def test_tree_matcher(self):
file1 = self.make_file("sub/file1.py")
file2 = self.make_file("sub/file2.c")
file3 = self.make_file("sub2/file3.h")
file4 = self.make_file("sub3/file4.py")
file5 = self.make_file("sub3/file5.c")
fl = FileLocator()
tm = TreeMatcher([
fl.canonical_filename("sub"),
fl.canonical_filename(file4),
])
self.assertTrue(tm.match(fl.canonical_filename(file1)))
self.assertTrue(tm.match(fl.canonical_filename(file2)))
self.assertFalse(tm.match(fl.canonical_filename(file3)))
self.assertTrue(tm.match(fl.canonical_filename(file4)))
self.assertFalse(tm.match(fl.canonical_filename(file5)))
def test_tree_matcher(self):
file1 = self.make_file("sub/file1.py")
file2 = self.make_file("sub/file2.c")
file3 = self.make_file("sub2/file3.h")
file4 = self.make_file("sub3/file4.py")
file5 = self.make_file("sub3/file5.c")
fl = FileLocator()
tm = TreeMatcher([
fl.canonical_filename("sub"),
fl.canonical_filename(file4),
])
self.assertTrue(tm.match(fl.canonical_filename(file1)))
self.assertTrue(tm.match(fl.canonical_filename(file2)))
self.assertFalse(tm.match(fl.canonical_filename(file3)))
self.assertTrue(tm.match(fl.canonical_filename(file4)))
self.assertFalse(tm.match(fl.canonical_filename(file5)))
class Coverage(object):
"""Programmatic access to coverage.py.
To use::
from coverage import Coverage
cov = Coverage()
cov.start()
#.. call your code ..
cov.stop()
cov.html_report(directory='covhtml')
"""
def __init__(
self, data_file=None, data_suffix=None, cover_pylib=None,
auto_data=False, timid=None, branch=None, config_file=True,
source=None, omit=None, include=None, debug=None,
concurrency=None,
):
"""
`data_file` is the base name of the data file to use, defaulting to
".coverage". `data_suffix` is appended (with a dot) to `data_file` to
create the final file name. If `data_suffix` is simply True, then a
suffix is created with the machine and process identity included.
`cover_pylib` is a boolean determining whether Python code installed
with the Python interpreter is measured. This includes the Python
standard library and any packages installed with the interpreter.
If `auto_data` is true, then any existing data file will be read when
coverage measurement starts, and data will be saved automatically when
measurement stops.
If `timid` is true, then a slower and simpler trace function will be
used. This is important for some environments where manipulation of
tracing functions breaks the faster trace function.
If `branch` is true, then branch coverage will be measured in addition
to the usual statement coverage.
`config_file` determines what configuration file to read:
* If it is ".coveragerc", it is interpreted as if it were True,
for backward compatibility.
* If it is a string, it is the name of the file to read. If the
file can't be read, it is an error.
* If it is True, then a few standard files names are tried
(".coveragerc", "setup.cfg"). It is not an error for these files
to not be found.
* If it is False, then no configuration file is read.
`source` is a list of file paths or package names. Only code located
in the trees indicated by the file paths or package names will be
measured.
`include` and `omit` are lists of file name patterns. Files that match
`include` will be measured, files that match `omit` will not. Each
will also accept a single string argument.
`debug` is a list of strings indicating what debugging information is
desired.
`concurrency` is a string indicating the concurrency library being used
in the measured code. Without this, coverage.py will get incorrect
results. Valid strings are "greenlet", "eventlet", "gevent",
"multiprocessing", or "thread" (the default).
.. versionadded:: 4.0
The `concurrency` parameter.
"""
# Build our configuration from a number of sources:
# 1: defaults:
self.config = CoverageConfig()
# 2: from the rcfile, .coveragerc or setup.cfg file:
if config_file:
did_read_rc = False
# Some API users were specifying ".coveragerc" to mean the same as
# True, so make it so.
if config_file == ".coveragerc":
config_file = True
specified_file = (config_file is not True)
if not specified_file:
config_file = ".coveragerc"
did_read_rc = self.config.from_file(config_file)
if not did_read_rc:
if specified_file:
raise CoverageException(
"Couldn't read '%s' as a config file" % config_file
)
self.config.from_file("setup.cfg", section_prefix="coverage:")
# 3: from environment variables:
env_data_file = os.environ.get('COVERAGE_FILE')
if env_data_file:
self.config.data_file = env_data_file
debugs = os.environ.get('COVERAGE_DEBUG')
if debugs:
self.config.debug.extend(debugs.split(","))
# 4: from constructor arguments:
self.config.from_args(
data_file=data_file, cover_pylib=cover_pylib, timid=timid,
branch=branch, parallel=bool_or_none(data_suffix),
source=source, omit=omit, include=include, debug=debug,
concurrency=concurrency,
)
self._debug_file = None
self._auto_data = auto_data
self._data_suffix = data_suffix
# The matchers for _should_trace.
self.source_match = None
self.source_pkgs_match = None
self.pylib_match = self.cover_match = None
self.include_match = self.omit_match = None
# Is it ok for no data to be collected?
self._warn_no_data = True
self._warn_unimported_source = True
# A record of all the warnings that have been issued.
self._warnings = []
# Other instance attributes, set later.
self.omit = self.include = self.source = None
self.source_pkgs = None
self.data = self.data_files = self.collector = None
self.plugins = None
self.pylib_dirs = self.cover_dirs = None
self.data_suffix = self.run_suffix = None
self._exclude_re = None
self.debug = None
# State machine variables:
# Have we initialized everything?
self._inited = False
# Have we started collecting and not stopped it?
self._started = False
# Have we measured some data and not harvested it?
self._measured = False
def _init(self):
"""Set all the initial state.
This is called by the public methods to initialize state. This lets us
construct a :class:`Coverage` object, then tweak its state before this
function is called.
"""
if self._inited:
return
# Create and configure the debugging controller. COVERAGE_DEBUG_FILE
# is an environment variable, the name of a file to append debug logs
# to.
if self._debug_file is None:
debug_file_name = os.environ.get("COVERAGE_DEBUG_FILE")
if debug_file_name:
self._debug_file = open(debug_file_name, "a")
else:
self._debug_file = sys.stderr
self.debug = DebugControl(self.config.debug, self._debug_file)
# Load plugins
self.plugins = Plugins.load_plugins(self.config.plugins, self.config, self.debug)
# _exclude_re is a dict that maps exclusion list names to compiled
# regexes.
self._exclude_re = {}
self._exclude_regex_stale()
files.set_relative_directory()
# The source argument can be directories or package names.
self.source = []
self.source_pkgs = []
for src in self.config.source or []:
if os.path.exists(src):
self.source.append(files.canonical_filename(src))
else:
self.source_pkgs.append(src)
self.omit = prep_patterns(self.config.omit)
self.include = prep_patterns(self.config.include)
concurrency = self.config.concurrency
if concurrency == "multiprocessing":
patch_multiprocessing()
concurrency = None
self.collector = Collector(
should_trace=self._should_trace,
check_include=self._check_include_omit_etc,
timid=self.config.timid,
branch=self.config.branch,
warn=self._warn,
concurrency=concurrency,
)
# Early warning if we aren't going to be able to support plugins.
if self.plugins.file_tracers and not self.collector.supports_plugins:
self._warn(
"Plugin file tracers (%s) aren't supported with %s" % (
", ".join(
plugin._coverage_plugin_name
for plugin in self.plugins.file_tracers
),
self.collector.tracer_name(),
)
)
for plugin in self.plugins.file_tracers:
plugin._coverage_enabled = False
# Suffixes are a bit tricky. We want to use the data suffix only when
# collecting data, not when combining data. So we save it as
# `self.run_suffix` now, and promote it to `self.data_suffix` if we
# find that we are collecting data later.
if self._data_suffix or self.config.parallel:
if not isinstance(self._data_suffix, string_class):
# if data_suffix=True, use .machinename.pid.random
self._data_suffix = True
else:
self._data_suffix = None
self.data_suffix = None
self.run_suffix = self._data_suffix
# Create the data file. We do this at construction time so that the
# data file will be written into the directory where the process
# started rather than wherever the process eventually chdir'd to.
self.data = CoverageData(debug=self.debug)
self.data_files = CoverageDataFiles(basename=self.config.data_file, warn=self._warn)
# The directories for files considered "installed with the interpreter".
self.pylib_dirs = set()
if not self.config.cover_pylib:
# Look at where some standard modules are located. That's the
# indication for "installed with the interpreter". In some
# environments (virtualenv, for example), these modules may be
# spread across a few locations. Look at all the candidate modules
# we've imported, and take all the different ones.
for m in (atexit, inspect, os, platform, re, _structseq, traceback):
if m is not None and hasattr(m, "__file__"):
self.pylib_dirs.add(self._canonical_dir(m))
if _structseq and not hasattr(_structseq, '__file__'):
# PyPy 2.4 has no __file__ in the builtin modules, but the code
# objects still have the file names. So dig into one to find
# the path to exclude.
structseq_new = _structseq.structseq_new
try:
structseq_file = structseq_new.func_code.co_filename
except AttributeError:
structseq_file = structseq_new.__code__.co_filename
self.pylib_dirs.add(self._canonical_dir(structseq_file))
# To avoid tracing the coverage.py code itself, we skip anything
# located where we are.
self.cover_dirs = [self._canonical_dir(__file__)]
if env.TESTING:
# When testing, we use PyContracts, which should be considered
# part of coverage.py, and it uses six. Exclude those directories
# just as we exclude ourselves.
import contracts, six
for mod in [contracts, six]:
self.cover_dirs.append(self._canonical_dir(mod))
# Set the reporting precision.
Numbers.set_precision(self.config.precision)
atexit.register(self._atexit)
self._inited = True
# Create the matchers we need for _should_trace
if self.source or self.source_pkgs:
self.source_match = TreeMatcher(self.source)
self.source_pkgs_match = ModuleMatcher(self.source_pkgs)
else:
if self.cover_dirs:
self.cover_match = TreeMatcher(self.cover_dirs)
if self.pylib_dirs:
self.pylib_match = TreeMatcher(self.pylib_dirs)
if self.include:
self.include_match = FnmatchMatcher(self.include)
if self.omit:
self.omit_match = FnmatchMatcher(self.omit)
# The user may want to debug things, show info if desired.
wrote_any = False
if self.debug.should('config'):
config_info = sorted(self.config.__dict__.items())
self.debug.write_formatted_info("config", config_info)
wrote_any = True
if self.debug.should('sys'):
self.debug.write_formatted_info("sys", self.sys_info())
for plugin in self.plugins:
header = "sys: " + plugin._coverage_plugin_name
info = plugin.sys_info()
self.debug.write_formatted_info(header, info)
wrote_any = True
if wrote_any:
self.debug.write_formatted_info("end", ())
def _canonical_dir(self, morf):
"""Return the canonical directory of the module or file `morf`."""
morf_filename = PythonFileReporter(morf, self).filename
return os.path.split(morf_filename)[0]
def _source_for_file(self, filename):
"""Return the source file for `filename`.
