Upgrade and clean up sqlite package

[jdemeyer]

spkg: ​http://boxen.math.washington.edu/home/jdemeyer/spkg/sqlite-3.7.17.spkg (spkg diff​)

apply: 14629_sqlite.patch​ to the Sage library.

sqlite-3.7.17 (Jeroen Demeyer, 24 May 2013)

• #14629: upgrade to version 3.7.17
• Use standard template for SPKG.txt and spkg-install, use $MAKE instead of make
• Remove Fortran and C++ stuff from spkg-install
• Compile with -DSQLITE_WITHOUT_ZONEMALLOC on OS X <= 10.4

[jdemeyer]

sage -t devel/sage/sage/databases/cremona.py
**********************************************************************
File "devel/sage/sage/databases/cremona.py", line 636, in sage.databases.cremona.MiniCremonaDatabase.__iter__
Failed example:
    it.next().label()
Expected:
    '11a1'
Got:
    '1001a1'
**********************************************************************
File "devel/sage/sage/databases/cremona.py", line 638, in sage.databases.cremona.MiniCremonaDatabase.__iter__
Failed example:
    it.next().label()
Expected:
    '11a2'
Got:
    '1001b1'
**********************************************************************
File "devel/sage/sage/databases/cremona.py", line 640, in sage.databases.cremona.MiniCremonaDatabase.__iter__
Failed example:
    it.next().label()
Expected:
    '11a3'
Got:
    '1001b2'
**********************************************************************
File "devel/sage/sage/databases/cremona.py", line 642, in sage.databases.cremona.MiniCremonaDatabase.__iter__
Failed example:
    it.next().label()
Expected:
    '14a1'
Got:
    '1001b3'
**********************************************************************
File "devel/sage/sage/databases/cremona.py", line 645, in sage.databases.cremona.MiniCremonaDatabase.__iter__
Failed example:
    it.next().label()
Expected:
    '45a3'
Got:
    '1014g2'
**********************************************************************

[fbissey]

Funny that. We did face the upgrade in sage-on-gentoo in sage 5.8 and we had the same issue. Christopher was supposed to open a ticket with his solution but he never did.

--- sage/databases/cremona.py.orig	2013-02-19 10:33:56.000000000 +0000
+++ sage/databases/cremona.py	2013-03-18 16:51:58.950097433 +0000
@@ -634,19 +634,19 @@
 
             sage: it = CremonaDatabase().__iter__()
             sage: it.next().label()
-            '11a1'
+            '1001a1'
             sage: it.next().label()
-            '11a2'
+            '1001b1'
             sage: it.next().label()
-            '11a3'
+            '1001b2'
             sage: it.next().label()
-            '14a1'
+            '1001b3'
             sage: skip = [it.next() for _ in range(100)]
             sage: it.next().label()
-            '45a3'
+            '1014g2'
         """
         for c in self.__connection__.cursor().execute('SELECT curve FROM ' \
-            + 't_curve'):
+            + 't_curve ORDER BY curve'):
             yield self.elliptic_curve(c[0])
 
     def __getitem__(self, N):

Anyway, your solution looks good too. One concern we had: is it fast?

[jdemeyer]

Replying to fbissey:

One concern we had: is it fast?

I have no idea, but I assume that SQLite is sufficiently fast. There are other place too where ORDER BY is used. In any case, your patch is not good because it assumes that there is no curve of conductor 10001 for example. This assumption is false if the large Cremona database has been installed.

[fbissey]

Roger that, of course we don't currently package the large Cremona database so it was easy to miss. Looks all good to me, positive review.