Adds TestSuite(object).run() generic testing framework

[nthiery]

This patch implements TestSuite(object).run() which runs
systematic checks on the object. Here is a typical call:

     sage: TestSuite(ZZ).run(verbose = True)
     running ._test_an_element() ... done
     running ._test_element_pickling() ... done
     running ._test_not_implemented_methods() ... done
     running ._test_pickling() ... done

In practice, TestSuite(o).run() runs all the methods named test* of the object o.
The test* methods are typically implemented by abstract super classes
and in particular via categories, in order to enforce standard
behavior and API (_test_pickling, _test_an_element), or provide
mathematical sanity checks (_test_associativity).

For consistent error reporting, the test* methods in turn must use
the new gadget sage.misc.sage_unittest.InstanceTester to actually
run the tests.

This is used by the category patches #5891 and followers

CC: @sagetrac-sage-combinat @sagetrac-cwitty @roed314 @saliola @sagetrac-mvngu

Component: doctest coverage

Keywords: testunit

Author: Nicolas M. Thiéry

Reviewer: Mike Hansen, Minh Van Nguyen

Merged: Sage 4.1.2.alpha2

Issue created by migration from https://trac.sagemath.org/ticket/6343

[nthiery]

Description changed:

--- 
+++ 
@@ -1,4 +1,4 @@
-This patch adds a .check() method in SageObject which runs systematic checks on the object. Here is a typical call:
+This patch implements a .check() method in SageObject which runs systematic checks on the object. Here is a typical call:
 
      sage: ZZ.check(verbose = True)
      running test_an_element ... done
@@ -10,6 +10,6 @@
 
 The test_* methods are typically implemented by abstract super classes and in particular via categories, in order to enforce standard behavior and API (test_pickling, test_an_element), or provide mathematical sanity checks (test_associativity).
 
-For consistent error reporting, the test_* methods in turn must use the gadget sage.misc.instance_tester.InstanceTester to actually run the tests.
+For consistent error reporting, the test_* methods in turn must use the new gadget sage.misc.instance_tester.InstanceTester to actually run the tests.
 
 This patch is a prerequisite for #5891 (category framework), which uses it intensively. Some examples in the patch illustrate this, and are therefore temporarily disabled.

[williamstein]

comment:2

Change the name from obj.check() to obj._check(). It is not reasonable that if one does obj. on any Sage object, one sees check.

William

[williamstein]

comment:4

Change the name from obj.check() to obj._check(). It is not reasonable that if one does obj. on any Sage object, one sees check.

[nthiery]

Changed keywords from none to testunit

[nthiery]

Description changed:

--- 
+++ 
@@ -1,15 +1,21 @@
-This patch implements a .check() method in SageObject which runs systematic checks on the object. Here is a typical call:
+This patch implements TestSuite(object).run()  which runs
+systematic checks on the object. Here is a typical call:
 
-     sage: ZZ.check(verbose = True)
-     running test_an_element ... done
-     running test_element_pickling ... done
-     running test_not_implemented_methods ... done
-     running test_pickling ... done
+     sage: TestSuite(ZZ).run(verbose = True)
+     running ._test_an_element() ... done
+     running ._test_element_pickling() ... done
+     running ._test_not_implemented_methods() ... done
+     running ._test_pickling() ... done
 
-In practice, o.check() runs all the methods named test_* of the object o.
+In practice, TestSuite(o).run() runs all the methods named _test_* of the object o.
+The _test_* methods are typically implemented by abstract super classes
+and in particular via categories, in order to enforce standard
+behavior and API (_test_pickling, _test_an_element), or provide
+mathematical sanity checks (_test_associativity).
 
-The test_* methods are typically implemented by abstract super classes and in particular via categories, in order to enforce standard behavior and API (test_pickling, test_an_element), or provide mathematical sanity checks (test_associativity).
+For consistent error reporting, the _test_* methods in turn must use
+the new gadget sage.misc.sage_unittest.InstanceTester to actually
+run the tests.
 
-For consistent error reporting, the test_* methods in turn must use the new gadget sage.misc.instance_tester.InstanceTester to actually run the tests.
+This is used by the category patches #5891 and followers
 
-This patch is a prerequisite for #5891 (category framework), which uses it intensively. Some examples in the patch illustrate this, and are therefore temporarily disabled.