Given a file name being traced, return the best guess as to the source
file to attribute it to.
"""
if filename.endswith(".py"):
# .py files are themselves source files.
return filename
elif filename.endswith((".pyc", ".pyo")):
# Bytecode files probably have source files near them.
py_filename = filename[:-1]
if os.path.exists(py_filename):
# Found a .py file, use that.
return py_filename
if env.WINDOWS:
# On Windows, it could be a .pyw file.
pyw_filename = py_filename + "w"
if os.path.exists(pyw_filename):
return pyw_filename
# Didn't find source, but it's probably the .py file we want.
return py_filename
elif filename.endswith("$py.class"):
# Jython is easy to guess.
return filename[:-9] + ".py"
# No idea, just use the file name as-is.
return filename
def _name_for_module(self, module_globals, filename):
"""Get the name of the module for a set of globals and file name.
For configurability's sake, we allow __main__ modules to be matched by
their importable name.
If loaded via runpy (aka -m), we can usually recover the "original"
full dotted module name, otherwise, we resort to interpreting the
file name to get the module's name. In the case that the module name
can't be determined, None is returned.
"""
dunder_name = module_globals.get('__name__', None)
if isinstance(dunder_name, str) and dunder_name != '__main__':
# This is the usual case: an imported module.
return dunder_name
loader = module_globals.get('__loader__', None)
for attrname in ('fullname', 'name'): # attribute renamed in py3.2
if hasattr(loader, attrname):
fullname = getattr(loader, attrname)
else:
continue
if isinstance(fullname, str) and fullname != '__main__':
# Module loaded via: runpy -m
return fullname
# Script as first argument to Python command line.
inspectedname = inspect.getmodulename(filename)
if inspectedname is not None:
return inspectedname
else:
return dunder_name
def _should_trace_internal(self, filename, frame):
"""Decide whether to trace execution in `filename`, with a reason.
This function is called from the trace function. As each new file name
is encountered, this function determines whether it is traced or not.
Returns a FileDisposition object.
"""
original_filename = filename
disp = _disposition_init(self.collector.file_disposition_class, filename)
def nope(disp, reason):
"""Simple helper to make it easy to return NO."""
disp.trace = False
disp.reason = reason
return disp
# Compiled Python files have two file names: frame.f_code.co_filename is
# the file name at the time the .pyc was compiled. The second name is
# __file__, which is where the .pyc was actually loaded from. Since
# .pyc files can be moved after compilation (for example, by being
# installed), we look for __file__ in the frame and prefer it to the
# co_filename value.
dunder_file = frame.f_globals.get('__file__')
if dunder_file:
filename = self._source_for_file(dunder_file)
if original_filename and not original_filename.startswith('<'):
orig = os.path.basename(original_filename)
if orig != os.path.basename(filename):
# Files shouldn't be renamed when moved. This happens when
# exec'ing code. If it seems like something is wrong with
# the frame's file name, then just use the original.
filename = original_filename
if not filename:
# Empty string is pretty useless.
return nope(disp, "empty string isn't a file name")
if filename.startswith('memory:'):
return nope(disp, "memory isn't traceable")
if filename.startswith('<'):
# Lots of non-file execution is represented with artificial
# file names like "<string>", "<doctest readme.txt[0]>", or
# "<exec_function>". Don't ever trace these executions, since we
# can't do anything with the data later anyway.
return nope(disp, "not a real file name")
# pyexpat does a dumb thing, calling the trace function explicitly from
# C code with a C file name.
if re.search(r"[/\\]Modules[/\\]pyexpat.c", filename):
return nope(disp, "pyexpat lies about itself")
# Jython reports the .class file to the tracer, use the source file.
if filename.endswith("$py.class"):
filename = filename[:-9] + ".py"
canonical = files.canonical_filename(filename)
disp.canonical_filename = canonical
# Try the plugins, see if they have an opinion about the file.
plugin = None
for plugin in self.plugins.file_tracers:
if not plugin._coverage_enabled:
continue
try:
file_tracer = plugin.file_tracer(canonical)
if file_tracer is not None:
file_tracer._coverage_plugin = plugin
disp.trace = True
disp.file_tracer = file_tracer
if file_tracer.has_dynamic_source_filename():
disp.has_dynamic_filename = True
else:
disp.source_filename = files.canonical_filename(
file_tracer.source_filename()
)
break
except Exception:
self._warn(
"Disabling plugin %r due to an exception:" % (
plugin._coverage_plugin_name
)
)
traceback.print_exc()
plugin._coverage_enabled = False
continue
else:
# No plugin wanted it: it's Python.
disp.trace = True
disp.source_filename = canonical
if not disp.has_dynamic_filename:
if not disp.source_filename:
raise CoverageException(
"Plugin %r didn't set source_filename for %r" %
(plugin, disp.original_filename)
)
reason = self._check_include_omit_etc_internal(
disp.source_filename, frame,
)
if reason:
nope(disp, reason)
return disp
def _check_include_omit_etc_internal(self, filename, frame):
"""Check a file name against the include, omit, etc, rules.
Returns a string or None. String means, don't trace, and is the reason
why. None means no reason found to not trace.
"""
modulename = self._name_for_module(frame.f_globals, filename)
# If the user specified source or include, then that's authoritative
# about the outer bound of what to measure and we don't have to apply
# any canned exclusions. If they didn't, then we have to exclude the
# stdlib and coverage.py directories.
if self.source_match:
if self.source_pkgs_match.match(modulename):
if modulename in self.source_pkgs:
self.source_pkgs.remove(modulename)
return None # There's no reason to skip this file.
if not self.source_match.match(filename):
return "falls outside the --source trees"
elif self.include_match:
if not self.include_match.match(filename):
return "falls outside the --include trees"
else:
# If we aren't supposed to trace installed code, then check if this
# is near the Python standard library and skip it if so.
if self.pylib_match and self.pylib_match.match(filename):
return "is in the stdlib"
# We exclude the coverage.py code itself, since a little of it
# will be measured otherwise.
if self.cover_match and self.cover_match.match(filename):
return "is part of coverage.py"
# Check the file against the omit pattern.
if self.omit_match and self.omit_match.match(filename):
return "is inside an --omit pattern"
# No reason found to skip this file.
return None
def _should_trace(self, filename, frame):
"""Decide whether to trace execution in `filename`.
Calls `_should_trace_internal`, and returns the FileDisposition.
"""
disp = self._should_trace_internal(filename, frame)
if self.debug.should('trace'):
self.debug.write(_disposition_debug_msg(disp))
return disp
def _check_include_omit_etc(self, filename, frame):
"""Check a file name against the include/omit/etc, rules, verbosely.
Returns a boolean: True if the file should be traced, False if not.
"""
reason = self._check_include_omit_etc_internal(filename, frame)
if self.debug.should('trace'):
if not reason:
msg = "Including %r" % (filename,)
else:
msg = "Not including %r: %s" % (filename, reason)
self.debug.write(msg)
return not reason
def _warn(self, msg):
"""Use `msg` as a warning."""
self._warnings.append(msg)
if self.debug.should('pid'):
msg = "[%d] %s" % (os.getpid(), msg)
sys.stderr.write("Coverage.py warning: %s\n" % msg)
def get_option(self, option_name):
"""Get an option from the configuration.
`option_name` is a colon-separated string indicating the section and
option name. For example, the ``branch`` option in the ``[run]``
section of the config file would be indicated with `"run:branch"`.
Returns the value of the option.
.. versionadded:: 4.0
"""
return self.config.get_option(option_name)
def set_option(self, option_name, value):
"""Set an option in the configuration.
`option_name` is a colon-separated string indicating the section and
option name. For example, the ``branch`` option in the ``[run]``
section of the config file would be indicated with ``"run:branch"``.
`value` is the new value for the option. This should be a Python
value where appropriate. For example, use True for booleans, not the
string ``"True"``.
As an example, calling::
cov.set_option("run:branch", True)
has the same effect as this configuration file::
[run]
branch = True
.. versionadded:: 4.0
"""
self.config.set_option(option_name, value)
def use_cache(self, usecache):
"""Obsolete method."""
self._init()
if not usecache:
self._warn("use_cache(False) is no longer supported.")
def load(self):
"""Load previously-collected coverage data from the data file."""
self._init()
self.collector.reset()
self.data_files.read(self.data)
def start(self):
"""Start measuring code coverage.
