Hi Jakub, On Nov 30, 2010, at 08:48 PM, Jakub Wilk wrote:
>I'm not maintainer of this package, but... > >* Barry Warsaw <ba...@python.org>, 2010-11-29, 15:05: >>Package: cheetah >>Version: 2.4.3 > >There is no such version in Debian: >$ rmadison cheetah > cheetah | 2.0~rc7-1 | etch | source > cheetah | 2.0.1-2 | lenny | source > cheetah | 2.4.2.1-1 | squeeze | source > cheetah | 2.4.2.1-1 | sid | source Ah, right sorry about that. Ubuntu has 2.4.3 and I forgot to check the version in Debian when I submitted the bug report. Upstream released Cheetah 2.4.3 on 2010-09-16 so Debian should update, though that's probably not immediately necessary (i.e. maybe it can wait until Squeeze is released). http://pypi.python.org/pypi/Cheetah/2.4.3 But oddly enough, both the online change log and in-tarball change log only go up to 2.4.2: http://www.cheetahtemplate.org/CHANGES.txt I have no idea why that is. >>02_clean_modules.patch and 03_python-2.6-warning-fix.patch are no >longer >>necessary. > >The latest version of cheetah in Debian doesn't carry any of these patches >AFAICS. Cool. I'm betting they were necessary for 2.0.x but not 2.4.x. >>diff -uNr Cheetah-2.4.2.1.orig/cheetah/Template.py >>Cheetah-2.4.2.1/cheetah/Template.py >>--- Cheetah-2.4.2.1.orig/cheetah/Template.py 2010-02-07 23:17:23.000000000 >>-0500 >>+++ Cheetah-2.4.2.1/cheetah/Template.py 2010-10-18 14:06:54.439935114 >>-0400 >>@@ -32,7 +32,7 @@ >> > filetype = None >> >-if isinstance(sys.version_info, tuple): >>+if isinstance(sys.version_info[:], tuple): >> # Python 2.xx >> filetype = types.FileType >> def createMethod(func, cls): > >This hunk looks good, as sys.version_info is not a tuple in Python 2.7. (Don't >ask me why...) Heh, it became a namedtuple, presumably for readability: % python2.7 Python 2.7.0+ (r27:82500, Sep 15 2010, 18:14:55) [GCC 4.4.5] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> import sys >>> type(sys.version_info) <type 'sys.version_info'> >>> sys.version_info sys.version_info(major=2, minor=7, micro=0, releaselevel='final', serial=0) >>> sys.version_info.minor 7 >>> sys.version_info[1] 7 >>> type(sys.version_info[:]) <type 'tuple'> >>> Anyway, this is the fix suggested in https://github.com/rtyler/cheetah/issues/#issue/2 >>diff -uNr Cheetah-2.4.2.1.orig/cheetah/Tests/SyntaxAndOutput.py >>Cheetah-2.4.2.1/cheetah/Tests/SyntaxAndOutput.py >>--- Cheetah-2.4.2.1.orig/cheetah/Tests/SyntaxAndOutput.py 2010-10-18 >>15:00:44.668935001 -0400 >>+++ Cheetah-2.4.2.1/cheetah/Tests/SyntaxAndOutput.py 2010-10-18 >>14:43:56.500934997 -0400 >>@@ -847,17 +847,17 @@ >> def test10(self): >> r"""func placeholder - with ('''\nstring\n''')""" >> self.verify("$aFunc('''\naoeu\n''')", >>- "\naoeu\n") >>+ "\naoeu\n", convertEOLs=False) >> > def test11(self): >> r"""func placeholder - with ('''\nstring'\n''')""" >> self.verify("$aFunc('''\naoeu'\n''')", >>- "\naoeu'\n") >>+ "\naoeu'\n", convertEOLs=False) >> > def test12(self): >> r'''func placeholder - with ("""\nstring\n""")''' >> self.verify('$aFunc("""\naoeu\n""")', >>- "\naoeu\n") >>+ "\naoeu\n", convertEOLs=False) >> > def test13(self): >> """func placeholder - with (string*int)""" > >This one looks wrong. I believe that the tests were correct and the bug is >somewhere else. Given the upstream bug report, this doesn't seem like a totally unreasonable way to solve the bug across Python 2.6 and 2.7 for now. I haven't looked at the code though. Also note that this was a fairly straight rip from the Fedora patch. It's probably fine until upstream fixes it for real. Cheers, -Barry
signature.asc
Description: PGP signature
