[test-suite] Fix ambigous call to overloaded function isnan

[sdardis]

Summary

Older versions of glibc expose isnan as C99 macros and C++ functions, causing
a compilation failure of the CLAMR benchmark. Address this by explicitly calling
the c++ version of this function.

For reference: https://sourceware.org/bugzilla/show_bug.cgi?id=19439,
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=48891

[hfinkel]

LGTM

[MatzeB]

I guess a more C++y solution would be to #include <cmath> and using std::isnan() instead of using the C headers. But I guess this is fine for now.

[sdardis]

Update as per @MatzeB 's comment, checking before commit showed it broke the build on OSX.

[MatzeB]

In D42411#987999, @sdardis wrote:

Update as per @MatzeB 's comment, checking before commit showed it broke the build on OSX.

It worked for me like this, but that may be because I have a new clang version installed. This may indeed fail with older versions unless -std=c++11 is passed to the compiler.

Anyway let's leave it as is, if it breaks the build on your side.

[sdardis]

Anyway let's leave it as is, if it breaks the build on your side.

Just to double check, the 'it' you're referring to is the build system / compilation options for this test and you're happy with the patch as is?

[MatzeB]

In D42411#989024, @sdardis wrote:

Anyway let's leave it as is, if it breaks the build on your side.

Just to double check, the 'it' you're referring to is the build system / compilation options for this test and you're happy with the patch as is?

I'm fine with leaving the code in the test-suite as is (no need for this followup patch).

[sdardis]

I'm fine with leaving the code in the test-suite as is (no need for this followup patch).

Ah, I think I might have lead you to believe I had committed the previous version of this patch, which just used:

::isnan

to work around the failure with glibc.

Update as per @MatzeB 's comment, checking before commit showed it broke the build on OSX.

Here I should have stated that I checked building the test-suite on OSX before committing the previous version of this patch but when I saw it broke the build with on OSX, I did not commit that version and posted the current version which is against trunk.

My apologies.

[sdardis]

@MatzeB ping?

[MatzeB]

Just commit whatever version you see fit, I'm fine with any of the proposals; the bots will come back to you and tell you if things don't build somewhere :)

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C++/CLAMR/clamr_cpuonly.cpp

  *  Authors: Bob Robey       XCP-2   [e-mail]
  *           Neal Davis              [e-mail], [e-mail]
  *           David Nicholaeff        [e-mail], [e-mail]
  *           Dennis Trujillo         [e-mail], [e-mail]
  * 
  */
 
 #include <algorithm>
-#include <math.h>
+#include <cmath>
 #include <stdio.h>
 #include <stdlib.h>
 #include <sys/time.h>
 #include <unistd.h>
 #include <vector>
 //#include "graphics/display.h"
 #include "graphics.h"
 #include "input.h"
#endif
    } // End burst loop
 
    cpu_time_calcs += cpu_timer_stop(tstart_cpu);
 
    H_sum = state->mass_sum(enhanced_precision_sum);
 
    int error_status = STATUS_OK;
 
-   if (isnan(H_sum)) {
+   if (std::isnan(H_sum)) {
       printf("Got a NAN on cycle %d\n",ncycle);
       error_status = STATUS_NAN;
    }
 
    double percent_mass_diff = fabs(H_sum - H_sum_initial)/H_sum_initial * 100.0;
    if (percent_mass_diff >= upper_mass_diff_percentage) {
       printf("Mass difference outside of acceptable range on cycle %d percent_mass_diff %lg upper limit %lg\n",ncycle,percent_mass_diff, upper_mass_diff_percentage);
       error_status = STATUS_MASS_LOSS;