Coverage measurement actually occurs in functions called after
:meth:`start` is invoked. Statements in the same scope as
:meth:`start` won't be measured.
Once you invoke :meth:`start`, you must also call :meth:`stop`
eventually, or your process might not shut down cleanly.
"""
self._init()
if self.run_suffix:
# Calling start() means we're running code, so use the run_suffix
# as the data_suffix when we eventually save the data.
self.data_suffix = self.run_suffix
if self._auto_data:
self.load()
self.collector.start()
self._started = True
self._measured = True
def stop(self):
"""Stop measuring code coverage."""
if self._started:
self.collector.stop()
self._started = False
def _atexit(self):
"""Clean up on process shutdown."""
if self._started:
self.stop()
if self._auto_data:
self.save()
def erase(self):
"""Erase previously-collected coverage data.
This removes the in-memory data collected in this session as well as
discarding the data file.
"""
self._init()
self.collector.reset()
self.data.erase()
self.data_files.erase(parallel=self.config.parallel)
def clear_exclude(self, which='exclude'):
"""Clear the exclude list."""
self._init()
setattr(self.config, which + "_list", [])
self._exclude_regex_stale()
def exclude(self, regex, which='exclude'):
"""Exclude source lines from execution consideration.
A number of lists of regular expressions are maintained. Each list
selects lines that are treated differently during reporting.
`which` determines which list is modified. The "exclude" list selects
lines that are not considered executable at all. The "partial" list
indicates lines with branches that are not taken.
`regex` is a regular expression. The regex is added to the specified
list. If any of the regexes in the list is found in a line, the line
is marked for special treatment during reporting.
"""
self._init()
excl_list = getattr(self.config, which + "_list")
excl_list.append(regex)
self._exclude_regex_stale()
def _exclude_regex_stale(self):
"""Drop all the compiled exclusion regexes, a list was modified."""
self._exclude_re.clear()
def _exclude_regex(self, which):
"""Return a compiled regex for the given exclusion list."""
if which not in self._exclude_re:
excl_list = getattr(self.config, which + "_list")
self._exclude_re[which] = join_regex(excl_list)
return self._exclude_re[which]
def get_exclude_list(self, which='exclude'):
"""Return a list of excluded regex patterns.
`which` indicates which list is desired. See :meth:`exclude` for the
lists that are available, and their meaning.
"""
self._init()
return getattr(self.config, which + "_list")
def save(self):
"""Save the collected coverage data to the data file."""
self._init()
self.get_data()
self.data_files.write(self.data, suffix=self.data_suffix)
def combine(self, data_paths=None):
"""Combine together a number of similarly-named coverage data files.
All coverage data files whose name starts with `data_file` (from the
coverage() constructor) will be read, and combined together into the
current measurements.
`data_paths` is a list of files or directories from which data should
be combined. If no list is passed, then the data files from the
directory indicated by the current data file (probably the current
directory) will be combined.
.. versionadded:: 4.0
The `data_paths` parameter.
"""
self._init()
self.get_data()
aliases = None
if self.config.paths:
aliases = PathAliases()
for paths in self.config.paths.values():
result = paths[0]
for pattern in paths[1:]:
aliases.add(pattern, result)
self.data_files.combine_parallel_data(self.data, aliases=aliases, data_paths=data_paths)
def get_data(self):
"""Get the collected data and reset the collector.
Also warn about various problems collecting data.
Returns a :class:`coverage.CoverageData`, the collected coverage data.
.. versionadded:: 4.0
"""
self._init()
if not self._measured:
return self.data
self.collector.save_data(self.data)
# If there are still entries in the source_pkgs list, then we never
# encountered those packages.
if self._warn_unimported_source:
for pkg in self.source_pkgs:
if pkg not in sys.modules:
self._warn("Module %s was never imported." % pkg)
elif not (
hasattr(sys.modules[pkg], '__file__') and
os.path.exists(sys.modules[pkg].__file__)
):
self._warn("Module %s has no Python source." % pkg)
else:
self._warn("Module %s was previously imported, but not measured." % pkg)
# Find out if we got any data.
if not self.data and self._warn_no_data:
self._warn("No data was collected.")
# Find files that were never executed at all.
for src in self.source:
for py_file in find_python_files(src):
py_file = files.canonical_filename(py_file)
if self.omit_match and self.omit_match.match(py_file):
# Turns out this file was omitted, so don't pull it back
# in as unexecuted.
continue
self.data.touch_file(py_file)
if self.config.note:
self.data.add_run_info(note=self.config.note)
self._measured = False
return self.data
# Backward compatibility with version 1.
def analysis(self, morf):
"""Like `analysis2` but doesn't return excluded line numbers."""
f, s, _, m, mf = self.analysis2(morf)
return f, s, m, mf
def analysis2(self, morf):
"""Analyze a module.
`morf` is a module or a file name. It will be analyzed to determine
its coverage statistics. The return value is a 5-tuple:
* The file name for the module.
* A list of line numbers of executable statements.
* A list of line numbers of excluded statements.
* A list of line numbers of statements not run (missing from
execution).
* A readable formatted string of the missing line numbers.
The analysis uses the source file itself and the current measured
coverage data.
"""
self._init()
analysis = self._analyze(morf)
return (
analysis.filename,
sorted(analysis.statements),
sorted(analysis.excluded),
sorted(analysis.missing),
analysis.missing_formatted(),
)
def _analyze(self, it):
"""Analyze a single morf or code unit.
Returns an `Analysis` object.
"""
self.get_data()
if not isinstance(it, FileReporter):
it = self._get_file_reporter(it)
return Analysis(self.data, it)
def _get_file_reporter(self, morf):
"""Get a FileReporter for a module or file name."""
plugin = None
file_reporter = "python"
if isinstance(morf, string_class):
abs_morf = abs_file(morf)
plugin_name = self.data.file_tracer(abs_morf)
if plugin_name:
plugin = self.plugins.get(plugin_name)
if plugin:
file_reporter = plugin.file_reporter(abs_morf)
if file_reporter is None:
raise CoverageException(
"Plugin %r did not provide a file reporter for %r." % (
plugin._coverage_plugin_name, morf
)
)
if file_reporter == "python":
file_reporter = PythonFileReporter(morf, self)
return file_reporter
def _get_file_reporters(self, morfs=None):
"""Get a list of FileReporters for a list of modules or file names.
For each module or file name in `morfs`, find a FileReporter. Return
the list of FileReporters.
If `morfs` is a single module or file name, this returns a list of one
FileReporter. If `morfs` is empty or None, then the list of all files
measured is used to find the FileReporters.
"""
if not morfs:
morfs = self.data.measured_files()
# Be sure we have a list.
if not isinstance(morfs, (list, tuple)):
morfs = [morfs]
file_reporters = []
for morf in morfs:
file_reporter = self._get_file_reporter(morf)
file_reporters.append(file_reporter)
return file_reporters
def report(
self, morfs=None, show_missing=None, ignore_errors=None,
file=None, # pylint: disable=redefined-builtin
omit=None, include=None, skip_covered=None,
):
"""Write a summary report to `file`.
Each module in `morfs` is listed, with counts of statements, executed
statements, missing statements, and a list of lines missed.
`include` is a list of file name patterns. Files that match will be
included in the report. Files matching `omit` will not be included in
the report.
Returns a float, the total percentage covered.
"""
self.get_data()
self.config.from_args(
ignore_errors=ignore_errors, omit=omit, include=include,
show_missing=show_missing, skip_covered=skip_covered,
)
reporter = SummaryReporter(self, self.config)
return reporter.report(morfs, outfile=file)
def annotate(
self, morfs=None, directory=None, ignore_errors=None,
omit=None, include=None,
):
"""Annotate a list of modules.
Each module in `morfs` is annotated. The source is written to a new
file, named with a ",cover" suffix, with each line prefixed with a
marker to indicate the coverage of the line. Covered lines have ">",
excluded lines have "-", and missing lines have "!".
See :meth:`report` for other arguments.
"""
self.get_data()
self.config.from_args(
ignore_errors=ignore_errors, omit=omit, include=include
)
reporter = AnnotateReporter(self, self.config)
reporter.report(morfs, directory=directory)
def html_report(self, morfs=None, directory=None, ignore_errors=None,
omit=None, include=None, extra_css=None, title=None):
"""Generate an HTML report.
The HTML is written to `directory`. The file "index.html" is the
overview starting point, with links to more detailed pages for
individual modules.
`extra_css` is a path to a file of other CSS to apply on the page.
It will be copied into the HTML directory.
`title` is a text string (not HTML) to use as the title of the HTML
report.
See :meth:`report` for other arguments.
Returns a float, the total percentage covered.
"""
self.get_data()
self.config.from_args(
ignore_errors=ignore_errors, omit=omit, include=include,
html_dir=directory, extra_css=extra_css, html_title=title,
)
reporter = HtmlReporter(self, self.config)
return reporter.report(morfs)
def xml_report(
self, morfs=None, outfile=None, ignore_errors=None,
omit=None, include=None,
):
"""Generate an XML report of coverage results.
The report is compatible with Cobertura reports.
Each module in `morfs` is included in the report. `outfile` is the
path to write the file to, "-" will write to stdout.
See :meth:`report` for other arguments.
Returns a float, the total percentage covered.
