minor fix and update to test code

svn: r9345

ChangeLog

@@ -1,3 +1,11 @@
+2007-11-13 Jim Sack <jgsack@san.rr.com>
+	* src/test/regrtest.py: undo removal of module imports in scan-for-tests
+	* src/GrampsDbUtils/test/_GrampsDbWRFactories_test.py: minor improve.
+	* src/GrampsDbUtils/test/_GedcomChar_test.py: add tests for utf8
+	The regrtest had a unwise "optimization" that actually broke the
+	..Factories test. A comment was left in place where the un-doing was
+	done because it seems like a good thing to document.
+
 2007-11-13 Martin Hawlisch  <Martin.Hawlisch@gmx.de>
 	* src/TipOfDay.py: Catch exceptions when parsing the tips file.
 

src/GrampsDbUtils/test/_GedcomChar_test.py

@@ -32,8 +32,11 @@ def gen_chars(filename, encoding):
     if not os.path.exists(filename):
         codecs.open(filename, "wb", encoding).write(utest_chars)
 
+###
 class Test1_ansi(unittest.TestCase):
+    """Test original "ANSI" reader and codecs: latin, cp1252"""
     enc = "latin-1"
+    cp = "cp1252"
     fil = os.path.join(cdir,enc)
     exp  = utest_chars
     
@@ -50,7 +53,13 @@ class Test1_ansi(unittest.TestCase):
         got=codecs.open(s.fil, encoding=s.enc).read()
         s.assertEquals(got,s.exp, m(got,s.exp, "using codec %s" % s.enc))
 
+    def test1c_read_codec_cp1252(s):
+        got=codecs.open(s.fil, encoding=s.cp).read()
+        s.assertEquals(got,s.exp, m(got,s.exp, "using codec %s" % s.cp))
+
+###
 class Test2_ansel(unittest.TestCase):
+    """Test original AnselReader (later: ansel codec)"""
     enc = "ansel"
     afil = os.path.join(cdir,enc)
     exp  = a2u
@@ -64,6 +73,90 @@ class Test2_ansel(unittest.TestCase):
         got = ra.readline()
         s.assertEquals(got,s.exp, m(got,s.exp, "AnselReader"))
 
