ceddcc30c3de07b58f, the error was already there before the 
commit.

Best regards,

Thomas

On 05/03/2017 03:38 AM, Thomas Preudhomme wrote:

> Thanks Keith. I've tested the patch on arm-none-eabi and this solve
> the issue at end. I can also confirm that this is not a regression,
> the KFAIL->FAIL is only due to the the removal of setup_kfail in
> e15c3eb45bdc8bd5717fd5ceddcc30c3de07b58f, the error was already there
> before the commit.

Excellent. I've pushed the patch to both master and the 8.0 branch.

Thank you for following up on this.

Keith