"""
self.get_data()
self.config.from_args(
ignore_errors=ignore_errors, omit=omit, include=include,
xml_output=outfile,
)
file_to_close = None
delete_file = False
if self.config.xml_output:
if self.config.xml_output == '-':
outfile = sys.stdout
else:
# Ensure that the output directory is created; done here
# because this report pre-opens the output file.
# HTMLReport does this using the Report plumbing because
# its task is more complex, being multiple files.
output_dir = os.path.dirname(self.config.xml_output)
if output_dir and not os.path.isdir(output_dir):
os.makedirs(output_dir)
open_kwargs = {}
if env.PY3:
open_kwargs['encoding'] = 'utf8'
outfile = open(self.config.xml_output, "w", **open_kwargs)
file_to_close = outfile
try:
reporter = XmlReporter(self, self.config)
return reporter.report(morfs, outfile=outfile)
except CoverageException:
delete_file = True
raise
finally:
if file_to_close:
file_to_close.close()
if delete_file:
file_be_gone(self.config.xml_output)
def sys_info(self):
"""Return a list of (key, value) pairs showing internal information."""
import coverage as covmod
self._init()
ft_plugins = []
for ft in self.plugins.file_tracers:
ft_name = ft._coverage_plugin_name
if not ft._coverage_enabled:
ft_name += " (disabled)"
ft_plugins.append(ft_name)
info = [
('version', covmod.__version__),
('coverage', covmod.__file__),
('cover_dirs', self.cover_dirs),
('pylib_dirs', self.pylib_dirs),
('tracer', self.collector.tracer_name()),
('plugins.file_tracers', ft_plugins),
('config_files', self.config.attempted_config_files),
('configs_read', self.config.config_files),
('data_path', self.data_files.filename),
('python', sys.version.replace('\n', '')),
('platform', platform.platform()),
('implementation', platform.python_implementation()),
('executable', sys.executable),
('cwd', os.getcwd()),
('path', sys.path),
('environment', sorted(
("%s = %s" % (k, v))
for k, v in iitems(os.environ)
if k.startswith(("COV", "PY"))
)),
('command_line', " ".join(getattr(sys, 'argv', ['???']))),
]
matcher_names = [
'source_match', 'source_pkgs_match',
'include_match', 'omit_match',
'cover_match', 'pylib_match',
]
for matcher_name in matcher_names:
matcher = getattr(self, matcher_name)
if matcher:
matcher_info = matcher.info()
else:
matcher_info = '-none-'
info.append((matcher_name, matcher_info))
return info
class InOrOut(object):
"""Machinery for determining what files to measure."""
def __init__(self, warn, debug):
self.warn = warn
self.debug = debug
# The matchers for should_trace.
self.source_match = None
self.source_pkgs_match = None
self.pylib_paths = self.cover_paths = None
self.pylib_match = self.cover_match = None
self.include_match = self.omit_match = None
self.plugins = []
self.disp_class = FileDisposition
# The source argument can be directories or package names.
self.source = []
self.source_pkgs = []
self.source_pkgs_unmatched = []
self.omit = self.include = None
def configure(self, config):
"""Apply the configuration to get ready for decision-time."""
self.source_pkgs.extend(config.source_pkgs)
for src in config.source or []:
if os.path.isdir(src):
self.source.append(canonical_filename(src))
else:
self.source_pkgs.append(src)
self.source_pkgs_unmatched = self.source_pkgs[:]
self.omit = prep_patterns(config.run_omit)
self.include = prep_patterns(config.run_include)
# The directories for files considered "installed with the interpreter".
self.pylib_paths = set()
if not config.cover_pylib:
# Look at where some standard modules are located. That's the
# indication for "installed with the interpreter". In some
# environments (virtualenv, for example), these modules may be
# spread across a few locations. Look at all the candidate modules
# we've imported, and take all the different ones.
for m in (atexit, inspect, os, platform, _pypy_irc_topic, re, _structseq, traceback):
if m is not None and hasattr(m, "__file__"):
self.pylib_paths.add(canonical_path(m, directory=True))
if _structseq and not hasattr(_structseq, '__file__'):
# PyPy 2.4 has no __file__ in the builtin modules, but the code
# objects still have the file names. So dig into one to find
# the path to exclude. The "filename" might be synthetic,
# don't be fooled by those.
structseq_file = code_object(_structseq.structseq_new).co_filename
if not structseq_file.startswith("<"):
self.pylib_paths.add(canonical_path(structseq_file))
# To avoid tracing the coverage.py code itself, we skip anything
# located where we are.
self.cover_paths = [canonical_path(__file__, directory=True)]
if env.TESTING:
# Don't include our own test code.
self.cover_paths.append(os.path.join(self.cover_paths[0], "tests"))
# When testing, we use PyContracts, which should be considered
# part of coverage.py, and it uses six. Exclude those directories
# just as we exclude ourselves.
import contracts
import six
for mod in [contracts, six]:
self.cover_paths.append(canonical_path(mod))
def debug(msg):
if self.debug:
self.debug.write(msg)
# Create the matchers we need for should_trace
if self.source or self.source_pkgs:
against = []
if self.source:
self.source_match = TreeMatcher(self.source)
against.append("trees {!r}".format(self.source_match))
if self.source_pkgs:
self.source_pkgs_match = ModuleMatcher(self.source_pkgs)
against.append("modules {!r}".format(self.source_pkgs_match))
debug("Source matching against " + " and ".join(against))
else:
if self.cover_paths:
self.cover_match = TreeMatcher(self.cover_paths)
debug("Coverage code matching: {!r}".format(self.cover_match))
if self.pylib_paths:
self.pylib_match = TreeMatcher(self.pylib_paths)
debug("Python stdlib matching: {!r}".format(self.pylib_match))
if self.include:
self.include_match = FnmatchMatcher(self.include)
debug("Include matching: {!r}".format(self.include_match))
if self.omit:
self.omit_match = FnmatchMatcher(self.omit)
debug("Omit matching: {!r}".format(self.omit_match))
def should_trace(self, filename, frame=None):
"""Decide whether to trace execution in `filename`, with a reason.
This function is called from the trace function. As each new file name
is encountered, this function determines whether it is traced or not.
Returns a FileDisposition object.
"""
original_filename = filename
disp = disposition_init(self.disp_class, filename)
def nope(disp, reason):
"""Simple helper to make it easy to return NO."""
disp.trace = False
disp.reason = reason
return disp
if frame is not None:
# Compiled Python files have two file names: frame.f_code.co_filename is
# the file name at the time the .pyc was compiled. The second name is
# __file__, which is where the .pyc was actually loaded from. Since
# .pyc files can be moved after compilation (for example, by being
# installed), we look for __file__ in the frame and prefer it to the
# co_filename value.
dunder_file = frame.f_globals and frame.f_globals.get('__file__')
if dunder_file:
filename = source_for_file(dunder_file)
if original_filename and not original_filename.startswith('<'):
orig = os.path.basename(original_filename)
if orig != os.path.basename(filename):
# Files shouldn't be renamed when moved. This happens when
# exec'ing code. If it seems like something is wrong with
# the frame's file name, then just use the original.
filename = original_filename
if not filename:
# Empty string is pretty useless.
return nope(disp, "empty string isn't a file name")
if filename.startswith('memory:'):
return nope(disp, "memory isn't traceable")
if filename.startswith('<'):
# Lots of non-file execution is represented with artificial
# file names like "<string>", "<doctest readme.txt[0]>", or
# "<exec_function>". Don't ever trace these executions, since we
# can't do anything with the data later anyway.
return nope(disp, "not a real file name")
# pyexpat does a dumb thing, calling the trace function explicitly from
# C code with a C file name.
if re.search(r"[/\\]Modules[/\\]pyexpat.c", filename):
return nope(disp, "pyexpat lies about itself")
# Jython reports the .class file to the tracer, use the source file.
if filename.endswith("$py.class"):
filename = filename[:-9] + ".py"
canonical = canonical_filename(filename)
disp.canonical_filename = canonical
# Try the plugins, see if they have an opinion about the file.
plugin = None
for plugin in self.plugins.file_tracers:
if not plugin._coverage_enabled:
continue
try:
file_tracer = plugin.file_tracer(canonical)
if file_tracer is not None:
file_tracer._coverage_plugin = plugin
disp.trace = True
disp.file_tracer = file_tracer
if file_tracer.has_dynamic_source_filename():
disp.has_dynamic_filename = True
else:
disp.source_filename = canonical_filename(
file_tracer.source_filename()
)
break
except Exception:
self.warn(
"Disabling plug-in %r due to an exception:" % (plugin._coverage_plugin_name)
)
traceback.print_exc()
plugin._coverage_enabled = False
continue
else:
# No plugin wanted it: it's Python.
disp.trace = True
disp.source_filename = canonical
if not disp.has_dynamic_filename:
if not disp.source_filename:
raise CoverageException(
"Plugin %r didn't set source_filename for %r" %
(plugin, disp.original_filename)
)
reason = self.check_include_omit_etc(disp.source_filename, frame)
if reason:
nope(disp, reason)
return disp
def check_include_omit_etc(self, filename, frame):
"""Check a file name against the include, omit, etc, rules.
Returns a string or None. String means, don't trace, and is the reason
why. None means no reason found to not trace.
