[PATCH] fix uninitialized variable in testsuite

[PATCH] fix uninitialized variable in testsuite

[Sandra Loosemore]

[-- Attachment #1: Type: text/plain, Size: 127 bytes --]

I found that the test cls_many_mixed_args.c was failing due to an 
uninitialized variable.  This patch fixes it.  OK?

-Sandra

[-- Attachment #2: libffi-testsuite-fix.log --]
[-- Type: text/x-log, Size: 154 bytes --]

2013-10-10  Sandra Loosemore  <sandra@codesourcery.com>

	* testsuite/libffi.call/cls_many_mixed_args.c (cls_ret_double_fn):
	Fix uninitialized variable.

[-- Attachment #3: libffi-testsuite-fix.patch --]
[-- Type: text/x-patch, Size: 461 bytes --]

diff --git a/testsuite/libffi.call/cls_many_mixed_args.c b/testsuite/libffi.call/cls_many_mixed_args.c
index fc75b1d..18cdfcc 100644
--- a/testsuite/libffi.call/cls_many_mixed_args.c
+++ b/testsuite/libffi.call/cls_many_mixed_args.c
@@ -15,7 +15,7 @@ static void cls_ret_double_fn(ffi_cif* cif __UNUSED__, void* resp, void** args,
 			      void* userdata __UNUSED__)
 {
   int i;
-  double r;
+  double r = 0;
   double t;
   for(i = 0; i < NARGS; i++)
     {

Re: [PATCH] fix uninitialized variable in testsuite

[Anthony Green]

Good catch, Sandra.  I've committed this and the nios2 port.  Out of
curiosity, how did you test the nios2 port?

Thanks!

AG


On Thu, Oct 10, 2013 at 9:40 PM, Sandra Loosemore
<sandra@codesourcery.com> wrote:
> I found that the test cls_many_mixed_args.c was failing due to an
> uninitialized variable.  This patch fixes it.  OK?
>
> -Sandra

Re: [PATCH] fix uninitialized variable in testsuite

[Jay]

From her other mail:
"Unfortunately, the Nios II GCC port hasn't been committed to the FSF repository yet; it's been accepted by the steering committee but is still stuck awaiting technical review.  I used Mentor's GCC 4.7-based toolchain for testing instead."

 - Jay

On Oct 15, 2013, at 12:39 PM, Anthony Green <green@moxielogic.com> wrote:

> Good catch, Sandra.  I've committed this and the nios2 port.  Out of
> curiosity, how did you test the nios2 port?
> 
> Thanks!
> 
> AG
> 
> 
> On Thu, Oct 10, 2013 at 9:40 PM, Sandra Loosemore
> <sandra@codesourcery.com> wrote:
>> I found that the test cls_many_mixed_args.c was failing due to an
>> uninitialized variable.  This patch fixes it.  OK?
>> 
>> -Sandra

Re: [PATCH] fix uninitialized variable in testsuite

[Anthony Green]

Sure - I saw that. I was just wondering if it was against a simulator,
real hardware, or..?  I work on a Nios-like soft-core as well (moxie),
and am just curious.

On Tue, Oct 15, 2013 at 9:41 PM, Jay <jay.krell@cornell.edu> wrote:
> From her other mail:
> "Unfortunately, the Nios II GCC port hasn't been committed to the FSF repository yet; it's been accepted by the steering committee but is still stuck awaiting technical review.  I used Mentor's GCC 4.7-based toolchain for testing instead."
>
>  - Jay
>
> On Oct 15, 2013, at 12:39 PM, Anthony Green <green@moxielogic.com> wrote:
>
>> Good catch, Sandra.  I've committed this and the nios2 port.  Out of
>> curiosity, how did you test the nios2 port?
>>
>> Thanks!
>>
>> AG
>>
>>
>> On Thu, Oct 10, 2013 at 9:40 PM, Sandra Loosemore
>> <sandra@codesourcery.com> wrote:
>>> I found that the test cls_many_mixed_args.c was failing due to an
>>> uninitialized variable.  This patch fixes it.  OK?
>>>
>>> -Sandra

Re: [PATCH] fix uninitialized variable in testsuite

[Sandra Loosemore]

On 10/15/2013 08:15 PM, Anthony Green wrote:
> Sure - I saw that. I was just wondering if it was against a simulator,
> real hardware, or..?  I work on a Nios-like soft-core as well (moxie),
> and am just curious.

We're using real hardware for testing; a 3c120 board provided to us by 
Altera.

-Sandra