[Piglit] piglit-summary-html.py currently broken
Vinson Lee
vlee at freedesktop.org
Tue Jun 17 12:06:11 PDT 2014
On Tue, Jun 17, 2014 at 4:44 AM, Dylan Baker <baker.dylan.c at gmail.com> wrote:
> On Tuesday, June 17, 2014 11:26:11 AM Thomas Wood wrote:
>> On 17 June 2014 00:51, Ian Romanick <idr at freedesktop.org> wrote:
>> > On 06/16/2014 02:19 PM, Dylan Baker wrote:
>> >> On Monday, June 16, 2014 12:32:35 PM Ian Romanick wrote:
>> >>> When I run piglit-summary-html.py, it crashes with the following trace:
>> >>>
>> >>> Traceback (most recent call last):
>> >>> File "./piglit-summary-html.py", line 28, in <module>
>> >>>
>> >>> html(sys.argv[1:])
>> >>>
>> >>> File "/home/idr/devel/graphics/piglit/framework/programs/summary.py",
>> >>> line
>> >>>
>> >>> 97, in html output.generate_html(args.summaryDir, args.exclude_details)
>> >>>
>> >>> File "/home/idr/devel/graphics/piglit/framework/summary.py", line 436,
>> >>> in
>> >>>
>> >>> generate_html module_directory=self.TEMP_DIR)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/template.py", line 224, in
>> >>>
>> >>> __init__ module = self._compile_from_file(path, filename)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/template.py", line 250, in
>> >>>
>> >>> _compile_from_file path)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/template.py", line 513, in
>> >>>
>> >>> _compile_module_file node = lexer.parse()
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/lexer.py", line 211, in
>> >>> parse
>> >>>
>> >>> if self.match_control_line():
>> >>> File "/usr/lib/python2.7/site-packages/mako/lexer.py", line 400, in
>> >>>
>> >>> match_control_line self.append_node(parsetree.ControlLine, keyword,
>> >>> isend,
>> >>> text)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/lexer.py", line 125, in
>> >>>
>> >>> append_node node = nodecls(*args, **kwargs)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/parsetree.py", line 72, in
>> >>>
>> >>> __init__ code = ast.PythonFragment(text, **self.exception_kwargs)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/ast.py", line 90, in
>> >>> __init__
>> >>>
>> >>> super(PythonFragment, self).__init__(code, **exception_kwargs)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/ast.py", line 39, in
>> >>> __init__
>> >>>
>> >>> f.visit(expr)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/_ast_util.py", line 279,
>> >>> in
>> >>>
>> >>> visit return self.generic_visit(node)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/_ast_util.py", line 287,
>> >>> in
>> >>>
>> >>> generic_visit self.visit(item)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/_ast_util.py", line 278,
>> >>> in
>> >>>
>> >>> visit return f(node)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/pyparser.py", line 143, in
>> >>>
>> >>> visit_For self.visit(node.iter)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/_ast_util.py", line 279,
>> >>> in
>> >>>
>> >>> visit return self.generic_visit(node)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/_ast_util.py", line 287,
>> >>> in
>> >>>
>> >>> generic_visit self.visit(item)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/_ast_util.py", line 279,
>> >>> in
>> >>>
>> >>> visit return self.generic_visit(node)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/_ast_util.py", line 289,
>> >>> in
>> >>>
>> >>> generic_visit self.visit(value)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/_ast_util.py", line 278,
>> >>> in
>> >>>
>> >>> visit return f(node)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/pyparser.py", line 107, in
>> >>>
>> >>> visit_Lambda self._visit_function(node, True)
>> >>>
>> >>> File "/usr/lib/python2.7/site-packages/mako/pyparser.py", line 125, in
>> >>>
>> >>> _visit_function if arg_id(arg) in self.local_ident_stack:
>> >>> AttributeError: 'Tuple' object has no attribute 'id'
>> >>>
>> >>> I bisected this to the following commit:
>> >>>
>> >>> commit 6755d9195bc05bbee2445eeac6af6b0e0ecf5160
>> >>> Author: Thomas Wood <thomas.wood at intel.com>
>> >>> Date: Wed Jun 4 11:07:30 2014 +0100
>> >>>
>> >>> templates: add the totals to testrun_info
>> >>>
>> >>> Add the totals to the testrun_info template, including the total
>> >>> number
>> >>> of tests evaluated.
>> >>>
>> >>> Signed-off-by: Thomas Wood <thomas.wood at intel.com>
>> >>> Reviewed-by: Dylan Baker <baker.dylan.c at gmail.com>
>> >>
>> >> I cannot replicate this bug, can you upload your results files somewhere
>> >> or
>> >> send them to me?
>> >
>> > http://people.freedesktop.org/~idr/10.2-12/
>> >
>> > I'm suspicious that it depends on some software versions. I've got:
>> >
>> > python-2.7.3-13.fc18.x86_64
>> > python-mako-0.5.0-5.fc18.noarch
>> >
>> > Are other things important?
>>
>> I recently encountered this problem on a machine using Mako 0.5 and
>> came across the following bug report:
>>
>> https://bitbucket.org/zzzeek/mako/issue/201/
>>
>> Installing a newer version of Mako (1.0.0) fixed the problem. Version
>> 0.7.3 available in Fedora 20 also doesn't have this issue.
>>
>> The bug report suggests the issue is present in Mako 0.7.2, so perhaps
>> the following version check should be added to summary.py:
>>
>> diff --git a/framework/summary.py b/framework/summary.py
>> index e5eb156..bf63777 100644
>> --- a/framework/summary.py
>> +++ b/framework/summary.py
>> @@ -29,8 +29,15 @@ import collections
>> import tempfile
>> import datetime
>> import re
>> +from distutils.version import LooseVersion
>> +import mako
>> from mako.template import Template
>>
>> +if LooseVersion(mako.__version__) < '0.7.3':
>> + print("Please install Mako >= 0.7.3")
>> + exit(1)
>> +
>> +
>> # a local variable status exists, prevent accidental overloading by
>> renaming # the module
>> import framework.status as so
>
> Being that mako 7.3 has a release date of 2012/10/07 I think it's a little bit
> silly to check for it (Mako 1.0.0 is current, but I haven't tested that it
> works with anything except 7.3 and 9.x). If we are going to have that we
> should put it in the cmake where we check for mako rather than in the script
> IMO.
> _______________________________________________
> Piglit mailing list
> Piglit at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/piglit
>
I disagree with a CMake change. Putting the check in CMake would
disable building and running any individual piglit test on older
distros.
More information about the Piglit
mailing list