"""
modulename = name_for_module(filename, frame)
# If the user specified source or include, then that's authoritative
# about the outer bound of what to measure and we don't have to apply
# any canned exclusions. If they didn't, then we have to exclude the
# stdlib and coverage.py directories.
if self.source_match or self.source_pkgs_match:
extra = ""
ok = False
if self.source_pkgs_match:
if self.source_pkgs_match.match(modulename):
ok = True
if modulename in self.source_pkgs_unmatched:
self.source_pkgs_unmatched.remove(modulename)
else:
extra = "module {!r} ".format(modulename)
if not ok and self.source_match:
if self.source_match.match(filename):
ok = True
if not ok:
return extra + "falls outside the --source spec"
elif self.include_match:
if not self.include_match.match(filename):
return "falls outside the --include trees"
else:
# If we aren't supposed to trace installed code, then check if this
# is near the Python standard library and skip it if so.
if self.pylib_match and self.pylib_match.match(filename):
return "is in the stdlib"
# We exclude the coverage.py code itself, since a little of it
# will be measured otherwise.
if self.cover_match and self.cover_match.match(filename):
return "is part of coverage.py"
# Check the file against the omit pattern.
if self.omit_match and self.omit_match.match(filename):
return "is inside an --omit pattern"
# No point tracing a file we can't later write to SQLite.
try:
filename.encode("utf8")
except UnicodeEncodeError:
return "non-encodable filename"
# No reason found to skip this file.
return None
def warn_conflicting_settings(self):
"""Warn if there are settings that conflict."""
if self.include:
if self.source or self.source_pkgs:
self.warn("--include is ignored because --source is set", slug="include-ignored")
def warn_already_imported_files(self):
"""Warn if files have already been imported that we will be measuring."""
if self.include or self.source or self.source_pkgs:
warned = set()
for mod in list(sys.modules.values()):
filename = getattr(mod, "__file__", None)
if filename is None:
continue
if filename in warned:
continue
disp = self.should_trace(filename)
if disp.trace:
msg = "Already imported a file that will be measured: {}".format(filename)
self.warn(msg, slug="already-imported")
warned.add(filename)
def warn_unimported_source(self):
"""Warn about source packages that were of interest, but never traced."""
for pkg in self.source_pkgs_unmatched:
self._warn_about_unmeasured_code(pkg)
def _warn_about_unmeasured_code(self, pkg):
"""Warn about a package or module that we never traced.
`pkg` is a string, the name of the package or module.
"""
mod = sys.modules.get(pkg)
if mod is None:
self.warn("Module %s was never imported." % pkg, slug="module-not-imported")
return
if module_is_namespace(mod):
# A namespace package. It's OK for this not to have been traced,
# since there is no code directly in it.
return
if not module_has_file(mod):
self.warn("Module %s has no Python source." % pkg, slug="module-not-python")
return
# The module was in sys.modules, and seems like a module with code, but
# we never measured it. I guess that means it was imported before
# coverage even started.
self.warn(
"Module %s was previously imported, but not measured" % pkg,
slug="module-not-measured",
)
def find_possibly_unexecuted_files(self):
"""Find files in the areas of interest that might be untraced.
Yields pairs: file path, and responsible plug-in name.
"""
for pkg in self.source_pkgs:
if (not pkg in sys.modules or
not module_has_file(sys.modules[pkg])):
continue
pkg_file = source_for_file(sys.modules[pkg].__file__)
for ret in self._find_executable_files(canonical_path(pkg_file)):
yield ret
for src in self.source:
for ret in self._find_executable_files(src):
yield ret
def _find_plugin_files(self, src_dir):
"""Get executable files from the plugins."""
for plugin in self.plugins.file_tracers:
for x_file in plugin.find_executable_files(src_dir):
yield x_file, plugin._coverage_plugin_name
def _find_executable_files(self, src_dir):
"""Find executable files in `src_dir`.
Search for files in `src_dir` that can be executed because they
are probably importable. Don't include ones that have been omitted
by the configuration.
Yield the file path, and the plugin name that handles the file.
"""
py_files = ((py_file, None) for py_file in find_python_files(src_dir))
plugin_files = self._find_plugin_files(src_dir)
for file_path, plugin_name in itertools.chain(py_files, plugin_files):
file_path = canonical_filename(file_path)
if self.omit_match and self.omit_match.match(file_path):
# Turns out this file was omitted, so don't pull it back
# in as unexecuted.
continue
yield file_path, plugin_name
def sys_info(self):
"""Our information for Coverage.sys_info.
Returns a list of (key, value) pairs.
"""
info = [
('cover_paths', self.cover_paths),
('pylib_paths', self.pylib_paths),
]
matcher_names = [
'source_match', 'source_pkgs_match',
'include_match', 'omit_match',
'cover_match', 'pylib_match',
]
for matcher_name in matcher_names:
matcher = getattr(self, matcher_name)
if matcher:
matcher_info = matcher.info()
else:
matcher_info = '-none-'
info.append((matcher_name, matcher_info))
return info
class Coverage(object):
"""Programmatic access to coverage.py.
To use::
from coverage import Coverage
cov = Coverage()
cov.start()
#.. call your code ..
cov.stop()
cov.html_report(directory='covhtml')
"""
def __init__(
self,
data_file=None,
data_suffix=None,
cover_pylib=None,
auto_data=False,
timid=None,
branch=None,
config_file=True,
source=None,
omit=None,
include=None,
debug=None,
concurrency=None,
):
"""
`data_file` is the base name of the data file to use, defaulting to
".coverage". `data_suffix` is appended (with a dot) to `data_file` to
create the final file name. If `data_suffix` is simply True, then a
suffix is created with the machine and process identity included.
`cover_pylib` is a boolean determining whether Python code installed
with the Python interpreter is measured. This includes the Python
standard library and any packages installed with the interpreter.
If `auto_data` is true, then any existing data file will be read when
coverage measurement starts, and data will be saved automatically when
measurement stops.
If `timid` is true, then a slower and simpler trace function will be
used. This is important for some environments where manipulation of
tracing functions breaks the faster trace function.
If `branch` is true, then branch coverage will be measured in addition
to the usual statement coverage.
`config_file` determines what configuration file to read:
* If it is ".coveragerc", it is interpreted as if it were True,
for backward compatibility.
* If it is a string, it is the name of the file to read. If the
file can't be read, it is an error.
* If it is True, then a few standard files names are tried
(".coveragerc", "setup.cfg"). It is not an error for these files
to not be found.
* If it is False, then no configuration file is read.
`source` is a list of file paths or package names. Only code located
in the trees indicated by the file paths or package names will be
measured.
`include` and `omit` are lists of file name patterns. Files that match
`include` will be measured, files that match `omit` will not. Each
will also accept a single string argument.
`debug` is a list of strings indicating what debugging information is
desired.
`concurrency` is a string indicating the concurrency library being used
in the measured code. Without this, coverage.py will get incorrect
results. Valid strings are "greenlet", "eventlet", "gevent", or
"thread" (the default).
.. versionadded:: 4.0
The `concurrency` parameter.
"""
# Build our configuration from a number of sources:
# 1: defaults:
self.config = CoverageConfig()
# 2: from the rcfile, .coveragerc or setup.cfg file:
if config_file:
did_read_rc = False
# Some API users were specifying ".coveragerc" to mean the same as
# True, so make it so.
if config_file == ".coveragerc":
config_file = True
specified_file = (config_file is not True)
if not specified_file:
config_file = ".coveragerc"
did_read_rc = self.config.from_file(config_file)
if not did_read_rc:
if specified_file:
raise CoverageException(
"Couldn't read '%s' as a config file" % config_file)
self.config.from_file("setup.cfg", section_prefix="coverage:")
# 3: from environment variables:
env_data_file = os.environ.get('COVERAGE_FILE')
if env_data_file:
self.config.data_file = env_data_file
debugs = os.environ.get('COVERAGE_DEBUG')
if debugs:
self.config.debug.extend(debugs.split(","))
# 4: from constructor arguments:
self.config.from_args(
data_file=data_file,
cover_pylib=cover_pylib,
timid=timid,
branch=branch,
parallel=bool_or_none(data_suffix),
source=source,
omit=omit,
include=include,
debug=debug,
concurrency=concurrency,
)
self._debug_file = None
self._auto_data = auto_data
self._data_suffix = data_suffix
# The matchers for _should_trace.
self.source_match = None
self.source_pkgs_match = None
self.pylib_match = self.cover_match = None
self.include_match = self.omit_match = None
# Is it ok for no data to be collected?
self._warn_no_data = True
self._warn_unimported_source = True
# A record of all the warnings that have been issued.
self._warnings = []
# Other instance attributes, set later.
self.omit = self.include = self.source = None
self.source_pkgs = None
self.data = self.data_files = self.collector = None
self.plugins = None
self.pylib_dirs = self.cover_dirs = None
self.data_suffix = self.run_suffix = None
self._exclude_re = None
self.debug = None
# State machine variables:
# Have we initialized everything?
self._inited = False
# Have we started collecting and not stopped it?
self._started = False
# Have we measured some data and not harvested it?
self._measured = False
def _init(self):
"""Set all the initial state.
This is called by the public methods to initialize state. This lets us
construct a :class:`Coverage` object, then tweak its state before this
function is called.
