Add some patches to the PARI sources

[jdemeyer]

1. GCC 4.6.3 miscompiles src/language/init.c on OS X 10.4 PPC (32-bit), because of a problem with pointer arithmetic in stack_base() called in pari_init_functions(). There is an easy work-around by reordering some code in the PARI sources. Related to Add a gcc package #12369.

This is upstream bug http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49330

---

2. PARI's ispower() function gives wrong results when applied to negative numbers. Related to perfect_power for integers #12116.

This is fixed upstream http://pari.math.u-bordeaux.fr/cgi-bin/bugreport.cgi?bug=1302

---

3. issquarefree(0) raises an error instead of returning 0. See 0.is_squarefree() raises a PariError #12636.

This is fixed upstream http://pari.math.u-bordeaux.fr/cgi-bin/bugreport.cgi?bug=1304

---

spkg: http://boxen.math.washington.edu/home/jdemeyer/spkg/pari-2.5.1.p3.spkg

Upstream: Reported upstream. Developers acknowledge bug.

Component: packages: standard

Author: Jeroen Demeyer

Reviewer: Johan Bosman

Merged: sage-5.0.beta12

Issue created by migration from https://trac.sagemath.org/ticket/12638

[jdemeyer]

Description changed:

--- 
+++ 
@@ -4,4 +4,4 @@
 
 There is an easy work-around by reordering some code.
 
-**spkg**: []
+**spkg**: [http://boxen.math.washington.edu/home/jdemeyer/spkg/pari-2.5.1.p1.spkg](http://boxen.math.washington.edu/home/jdemeyer/spkg/pari-2.5.1.p1.spkg)

[jdemeyer]

comment:2

Built successfully with SAGE_CHECK=yes using gcc-4.6.3 on OS X 10.4 PPC, a system where it failed before.

[jdemeyer]

Description changed:

--- 
+++ 
@@ -1,7 +1,15 @@
-GCC 4.6.3 miscompiles `src/language/init.c` on OS X 10.4 PPC (32-bit), because of a problem with pointer arithmetic in `stack_base()` called in `pari_init_functions()`.
+1. GCC 4.6.3 miscompiles `src/language/init.c` on OS X 10.4 PPC (32-bit), because of a problem with pointer arithmetic in `stack_base()` called in `pari_init_functions()`.
 
 This is upstream bug [http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49330](http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49330)
 
-There is an easy work-around by reordering some code.
+There is an easy work-around by reordering some code in the PARI sources.
+
+---
+
+2. PARI's `ispower()` function gives wrong results when applied to negative numbers.
+
+This is fixed upstream [http://pari.math.u-bordeaux.fr/cgi-bin/bugreport.cgi?bug=1302](http://pari.math.u-bordeaux.fr/cgi-bin/bugreport.cgi?bug=1302)
+
+---
 
 **spkg**: [http://boxen.math.washington.edu/home/jdemeyer/spkg/pari-2.5.1.p1.spkg](http://boxen.math.washington.edu/home/jdemeyer/spkg/pari-2.5.1.p1.spkg)

[jdemeyer]

Description changed:

--- 
+++ 
@@ -12,4 +12,4 @@
 
 ---
 
-**spkg**: [http://boxen.math.washington.edu/home/jdemeyer/spkg/pari-2.5.1.p1.spkg](http://boxen.math.washington.edu/home/jdemeyer/spkg/pari-2.5.1.p1.spkg)
+**spkg**: [http://boxen.math.washington.edu/home/jdemeyer/spkg/pari-2.5.1.p2.spkg](http://boxen.math.washington.edu/home/jdemeyer/spkg/pari-2.5.1.p2.spkg)

[jdemeyer]

Description changed:

--- 
+++ 
@@ -1,12 +1,10 @@
-1. GCC 4.6.3 miscompiles `src/language/init.c` on OS X 10.4 PPC (32-bit), because of a problem with pointer arithmetic in `stack_base()` called in `pari_init_functions()`.
+1. GCC 4.6.3 miscompiles `src/language/init.c` on OS X 10.4 PPC (32-bit), because of a problem with pointer arithmetic in `stack_base()` called in `pari_init_functions()`.  There is an easy work-around by reordering some code in the PARI sources.  Related to #12369.
 
 This is upstream bug [http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49330](http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49330)
 
-There is an easy work-around by reordering some code in the PARI sources.
-
 ---
 
-2. PARI's `ispower()` function gives wrong results when applied to negative numbers.
+2. PARI's `ispower()` function gives wrong results when applied to negative numbers.  Related to #12116.
 
 This is fixed upstream [http://pari.math.u-bordeaux.fr/cgi-bin/bugreport.cgi?bug=1302](http://pari.math.u-bordeaux.fr/cgi-bin/bugreport.cgi?bug=1302)