def Picire(sut_call,
sut_call_kwargs,
listener,
ident,
issue,
work_dir,
parallel=False,
combine_loops=False,
split_method='zeller',
subset_first=True,
subset_iterator='forward',
complement_iterator='forward',
jobs=os.cpu_count(),
max_utilization=100,
encoding=None,
atom='both',
cache_class='ContentCache',
cleanup=True,
**kwargs):
"""
Test case reducer based on the Picire Parallel Delta Debugging Framework.
**Optional parameters of the reducer:**
- ``parallel``, ``combine_loops``, ``split_method``, ``subset_first``,
``subset_iterator``, ``complement_iterator``, ``jobs``,
``max_utilization``, ``encoding``, ``atom``, ``cache_class``,
``cleanup``
Refer to https://github.com/renatahodovan/picire for configuring Picire.
Note: This reducer is capable of detecting new issues found during the test
reduction (if any).
**Example configuration snippet:**
.. code-block:: ini
[sut.foo]
#call=...
cost=1
reduce=fuzzinator.reduce.Picire
reduce_cost=4
[sut.foo.reduce]
parallel=True
jobs=4
subset_iterator=skip
"""
def eval_arg(arg):
return eval(arg) if isinstance(arg, str) else arg
logging.getLogger('picire').setLevel(logger.level)
src = issue['test']
file_name = issue.get('filename', 'test')
parallel = eval_arg(parallel)
jobs = 1 if not parallel else eval_arg(jobs)
encoding = encoding or chardet.detect(src)['encoding']
cleanup = eval_arg(cleanup)
combine_loops = eval_arg(combine_loops)
subset_first = eval_arg(subset_first)
max_utilization = eval_arg(max_utilization)
split_method = getattr(picire.config_splitters, split_method)
subset_iterator = getattr(picire.config_iterators, subset_iterator)
complement_iterator = getattr(picire.config_iterators, complement_iterator)
cache_class = getattr(picire, cache_class)
if parallel:
cache_class = picire.shared_cache_decorator(cache_class)
# Choose the reducer class that will be used and its configuration.
reduce_config = {'split': split_method}
if not parallel:
reduce_class = picire.LightDD
reduce_config['subset_iterator'] = subset_iterator
reduce_config['complement_iterator'] = complement_iterator
reduce_config['subset_first'] = subset_first
else:
reduce_config['proc_num'] = jobs
reduce_config['max_utilization'] = max_utilization
if combine_loops:
reduce_class = picire.CombinedParallelDD
reduce_config['config_iterator'] = picire.CombinedIterator(
subset_first, subset_iterator, complement_iterator)
else:
reduce_class = picire.ParallelDD
reduce_config['subset_iterator'] = subset_iterator
reduce_config['complement_iterator'] = complement_iterator
reduce_config['subset_first'] = subset_first
issues = dict()
tester_config = dict(sut_call=sut_call,
sut_call_kwargs=sut_call_kwargs,
enc=encoding,
expected=issue['id'],
listener=listener,
ident=ident,
issues=issues)
call_config = dict(reduce_class=reduce_class,
reduce_config=reduce_config,
tester_class=PicireTester,
tester_config=tester_config,
input=file_name,
src=src,
encoding=encoding,
out=work_dir,
atom=atom,
cache_class=cache_class,
cleanup=cleanup)
try:
reduced_file = picire.call(**call_config)
except Exception as e:
logger.warning('Exception in picire', exc_info=e)
return None, list(issues.values())
with open(reduced_file, 'rb') as f:
src = f.read()
return src, list(issues.values())
def Picire(sut_call,
sut_call_kwargs,
listener,
ident,
issue,
work_dir,
parallel=False,
combine_loops=False,
split_method='zeller',
subset_first=True,
subset_iterator='forward',
complement_iterator='forward',
jobs=os.cpu_count(),
max_utilization=100,
encoding=None,
atom='line',
disable_cache=False,
cleanup=True,
**kwargs):
"""
Test case reducer based on the Picire Parallel Delta Debugging Framework.
Optional parameters of the reducer:
- 'parallel', 'combine_loops', 'split_method', 'subset_first',
'subset_iterator', 'complement_iterator', 'jobs', 'max_utilization',
'encoding', 'atom', 'disable_cache', 'cleanup'
Refer to https://github.com/renatahodovan/picire for configuring Picire.
Note: This reducer is capable of detecting new issues found during the test
reduction (if any).
Example configuration snippet:
[sut.foo]
#call=...
cost=1
reduce=fuzzinator.reduce.Picire
reduce_cost=4
[sut.foo.reduce]
parallel=True
jobs=4
subset_iterator=skip
"""
def eval_arg(arg):
return eval(arg) if isinstance(arg, str) else arg
logging.getLogger('picire').setLevel(logger.level)
src = issue['test']
file_name = issue.get('filename', 'test')
parallel = eval_arg(parallel)
jobs = 1 if not parallel else eval_arg(jobs)
disable_cache = eval_arg(disable_cache)
encoding = encoding or chardet.detect(src)['encoding']
cleanup = eval_arg(cleanup)
combine_loops = eval_arg(combine_loops)
subset_first = eval_arg(subset_first)
max_utilization = eval_arg(max_utilization)
split_method = getattr(picire.config_splitters, split_method)
subset_iterator = getattr(picire.config_iterators, subset_iterator)
complement_iterator = getattr(picire.config_iterators, complement_iterator)
# Choose the reducer class that will be used and its configuration.
reduce_config = {'split': split_method}
if not parallel:
reduce_class = picire.LightDD
reduce_config['subset_iterator'] = subset_iterator
reduce_config['complement_iterator'] = complement_iterator
reduce_config['subset_first'] = subset_first
else:
reduce_config['proc_num'] = jobs
reduce_config['max_utilization'] = max_utilization
if combine_loops:
reduce_class = picire.CombinedParallelDD
reduce_config['config_iterator'] = picire.CombinedIterator(
subset_first, subset_iterator, complement_iterator)
else:
reduce_class = picire.ParallelDD
reduce_config['subset_iterator'] = subset_iterator
reduce_config['complement_iterator'] = complement_iterator
reduce_config['subset_first'] = subset_first
issues = dict()
tester_config = dict(sut_call=sut_call,
sut_call_kwargs=sut_call_kwargs,
enc=encoding,
expected=issue['id'],
listener=listener,
ident=ident,
issues=issues)
call_config = dict(reduce_class=reduce_class,
reduce_config=reduce_config,
tester_class=PicireTester,
tester_config=tester_config,
input=file_name,
src=src,
encoding=encoding,
out=work_dir,
atom=atom,
parallel=parallel,
disable_cache=disable_cache,
cleanup=cleanup)
try:
reduced_file = picire.call(**call_config)
except:
return None, list(issues.values())
# Reduce by char if we used line based reduce earlier.
if atom == 'line':
with open(reduced_file, 'rb') as f:
src = f.read()
call_config.update(dict(atom='char', src=src))
try:
reduced_file = picire.call(**call_config)
except:
return None, list(issues.values())
with open(reduced_file, 'rb') as f:
src = f.read()
return src, list(issues.values())