"""
if self._inited:
return
# Create and configure the debugging controller. COVERAGE_DEBUG_FILE
# is an environment variable, the name of a file to append debug logs
# to.
if self._debug_file is None:
debug_file_name = os.environ.get("COVERAGE_DEBUG_FILE")
if debug_file_name:
self._debug_file = open(debug_file_name, "a")
else:
self._debug_file = sys.stderr
self.debug = DebugControl(self.config.debug, self._debug_file)
# Load plugins
self.plugins = Plugins.load_plugins(self.config.plugins, self.config,
self.debug)
# _exclude_re is a dict that maps exclusion list names to compiled
# regexes.
self._exclude_re = {}
self._exclude_regex_stale()
files.set_relative_directory()
# The source argument can be directories or package names.
self.source = []
self.source_pkgs = []
for src in self.config.source or []:
if os.path.exists(src):
self.source.append(files.canonical_filename(src))
else:
self.source_pkgs.append(src)
self.omit = prep_patterns(self.config.omit)
self.include = prep_patterns(self.config.include)
concurrency = self.config.concurrency
if concurrency == "multiprocessing":
patch_multiprocessing()
concurrency = None
self.collector = Collector(
should_trace=self._should_trace,
check_include=self._check_include_omit_etc,
timid=self.config.timid,
branch=self.config.branch,
warn=self._warn,
concurrency=concurrency,
)
# Early warning if we aren't going to be able to support plugins.
if self.plugins.file_tracers and not self.collector.supports_plugins:
self._warn("Plugin file tracers (%s) aren't supported with %s" % (
", ".join(plugin._coverage_plugin_name
for plugin in self.plugins.file_tracers),
self.collector.tracer_name(),
))
for plugin in self.plugins.file_tracers:
plugin._coverage_enabled = False
# Suffixes are a bit tricky. We want to use the data suffix only when
# collecting data, not when combining data. So we save it as
# `self.run_suffix` now, and promote it to `self.data_suffix` if we
# find that we are collecting data later.
if self._data_suffix or self.config.parallel:
if not isinstance(self._data_suffix, string_class):
# if data_suffix=True, use .machinename.pid.random
self._data_suffix = True
else:
self._data_suffix = None
self.data_suffix = None
self.run_suffix = self._data_suffix
# Create the data file. We do this at construction time so that the
# data file will be written into the directory where the process
# started rather than wherever the process eventually chdir'd to.
self.data = CoverageData(debug=self.debug)
self.data_files = CoverageDataFiles(basename=self.config.data_file)
# The directories for files considered "installed with the interpreter".
self.pylib_dirs = set()
if not self.config.cover_pylib:
# Look at where some standard modules are located. That's the
# indication for "installed with the interpreter". In some
# environments (virtualenv, for example), these modules may be
# spread across a few locations. Look at all the candidate modules
# we've imported, and take all the different ones.
for m in (atexit, inspect, os, platform, _structseq, traceback):
if m is not None and hasattr(m, "__file__"):
self.pylib_dirs.add(self._canonical_dir(m))
if _structseq and not hasattr(_structseq, '__file__'):
# PyPy 2.4 has no __file__ in the builtin modules, but the code
# objects still have the file names. So dig into one to find
# the path to exclude.
structseq_new = _structseq.structseq_new
try:
structseq_file = structseq_new.func_code.co_filename
except AttributeError:
structseq_file = structseq_new.__code__.co_filename
self.pylib_dirs.add(self._canonical_dir(structseq_file))
# To avoid tracing the coverage.py code itself, we skip anything
# located where we are.
self.cover_dirs = [self._canonical_dir(__file__)]
if env.TESTING:
# When testing, we use PyContracts, which should be considered
# part of coverage.py, and it uses six. Exclude those directories
# just as we exclude ourselves.
import contracts, six
for mod in [contracts, six]:
self.cover_dirs.append(self._canonical_dir(mod))
# Set the reporting precision.
Numbers.set_precision(self.config.precision)
atexit.register(self._atexit)
self._inited = True
# Create the matchers we need for _should_trace
if self.source or self.source_pkgs:
self.source_match = TreeMatcher(self.source)
self.source_pkgs_match = ModuleMatcher(self.source_pkgs)
else:
if self.cover_dirs:
self.cover_match = TreeMatcher(self.cover_dirs)
if self.pylib_dirs:
self.pylib_match = TreeMatcher(self.pylib_dirs)
if self.include:
self.include_match = FnmatchMatcher(self.include)
if self.omit:
self.omit_match = FnmatchMatcher(self.omit)
# The user may want to debug things, show info if desired.
wrote_any = False
if self.debug.should('config'):
config_info = sorted(self.config.__dict__.items())
self.debug.write_formatted_info("config", config_info)
wrote_any = True
if self.debug.should('sys'):
self.debug.write_formatted_info("sys", self.sys_info())
for plugin in self.plugins:
header = "sys: " + plugin._coverage_plugin_name
info = plugin.sys_info()
self.debug.write_formatted_info(header, info)
wrote_any = True
if wrote_any:
self.debug.write_formatted_info("end", ())
def _canonical_dir(self, morf):
"""Return the canonical directory of the module or file `morf`."""
morf_filename = PythonFileReporter(morf, self).filename
return os.path.split(morf_filename)[0]
def _source_for_file(self, filename):
"""Return the source file for `filename`.
Given a file name being traced, return the best guess as to the source
file to attribute it to.
"""
if filename.endswith(".py"):
# .py files are themselves source files.
return filename
elif filename.endswith((".pyc", ".pyo")):
# Bytecode files probably have source files near them.
py_filename = filename[:-1]
if os.path.exists(py_filename):
# Found a .py file, use that.
return py_filename
if env.WINDOWS:
# On Windows, it could be a .pyw file.
pyw_filename = py_filename + "w"
if os.path.exists(pyw_filename):
return pyw_filename
# Didn't find source, but it's probably the .py file we want.
return py_filename
elif filename.endswith("$py.class"):
# Jython is easy to guess.
return filename[:-9] + ".py"
# No idea, just use the file name as-is.
return filename
def _name_for_module(self, module_globals, filename):
"""Get the name of the module for a set of globals and file name.
For configurability's sake, we allow __main__ modules to be matched by
their importable name.
If loaded via runpy (aka -m), we can usually recover the "original"
full dotted module name, otherwise, we resort to interpreting the
file name to get the module's name. In the case that the module name
can't be determined, None is returned.
"""
dunder_name = module_globals.get('__name__', None)
if isinstance(dunder_name, str) and dunder_name != '__main__':
# This is the usual case: an imported module.
return dunder_name
loader = module_globals.get('__loader__', None)
for attrname in ('fullname', 'name'): # attribute renamed in py3.2
if hasattr(loader, attrname):
fullname = getattr(loader, attrname)
else:
continue
if isinstance(fullname, str) and fullname != '__main__':
# Module loaded via: runpy -m
return fullname
# Script as first argument to Python command line.
inspectedname = inspect.getmodulename(filename)
if inspectedname is not None:
return inspectedname
else:
return dunder_name
def _should_trace_internal(self, filename, frame):
"""Decide whether to trace execution in `filename`, with a reason.
This function is called from the trace function. As each new file name
is encountered, this function determines whether it is traced or not.
Returns a FileDisposition object.
"""
original_filename = filename
disp = _disposition_init(self.collector.file_disposition_class,
filename)
def nope(disp, reason):
"""Simple helper to make it easy to return NO."""
disp.trace = False
disp.reason = reason
return disp
# Compiled Python files have two file names: frame.f_code.co_filename is
# the file name at the time the .pyc was compiled. The second name is
# __file__, which is where the .pyc was actually loaded from. Since
# .pyc files can be moved after compilation (for example, by being
# installed), we look for __file__ in the frame and prefer it to the
# co_filename value.
dunder_file = frame.f_globals.get('__file__')
if dunder_file:
filename = self._source_for_file(dunder_file)
if original_filename and not original_filename.startswith('<'):
orig = os.path.basename(original_filename)
if orig != os.path.basename(filename):
# Files shouldn't be renamed when moved. This happens when
# exec'ing code. If it seems like something is wrong with
# the frame's file name, then just use the original.
filename = original_filename
if not filename:
# Empty string is pretty useless.
return nope(disp, "empty string isn't a file name")
if filename.startswith('memory:'):
return nope(disp, "memory isn't traceable")
if filename.startswith('<'):
# Lots of non-file execution is represented with artificial
# file names like "<string>", "<doctest readme.txt[0]>", or
# "<exec_function>". Don't ever trace these executions, since we
# can't do anything with the data later anyway.
return nope(disp, "not a real file name")
# Jython reports the .class file to the tracer, use the source file.
if filename.endswith("$py.class"):
filename = filename[:-9] + ".py"
canonical = files.canonical_filename(filename)
disp.canonical_filename = canonical
# Try the plugins, see if they have an opinion about the file.
plugin = None
for plugin in self.plugins.file_tracers:
if not plugin._coverage_enabled:
continue
try:
file_tracer = plugin.file_tracer(canonical)
if file_tracer is not None:
file_tracer._coverage_plugin = plugin
disp.trace = True
disp.file_tracer = file_tracer
if file_tracer.has_dynamic_source_filename():
disp.has_dynamic_filename = True
else:
disp.source_filename = files.canonical_filename(
file_tracer.source_filename())
break
except Exception:
self._warn("Disabling plugin %r due to an exception:" %
(plugin._coverage_plugin_name))
traceback.print_exc()
plugin._coverage_enabled = False
continue
else:
# No plugin wanted it: it's Python.
disp.trace = True
disp.source_filename = canonical
if not disp.has_dynamic_filename:
if not disp.source_filename:
raise CoverageException(
"Plugin %r didn't set source_filename for %r" %
(plugin, disp.original_filename))
reason = self._check_include_omit_etc_internal(
disp.source_filename,
frame,
)
if reason:
nope(disp, reason)
return disp
def _check_include_omit_etc_internal(self, filename, frame):
"""Check a file name against the include, omit, etc, rules.
Returns a string or None. String means, don't trace, and is the reason
why. None means no reason found to not trace.
"""
modulename = self._name_for_module(frame.f_globals, filename)
# If the user specified source or include, then that's authoritative
# about the outer bound of what to measure and we don't have to apply
# any canned exclusions. If they didn't, then we have to exclude the
# stdlib and coverage.py directories.
if self.source_match:
if self.source_pkgs_match.match(modulename):
if modulename in self.source_pkgs:
self.source_pkgs.remove(modulename)
return None # There's no reason to skip this file.
if not self.source_match.match(filename):
return "falls outside the --source trees"
elif self.include_match:
if not self.include_match.match(filename):
return "falls outside the --include trees"
else:
# If we aren't supposed to trace installed code, then check if this
# is near the Python standard library and skip it if so.
if self.pylib_match and self.pylib_match.match(filename):
return "is in the stdlib"
# We exclude the coverage.py code itself, since a little of it
# will be measured otherwise.
if self.cover_match and self.cover_match.match(filename):
return "is part of coverage.py"
# Check the file against the omit pattern.
if self.omit_match and self.omit_match.match(filename):
return "is inside an --omit pattern"
# No reason found to skip this file.
return None
def _should_trace(self, filename, frame):
"""Decide whether to trace execution in `filename`.
