devel/py-numba: Update 0.47.0 -> 0.51.2

PR:		250937
Approved by:	dave@dal.ca (maintainer's timeout 21 days)

devel/py-numba/Makefile

@@ -2,8 +2,7 @@
 # $FreeBSD$
 
 PORTNAME=	numba
-DISTVERSION=	0.47.0
-PORTREVISION=	3
+DISTVERSION=	0.51.2
 CATEGORIES=	devel python
 MASTER_SITES=	CHEESESHOP
 PKGNAMEPREFIX=	${PYTHON_PKGNAMEPREFIX}

devel/py-numba/distinfo

@@ -1,3 +1,3 @@
-TIMESTAMP = 1578995274
-SHA256 (numba-0.47.0.tar.gz) = c0703df0a0ea2e29fbef7937d9849cc4734253066cb5820c5d6e0851876e3b0a
-SIZE (numba-0.47.0.tar.gz) = 1935290
+TIMESTAMP = 1604791415
+SHA256 (numba-0.51.2.tar.gz) = 16bd59572114adbf5f600ea383880d7b2071ae45477e84a24994e089ea390768
+SIZE (numba-0.51.2.tar.gz) = 2059680

devel/py-numba/files/patch-numba_np_ufunc_workqueue.c

@@ -0,0 +1,15 @@
+--- numba/np/ufunc/workqueue.c.orig	2020-11-07 23:26:52 UTC
++++ numba/np/ufunc/workqueue.c
+@@ -27,7 +27,11 @@ race conditions.
+ /* PThread */
+ #include <pthread.h>
+ #include <unistd.h>
+-#include <alloca.h>
++#if defined(__FreeBSD__)
++#  include <stdlib.h>
++#else
++#  include <alloca.h>
++#endif
+ #include <sys/types.h>
+ #include <unistd.h>
+ #include <signal.h>

devel/py-numba/files/patch-numba_npyufunc_workqueue.c

@@ -1,14 +0,0 @@
---- numba/npyufunc/workqueue.c.orig	2018-11-14 22:14:35 UTC
-+++ numba/npyufunc/workqueue.c
-@@ -19,7 +19,11 @@ race condition.
- /* PThread */
- #include <pthread.h>
- #include <unistd.h>
-+#if defined(__FreeBSD__)
-+#include <stdlib.h>
-+#else
- #include <alloca.h>
-+#endif
- #define NUMBA_PTHREAD
- #endif
- 

devel/py-numba/pkg-descr

@@ -1,7 +1,7 @@
-Numba gives you the power to speed up your applications with high performance 
-functions written directly in Python. With a few annotations, array-oriented 
-and math-heavy Python code can be just-in-time compiled to native machine 
-instructions, similar in performance to C, C++ and Fortran, without having to 
+Numba gives you the power to speed up your applications with high performance
+functions written directly in Python. With a few annotations, array-oriented
+and math-heavy Python code can be just-in-time compiled to native machine
+instructions, similar in performance to C, C++ and Fortran, without having to
 switch languages or Python interpreters.
 
 WWW: https://numba.pydata.org/