[patch, libgfortran] Bug 78123 - Short reads with T edit descriptor not padding correctly

[Jerry DeLisle]

On 11/04/2016 07:11 AM, Christophe Lyon wrote:
> On 4 November 2016 at 14:48, Christophe Lyon <christophe.lyon@linaro.org> wrote:
>> On 30 October 2016 at 21:40, Thomas Koenig <tkoenig@netcologne.de> wrote:
>>> Hi Jerry,
>>>
--- snip ---
>>
>> The new test fails at execution on arm and aarch64:
>>
>>   gfortran.dg/fmt_t_9.f   -O0  execution test
>>   gfortran.dg/fmt_t_9.f   -O1  execution test
>>   gfortran.dg/fmt_t_9.f   -O2  execution test
>>   gfortran.dg/fmt_t_9.f   -O3 -fomit-frame-pointer -funroll-loops
>> -fpeel-loops -ftracer -finline-functions  execution test
>>   gfortran.dg/fmt_t_9.f   -O3 -g  execution test
>>   gfortran.dg/fmt_t_9.f   -Os  execution test
>>
>> seen on gcc-5 and gcc-6 branches. Maybe I missed it on trunk due to
>> larger granularity in validations.

Hi Christophe,

The patch is very non target specific. All I can think at the moment is its 
related to buffering and flushing details, or an error in the test case.

When I get a chance, I would like to send you a test program to maybe help debug 
this.  Or, is there a machine available I cant ssh into and work on it there?

Jerry