Get Pynac functions to work on FreeBSD

[kcrisman]

This could just be related to the [ticket:9543 Cephes on FreeBSD ticket]. But it seems that using slightly different functions in one of the Pynac files helps on FreeBSD. As reported by Stephen Montgomery-Smith:

--- sage-4.8/sage/symbolic/pynac_cc.h.orig 2010-06-28 16:37:01.000000000 +0000
+++ sage-4.8/sage/symbolic/pynac_cc.h 2012-01-21 22:48:22.000000000 +0000
@@ -1,37 +1,21 @@
-#include <math.h>
+#include <cmath>
 
 inline long double sage_logl(long double x) 
 {
-#if defined(__CYGWIN__)
   return log(x);
-#else
-  return logl(x);
-#endif 
 }
 
 inline long double sage_sqrtl(long double x) 
 {
-#if defined(__CYGWIN__)
   return sqrt(x);
-#else
-  return sqrtl(x);
-#endif 
 } 
 
 inline long double sage_tgammal(long double x) 
 {
-#if defined(__CYGWIN__)
   return tgamma(x);
-#else
-  return tgammal(x);
-#endif 
 }
 
 inline long double sage_lgammal(long double x)
 {
-#if defined(__CYGWIN__)
   return lgamma(x);
-#else
-  return lgammal(x);
-#endif 
 }

Depends on #9543

CC: @peterjeremy

Component: porting: BSD

Reviewer: Stephen Montgomery-Smith

Issue created by migration from https://trac.sagemath.org/ticket/12401

[sagetrac-stephen]

Attachment: spkg-patch-sage-5.0.beta12_-_sage_symbolic_pynac_cc.h.gz

Updated patch for pynac_cc.h

[sagetrac-stephen]

comment:1

Please use the updated patch instead. It should be possible to add this patch without interfering with builds from other OS.

--- sage-5.0.beta12/sage/symbolic/pynac_cc.h-orig	2012-04-07 22:52:53.000000000 +0000
+++ sage-5.0.beta12/sage/symbolic/pynac_cc.h	2012-04-07 22:53:27.000000000 +0000
@@ -2,7 +2,7 @@
 
 inline long double sage_logl(long double x) 
 {
-#if defined(__CYGWIN__)
+#if defined(__CYGWIN__) || defined(__FreeBSD__)
   return log(x);
 #else
   return logl(x);
@@ -11,7 +11,7 @@
 
 inline long double sage_sqrtl(long double x) 
 {
-#if defined(__CYGWIN__)
+#if defined(__CYGWIN__) || defined(__FreeBSD__)
   return sqrt(x);
 #else
   return sqrtl(x);
@@ -20,7 +20,7 @@
 
 inline long double sage_tgammal(long double x) 
 {
-#if defined(__CYGWIN__)
+#if defined(__CYGWIN__) || defined(__FreeBSD__)
   return tgamma(x);
 #else
   return tgammal(x);
@@ -29,7 +29,7 @@
 
 inline long double sage_lgammal(long double x)
 {
-#if defined(__CYGWIN__)
+#if defined(__CYGWIN__) || defined(__FreeBSD__)
   return lgamma(x);
 #else
   return lgammal(x);

[sagetrac-stephen]

comment:2

The changes reported here do not need to be made if the following trac is committed:

#9543

[kcrisman]

Reviewer: Stephen Montgomery-Smith

[kcrisman]

Dependencies: #9543

[kcrisman]

comment:3

Okay, this sounds almost like a positive review. Once #9543 is in, we can probably close this, then.

[kcrisman]

comment:4

Well, #9543 is now in. Stephen, when you end up building Sage 5.6 (whenever that comes out), a quick switch to "positive review" is all that would be needed here.

[sagetrac-stephen]

comment:5

I have already checked it out using sage 5.6.rc0. Is it OK if I set it to "positive review" now?

[kcrisman]

comment:6

I have already checked it out using sage 5.6.rc0. Is it OK if I set it to "positive review" now?

That was what I meant :-) so I'll do it now. Thanks so much for your continuing help.