Calls `_should_trace_internal`, and returns the FileDisposition.
"""
disp = self._should_trace_internal(filename, frame)
if self.debug.should('trace'):
self.debug.write(_disposition_debug_msg(disp))
return disp
def _check_include_omit_etc(self, filename, frame):
"""Check a file name against the include/omit/etc, rules, verbosely.
Returns a boolean: True if the file should be traced, False if not.
"""
reason = self._check_include_omit_etc_internal(filename, frame)
if self.debug.should('trace'):
if not reason:
msg = "Including %r" % (filename, )
else:
msg = "Not including %r: %s" % (filename, reason)
self.debug.write(msg)
return not reason
def _warn(self, msg):
"""Use `msg` as a warning."""
self._warnings.append(msg)
if self.debug.should('pid'):
msg = "[%d] %s" % (os.getpid(), msg)
sys.stderr.write("Coverage.py warning: %s\n" % msg)
def get_option(self, option_name):
"""Get an option from the configuration.
`option_name` is a colon-separated string indicating the section and
option name. For example, the ``branch`` option in the ``[run]``
section of the config file would be indicated with `"run:branch"`.
Returns the value of the option.
.. versionadded:: 4.0
"""
return self.config.get_option(option_name)
def set_option(self, option_name, value):
"""Set an option in the configuration.
`option_name` is a colon-separated string indicating the section and
option name. For example, the ``branch`` option in the ``[run]``
section of the config file would be indicated with ``"run:branch"``.
`value` is the new value for the option. This should be a Python
value where appropriate. For example, use True for booleans, not the
string ``"True"``.
As an example, calling::
cov.set_option("run:branch", True)
has the same effect as this configuration file::
[run]
branch = True
.. versionadded:: 4.0
"""
self.config.set_option(option_name, value)
def use_cache(self, usecache):
"""Obsolete method."""
self._init()
if not usecache:
self._warn("use_cache(False) is no longer supported.")
def load(self):
"""Load previously-collected coverage data from the data file."""
self._init()
self.collector.reset()
self.data_files.read(self.data)
def start(self):
"""Start measuring code coverage.
Coverage measurement actually occurs in functions called after
:meth:`start` is invoked. Statements in the same scope as
:meth:`start` won't be measured.
Once you invoke :meth:`start`, you must also call :meth:`stop`
eventually, or your process might not shut down cleanly.
"""
self._init()
if self.run_suffix:
# Calling start() means we're running code, so use the run_suffix
# as the data_suffix when we eventually save the data.
self.data_suffix = self.run_suffix
if self._auto_data:
self.load()
self.collector.start()
self._started = True
self._measured = True
def stop(self):
"""Stop measuring code coverage."""
if self._started:
self.collector.stop()
self._started = False
def _atexit(self):
"""Clean up on process shutdown."""
if self._started:
self.stop()
if self._auto_data:
self.save()
def erase(self):
"""Erase previously-collected coverage data.
This removes the in-memory data collected in this session as well as
discarding the data file.
"""
self._init()
self.collector.reset()
self.data.erase()
self.data_files.erase(parallel=self.config.parallel)
def clear_exclude(self, which='exclude'):
"""Clear the exclude list."""
self._init()
setattr(self.config, which + "_list", [])
self._exclude_regex_stale()
def exclude(self, regex, which='exclude'):
"""Exclude source lines from execution consideration.
A number of lists of regular expressions are maintained. Each list
selects lines that are treated differently during reporting.
`which` determines which list is modified. The "exclude" list selects
lines that are not considered executable at all. The "partial" list
indicates lines with branches that are not taken.
`regex` is a regular expression. The regex is added to the specified
list. If any of the regexes in the list is found in a line, the line
is marked for special treatment during reporting.
"""
self._init()
excl_list = getattr(self.config, which + "_list")
excl_list.append(regex)
self._exclude_regex_stale()
def _exclude_regex_stale(self):
"""Drop all the compiled exclusion regexes, a list was modified."""
self._exclude_re.clear()
def _exclude_regex(self, which):
"""Return a compiled regex for the given exclusion list."""
if which not in self._exclude_re:
excl_list = getattr(self.config, which + "_list")
self._exclude_re[which] = join_regex(excl_list)
return self._exclude_re[which]
def get_exclude_list(self, which='exclude'):
"""Return a list of excluded regex patterns.
`which` indicates which list is desired. See :meth:`exclude` for the
lists that are available, and their meaning.
"""
self._init()
return getattr(self.config, which + "_list")
def save(self):
"""Save the collected coverage data to the data file."""
self._init()
self.get_data()
self.data_files.write(self.data, suffix=self.data_suffix)
def combine(self, data_paths=None):
"""Combine together a number of similarly-named coverage data files.
All coverage data files whose name starts with `data_file` (from the
coverage() constructor) will be read, and combined together into the
current measurements.
`data_paths` is a list of files or directories from which data should
be combined. If no list is passed, then the data files from the
directory indicated by the current data file (probably the current
directory) will be combined.
.. versionadded:: 4.0
The `data_paths` parameter.
"""
self._init()
self.get_data()
aliases = None
if self.config.paths:
aliases = PathAliases()
for paths in self.config.paths.values():
result = paths[0]
for pattern in paths[1:]:
aliases.add(pattern, result)
self.data_files.combine_parallel_data(self.data,
aliases=aliases,
data_paths=data_paths)
def get_data(self):
"""Get the collected data and reset the collector.
Also warn about various problems collecting data.
Returns a :class:`coverage.CoverageData`, the collected coverage data.
.. versionadded:: 4.0
"""
self._init()
if not self._measured:
return
self.collector.save_data(self.data)
# If there are still entries in the source_pkgs list, then we never
# encountered those packages.
if self._warn_unimported_source:
for pkg in self.source_pkgs:
if pkg not in sys.modules:
self._warn("Module %s was never imported." % pkg)
elif not (hasattr(sys.modules[pkg], '__file__')
and os.path.exists(sys.modules[pkg].__file__)):
self._warn("Module %s has no Python source." % pkg)
else:
self._warn(
"Module %s was previously imported, but not measured."
% pkg)
# Find out if we got any data.
if not self.data and self._warn_no_data:
self._warn("No data was collected.")
# Find files that were never executed at all.
for src in self.source:
for py_file in find_python_files(src):
py_file = files.canonical_filename(py_file)
if self.omit_match and self.omit_match.match(py_file):
# Turns out this file was omitted, so don't pull it back
# in as unexecuted.
continue
self.data.touch_file(py_file)
# Add run information.
self.data.add_run_info(brief_sys=" ".join([
platform.python_implementation(),
platform.python_version(),
platform.system(),
]))
if self.config.note:
self.data.add_run_info(note=self.config.note)
self._measured = False
return self.data
# Backward compatibility with version 1.
def analysis(self, morf):
"""Like `analysis2` but doesn't return excluded line numbers."""
f, s, _, m, mf = self.analysis2(morf)
return f, s, m, mf
def analysis2(self, morf):
"""Analyze a module.
`morf` is a module or a file name. It will be analyzed to determine
its coverage statistics. The return value is a 5-tuple:
* The file name for the module.
* A list of line numbers of executable statements.
* A list of line numbers of excluded statements.
* A list of line numbers of statements not run (missing from
execution).
* A readable formatted string of the missing line numbers.
The analysis uses the source file itself and the current measured
coverage data.
"""
self._init()
analysis = self._analyze(morf)
return (
analysis.filename,
sorted(analysis.statements),
sorted(analysis.excluded),
sorted(analysis.missing),
analysis.missing_formatted(),
)
def _analyze(self, it):
"""Analyze a single morf or code unit.
Returns an `Analysis` object.
"""
self.get_data()
if not isinstance(it, FileReporter):
it = self._get_file_reporter(it)
return Analysis(self.data, it)
def _get_file_reporter(self, morf):
"""Get a FileReporter for a module or file name."""
plugin = None
file_reporter = "python"
if isinstance(morf, string_class):
abs_morf = abs_file(morf)
plugin_name = self.data.file_tracer(abs_morf)
if plugin_name:
plugin = self.plugins.get(plugin_name)
if plugin:
file_reporter = plugin.file_reporter(abs_morf)
if file_reporter is None:
raise CoverageException(
"Plugin %r did not provide a file reporter for %r." %
(plugin._coverage_plugin_name, morf))
if file_reporter == "python":
file_reporter = PythonFileReporter(morf, self)
return file_reporter
def _get_file_reporters(self, morfs=None):
"""Get a list of FileReporters for a list of modules or file names.
For each module or file name in `morfs`, find a FileReporter. Return
the list of FileReporters.
If `morfs` is a single module or file name, this returns a list of one
FileReporter. If `morfs` is empty or None, then the list of all files
measured is used to find the FileReporters.
"""
if not morfs:
morfs = self.data.measured_files()
# Be sure we have a list.
if not isinstance(morfs, (list, tuple)):
morfs = [morfs]
file_reporters = []
for morf in morfs:
file_reporter = self._get_file_reporter(morf)
file_reporters.append(file_reporter)
return file_reporters
def report(
self,
morfs=None,
show_missing=True,
ignore_errors=None,
file=None, # pylint: disable=redefined-builtin
omit=None,
include=None,
skip_covered=False,
):
"""Write a summary report to `file`.
Each module in `morfs` is listed, with counts of statements, executed
statements, missing statements, and a list of lines missed.
