Summary: | Compile error on Solaris 8 - colorspaces/rgb_f32/kis_rgb_f32_colorspace.cc uses powf which is not available | ||
---|---|---|---|
Product: | [Applications] krita | Reporter: | Steve Evans <stevee> |
Component: | General | Assignee: | Adrian Page <adrian> |
Status: | RESOLVED FIXED | ||
Severity: | normal | ||
Priority: | NOR | ||
Version First Reported In: | unspecified | ||
Target Milestone: | --- | ||
Platform: | Compiled Sources | ||
OS: | Solaris | ||
Latest Commit: | Version Fixed In: | ||
Sentry Crash Report: |
Description
Steve Evans
2006-04-12 18:56:19 UTC
Does pow do the same thing as powf? I've assigned this bug to Adrian Page, since he knows this code best. SVN commit 529238 by page: Use pow if powf is not available. BUG: 125446 M +5 -0 colorspaces/rgb_f16half/kis_rgb_f16half_colorspace.cc M +5 -0 colorspaces/rgb_f32/kis_rgb_f32_colorspace.cc M +7 -0 configure.in.in --- branches/koffice/1.5/koffice/krita/colorspaces/rgb_f16half/kis_rgb_f16half_colorspace.cc #529237:529238 @@ -38,6 +38,11 @@ #include "kis_integer_maths.h" +#ifndef HAVE_POWF +#undef powf +#define powf pow +#endif + //#define HALF_MAX ((half)1.0f) //temp #define EPSILON HALF_EPSILON --- branches/koffice/1.5/koffice/krita/colorspaces/rgb_f32/kis_rgb_f32_colorspace.cc #529237:529238 @@ -39,6 +39,11 @@ #include "kis_integer_maths.h" +#ifndef HAVE_POWF +#undef powf +#define powf pow +#endif + #define FLOAT_MAX 1.0f //temp #define EPSILON 1e-6 --- branches/koffice/1.5/koffice/krita/configure.in.in #529237:529238 @@ -99,3 +99,10 @@ # --- End of OpenGL check --- +# Check for powf. + +AC_CHECK_FUNC(powf, [have_powf="yes"], [AC_CHECK_LIB(m, powf, [have_powf="yes"], [have_powf="no"])]) + +if test "$have_powf" = 'yes'; then + AC_DEFINE([HAVE_POWF], 1, [Define to 1 if your system has powf in <maths.h>]) +fi This problem still exists in koffice 1.6.0 If the above patch didn't fix it, we really need a patch by someone who actually works with Solaris -- we're coding in the dark here, not having either solaris or solaris expertise. To clarify things, the above patch does work, but it is not included in koffice 1.6.0 Oh dear... And I wasn't paying attention, so it's too late to put it in for 1.6.1, too. I'll commit for 1.6.2, which will be released in January. I saw too late that Adrian had only committed to the 1.5 branch. I'm not sure how to port this to CMake and trunk, though. It's probably still a problem there, so reassigning to Adrian. And, of course, this bug is not unconfirmed. Checked in for 1.6, btw. the patch was applied to krita 1.6 by Adrian. And as far as I can see is in. So there is an other problem, can we have the compiler output ? The patch was added by Adrian in 1.5, not 1.6. It was added to 1.6 on Nov 19th which was too late for inclusion in koffice 1.6.1, as detailed in comment #6. The patch does solve the problem, so hopefully 1.6.2 will be OK. 1.6.3 has been released. Can this bug be closed now? I think so. |