+###
+class Test3(unittest.TestCase):
+    """Test otriginal UTF8Reader and codecs: utf_8, utf_8_sig 
+    with no 'BOM' (sig) in input (the common case) 
+
+    out of curiosity, verify behavior reading a 1-byte file
+    """
+    enc = "utf_8"
+    enc_sig = enc + "_sig"
+    ufil = os.path.join(cdir, "chars.utf8")
+    f1byte = os.path.join(cdir, "1byte")
+    exp  = utest_chars
+    
+    def setUp(s):
+        gen_chars(s.ufil, s.enc)
+        if not os.path.exists(s.f1byte):
+            open(s.f1byte, "wb").write("1")
+ 
+    def test3a_u8_UTF8Reader_NO_BOM_sig(s):
+        f=open(s.ufil)
+        ra=G.UTF8Reader(f)
+        g = ra.readline()
+        s.assertEquals(g,s.exp, m(g,s.exp, "orig UTF8Reader"))
+        r2 = G.UTF8Reader(open(s.f1byte))
+        g = r2.readline()
+        s.assertEquals(g,"1", 
+            m(g,"1", "read 1-byte file"))
+  
+    # NB: utf_8 reads data and never expects a BOM-sig
+    def test3b_utf8_codec_NO_BOM_sig_as_expected(s):
+        g=codecs.open(s.ufil, encoding=s.enc).read()
+        s.assertEquals(g,s.exp, m(g,s.exp, "codec utf8"))
+        g=codecs.open(s.f1byte, encoding=s.enc).read()
+        s.assertEquals(g,"1", m(g,"1", "codec utf8"))
+  
+    # NB: utf_8_sig reads data even absent a BOM-sig (GOOD!)
+    def test3c_utf8_sig_codec_NO_BOM_sig_tolerated_GOOD(s):
+        g=codecs.open(s.ufil, encoding=s.enc_sig).read()
+        s.assertEquals(g,s.exp, 
+            m(g,s.exp, "codec utf_8_sig NO sig input"))
+        g=codecs.open(s.f1byte, encoding=s.enc_sig).read()
+        s.assertEquals(g,"1", 
+            m(g,"1", "codec utf_8_sig NO sig input"))
+
+###
+class Test4(unittest.TestCase):
+    """Test otriginal UTF8Reader and codecs: utf_8, utf_8_sig
+    with 'BOM' (sig) in input (uncommon, [some?] MS Windows only?) 
+    """
+    enc = "utf_8"
+    enc_sig = enc + "_sig"
+    ufil = os.path.join(cdir, "chars.utf8_sig")
+    exp  = utest_chars
+    
+    def setUp(s):
+        gen_chars(s.ufil, s.enc_sig)
+ 
+    def test4a_u8_UTF8Reader_WITH_BOM_sig(s):
+        f=open(s.ufil)
+        ra=G.UTF8Reader(f)
+        g = ra.readline()
+        s.assertEquals(g,s.exp, m(g,s.exp, "orig UTF8Reader"))
+   
+    # utf_8 reads an initial BOM-sig as data -- oops, pity
+    #  write the test to verify this known codec behavior 
+    # ==> Recommend: do not use utf8 as input codec (use utf_8_sig)
+    def test4b_utf8_codec_WITH_BOM_sig_reads_as_data_PITY(s):
+        g=codecs.open(s.ufil, encoding=s.enc).read()
+        e0=u'\ufeff'
+        s.assertEquals(g[0], e0, 
+            m(g[0],e0, "codec utf8 reads 'BOM'-sig as data" ))
+        g = g[1:]
+        s.assertEquals(g,s.exp, 
+            m(g,s.exp, "codec utf8 reads rest of data ok"))
+  
+    # utf_8_sig reads and ignores the BOM-sig
+    def test4c_utf8_sig_codec_WITH_BOM_sig_as_expected(s):
+        g=codecs.open(s.ufil, encoding=s.enc_sig).read()
+        s.assertEquals(g,s.exp, 
+            m(g,s.exp, "codec utf_8_sig NO sig input"))
+
+###
+
+
 
 if __name__ == "__main__":
     unittest.main()

src/GrampsDbUtils/test/_GrampsDbWRFactories_test.py

@@ -2,7 +2,7 @@
 import unittest
 
 from test import test_util as tu
-tu.path_append_parent()
+par = tu.path_append_parent()
 
 ###
 class Test1(unittest.TestCase):
@@ -13,6 +13,9 @@ class Test1(unittest.TestCase):
 
     """
     def test1a_buried_imports(s):
+        import sys
+        s.assertTrue(par in sys.path, 
+            "par %r has to be in path!" % par)
         ilist = (
             "_WriteGrdb",
             "_WriteXML",

src/test/regrtest.py

@@ -80,7 +80,8 @@ def getTestSuites(loc=gramps_root):
     paths = [(path,files) for path,dirs,files in os.walk(loc) \
         if test_mod(path,dirs) and match_mod(files)] 
 
-    oldpath = list(sys.path)
+    ## NO -- see explanation below
+    ##  oldpath = list(sys.path)
     for (dir,test_modules) in paths:
         sys.path.append(dir)
 
@@ -96,8 +97,16 @@ def getTestSuites(loc=gramps_root):
                 perf_suites.append(mod.perfSuite())
             except:
                 pass
-        # remove temporary paths added
+        # NO: was: remove temporary paths added 
-        sys.path = list(oldpath)
+        # this seems like it should be reasonable,
+        # but it causes failure in _GrampsDbWRFactories_test.py
+        #  (I suspect it is an actual bug in the runner
+        #   but the easiest fix is to keep the imports,
+        #   which is what other loaders seem to do)
+        # ==>  this aspect of test frameworks is *hard*
+        ## NO -- do NOT:
+        ## remove temporary paths added
+        ## sys.path = list(oldpath)
     return (test_suites,perf_suites)