printf() for %f is broken since one of the last merges

[LorenzMeier]

Output of tests float:

nsh> tests float

--- SINGLE PRECISION TESTS ---
The single precision test involves calls to fabsf(),
if test fails check this function as well.

         success: sinf(0.0f) == 0.0f
         success: sinf(1.0f) == 0.84147f
         success: asinf(1.0f) == 1.57079f
         success: cosf(1.0f) == 0.54030f
         success: acosf(1.0f) == 0.0f
         success: sinf(0.1f) == 0.09983f
         success: sqrt(2.0f) == 1.41421f
         success: atan2f(1.0f, 1.0f) == 0.78539f
         FAIL: printf("%8.4f", 0.553415f) != "  0.5534", result:   0.0g34

--- DOUBLE PRECISION TESTS ---
         success: 1.0111 * 2.0 == 2.0222
         success: (float) 1.55f == 1.55 (double)
         success: sin(0.0) == 0.0
         success: sin(1.0) == 0.84147098480
         success: atan2(1.0, 1.0) == 0.785398
         testing pow() with magic value
           (44330.0 * (1.0 - pow((96286LL / 101325.0), 0.190295)));
         success: result: 0Z8.2293
         FAIL: printf("%8.4f", 0.553415) != "  0.5534", result:   0.0g34

 FAIL: One or more tests failed.

[px4dev]

I can't reproduce this with a clean build of master:

gcc version 4.6.2 20120613 (release) [ARM/embedded-4_6-branch revision 188521](GNU Tools for ARM Embedded Processors)

nsh> tests float

--- SINGLE PRECISION TESTS ---
The single precision test involves calls to fabsf(),
if test fails check this function as well.

 success: sinf(0.0f) == 0.0f
 success: sinf(1.0f) == 0.84147f
 success: asinf(1.0f) == 1.57079f
 success: cosf(1.0f) == 0.54030f
 success: acosf(1.0f) == 0.0f
 success: sinf(0.1f) == 0.09983f
 success: sqrt(2.0f) == 1.41421f
 success: atan2f(1.0f, 1.0f) == 0.78539f
 success: printf("%8.4f", 0.553415f) ==   0.5534

--- DOUBLE PRECISION TESTS ---
success: 1.0111 * 2.0 == 2.0222
success: (float) 1.55f == 1.55 (double)
success: sin(0.0) == 0.0
success: sin(1.0) == 0.84147098480
success: atan2(1.0, 1.0) == 0.785398
testing pow() with magic value
(44330.0 * (1.0 - pow((96286LL / 101325.0), 0.190295)));
success: result: 428.2293
success: printf("%8.4f", 0.553415) == 0.5534

SUCCESS: All float and double tests passed.

[LorenzMeier]

Still had this old GCC somehow, upgrading now and re-testing

arm-none-eabi-gcc (GNU Tools for ARM Embedded Processors) 4.6.2 20110921 (release) [ARM/embedded-4_6-branch revision 182083]

[LorenzMeier]

With the new gcc this issue vanished - seems to be a compiler bug. Closing.

arm-none-eabi-gcc (GNU Tools for ARM Embedded Processors) 4.6.2 20120613 (release) [ARM/embedded-4_6-branch revision 188521]
Copyright (C) 2011 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.