`include` is a list of file name patterns. Files that match will be
included in the report. Files matching `omit` will not be included in
the report.
Returns a float, the total percentage covered.
"""
self.get_data()
self.config.from_args(
ignore_errors=ignore_errors,
omit=omit,
include=include,
show_missing=show_missing,
skip_covered=skip_covered,
)
reporter = SummaryReporter(self, self.config)
return reporter.report(morfs, outfile=file)
def annotate(
self,
morfs=None,
directory=None,
ignore_errors=None,
omit=None,
include=None,
):
"""Annotate a list of modules.
Each module in `morfs` is annotated. The source is written to a new
file, named with a ",cover" suffix, with each line prefixed with a
marker to indicate the coverage of the line. Covered lines have ">",
excluded lines have "-", and missing lines have "!".
See :meth:`report` for other arguments.
"""
self.get_data()
self.config.from_args(ignore_errors=ignore_errors,
omit=omit,
include=include)
reporter = AnnotateReporter(self, self.config)
reporter.report(morfs, directory=directory)
def html_report(self,
morfs=None,
directory=None,
ignore_errors=None,
omit=None,
include=None,
extra_css=None,
title=None):
"""Generate an HTML report.
The HTML is written to `directory`. The file "index.html" is the
overview starting point, with links to more detailed pages for
individual modules.
`extra_css` is a path to a file of other CSS to apply on the page.
It will be copied into the HTML directory.
`title` is a text string (not HTML) to use as the title of the HTML
report.
See :meth:`report` for other arguments.
Returns a float, the total percentage covered.
"""
self.get_data()
self.config.from_args(
ignore_errors=ignore_errors,
omit=omit,
include=include,
html_dir=directory,
extra_css=extra_css,
html_title=title,
)
reporter = HtmlReporter(self, self.config)
return reporter.report(morfs)
def xml_report(
self,
morfs=None,
outfile=None,
ignore_errors=None,
omit=None,
include=None,
):
"""Generate an XML report of coverage results.
The report is compatible with Cobertura reports.
Each module in `morfs` is included in the report. `outfile` is the
path to write the file to, "-" will write to stdout.
See :meth:`report` for other arguments.
Returns a float, the total percentage covered.
"""
self.get_data()
self.config.from_args(
ignore_errors=ignore_errors,
omit=omit,
include=include,
xml_output=outfile,
)
file_to_close = None
delete_file = False
if self.config.xml_output:
if self.config.xml_output == '-':
outfile = sys.stdout
else:
# Ensure that the output directory is created; done here
# because this report pre-opens the output file.
# HTMLReport does this using the Report plumbing because
# its task is more complex, being multiple files.
output_dir = os.path.dirname(self.config.xml_output)
if output_dir and not os.path.isdir(output_dir):
os.makedirs(output_dir)
outfile = open(self.config.xml_output, "w")
file_to_close = outfile
try:
reporter = XmlReporter(self, self.config)
return reporter.report(morfs, outfile=outfile)
except CoverageException:
delete_file = True
raise
finally:
if file_to_close:
file_to_close.close()
if delete_file:
file_be_gone(self.config.xml_output)
def sys_info(self):
"""Return a list of (key, value) pairs showing internal information."""
import coverage as covmod
self._init()
ft_plugins = []
for ft in self.plugins.file_tracers:
ft_name = ft._coverage_plugin_name
if not ft._coverage_enabled:
ft_name += " (disabled)"
ft_plugins.append(ft_name)
info = [
('version', covmod.__version__),
('coverage', covmod.__file__),
('cover_dirs', self.cover_dirs),
('pylib_dirs', self.pylib_dirs),
('tracer', self.collector.tracer_name()),
('plugins.file_tracers', ft_plugins),
('config_files', self.config.attempted_config_files),
('configs_read', self.config.config_files),
('data_path', self.data_files.filename),
('python', sys.version.replace('\n', '')),
('platform', platform.platform()),
('implementation', platform.python_implementation()),
('executable', sys.executable),
('cwd', os.getcwd()),
('path', sys.path),
('environment',
sorted(("%s = %s" % (k, v)) for k, v in iitems(os.environ)
if k.startswith(("COV", "PY")))),
('command_line', " ".join(getattr(sys, 'argv', ['???']))),
]
matcher_names = [
'source_match',
'source_pkgs_match',
'include_match',
'omit_match',
'cover_match',
'pylib_match',
]
for matcher_name in matcher_names:
matcher = getattr(self, matcher_name)
if matcher:
matcher_info = matcher.info()
else:
matcher_info = '-none-'
info.append((matcher_name, matcher_info))
return info
class InOrOut:
"""Machinery for determining what files to measure."""
def __init__(self, warn, debug):
self.warn = warn
self.debug = debug
# The matchers for should_trace.
self.source_match = None
self.source_pkgs_match = None
self.pylib_paths = self.cover_paths = self.third_paths = None
self.pylib_match = self.cover_match = self.third_match = None
self.include_match = self.omit_match = None
self.plugins = []
self.disp_class = FileDisposition
# The source argument can be directories or package names.
self.source = []
self.source_pkgs = []
self.source_pkgs_unmatched = []
self.omit = self.include = None
# Is the source inside a third-party area?
self.source_in_third = False
def configure(self, config):
"""Apply the configuration to get ready for decision-time."""
self.source_pkgs.extend(config.source_pkgs)
for src in config.source or []:
if os.path.isdir(src):
self.source.append(canonical_filename(src))
else:
self.source_pkgs.append(src)
self.source_pkgs_unmatched = self.source_pkgs[:]
self.omit = prep_patterns(config.run_omit)
self.include = prep_patterns(config.run_include)
# The directories for files considered "installed with the interpreter".
self.pylib_paths = set()
if not config.cover_pylib:
add_stdlib_paths(self.pylib_paths)
# To avoid tracing the coverage.py code itself, we skip anything
# located where we are.
self.cover_paths = set()
add_coverage_paths(self.cover_paths)
# Find where third-party packages are installed.
self.third_paths = set()
add_third_party_paths(self.third_paths)
def debug(msg):
if self.debug:
self.debug.write(msg)
# Generally useful information
debug("sys.path:" + "".join(f"\n {p}" for p in sys.path))
# Create the matchers we need for should_trace
if self.source or self.source_pkgs:
against = []
if self.source:
self.source_match = TreeMatcher(self.source, "source")
against.append(f"trees {self.source_match!r}")
if self.source_pkgs:
self.source_pkgs_match = ModuleMatcher(self.source_pkgs,
"source_pkgs")
against.append(f"modules {self.source_pkgs_match!r}")
debug("Source matching against " + " and ".join(against))
else:
if self.pylib_paths:
self.pylib_match = TreeMatcher(self.pylib_paths, "pylib")
debug(f"Python stdlib matching: {self.pylib_match!r}")
if self.include:
self.include_match = FnmatchMatcher(self.include, "include")
debug(f"Include matching: {self.include_match!r}")
if self.omit:
self.omit_match = FnmatchMatcher(self.omit, "omit")
debug(f"Omit matching: {self.omit_match!r}")
self.cover_match = TreeMatcher(self.cover_paths, "coverage")
debug(f"Coverage code matching: {self.cover_match!r}")
self.third_match = TreeMatcher(self.third_paths, "third")
debug(f"Third-party lib matching: {self.third_match!r}")
# Check if the source we want to measure has been installed as a
# third-party package.
with sys_modules_saved():
for pkg in self.source_pkgs:
try:
modfile, path = file_and_path_for_module(pkg)
debug(f"Imported source package {pkg!r} as {modfile!r}")
except CoverageException as exc:
debug(f"Couldn't import source package {pkg!r}: {exc}")
continue
if modfile:
if self.third_match.match(modfile):
debug(
f"Source is in third-party because of source_pkg {pkg!r} at {modfile!r}"
)
self.source_in_third = True
else:
for pathdir in path:
if self.third_match.match(pathdir):
debug(
f"Source is in third-party because of {pkg!r} path directory "
+ f"at {pathdir!r}")
self.source_in_third = True
for src in self.source:
if self.third_match.match(src):
debug(
f"Source is in third-party because of source directory {src!r}"
)
self.source_in_third = True
def should_trace(self, filename, frame=None):
"""Decide whether to trace execution in `filename`, with a reason.
This function is called from the trace function. As each new file name
is encountered, this function determines whether it is traced or not.
Returns a FileDisposition object.
"""
original_filename = filename
disp = disposition_init(self.disp_class, filename)
def nope(disp, reason):
"""Simple helper to make it easy to return NO."""
disp.trace = False
disp.reason = reason
return disp
if original_filename.startswith('<'):
return nope(disp, "not a real original file name")
if frame is not None:
# Compiled Python files have two file names: frame.f_code.co_filename is
# the file name at the time the .pyc was compiled. The second name is
# __file__, which is where the .pyc was actually loaded from. Since
# .pyc files can be moved after compilation (for example, by being
# installed), we look for __file__ in the frame and prefer it to the
# co_filename value.
dunder_file = frame.f_globals and frame.f_globals.get('__file__')
if dunder_file:
filename = source_for_file(dunder_file)
if original_filename and not original_filename.startswith('<'):
orig = os.path.basename(original_filename)
if orig != os.path.basename(filename):
# Files shouldn't be renamed when moved. This happens when
# exec'ing code. If it seems like something is wrong with
# the frame's file name, then just use the original.
filename = original_filename
if not filename:
# Empty string is pretty useless.
return nope(disp, "empty string isn't a file name")
if filename.startswith('memory:'):
return nope(disp, "memory isn't traceable")
if filename.startswith('<'):
# Lots of non-file execution is represented with artificial
# file names like "<string>", "<doctest readme.txt[0]>", or
# "<exec_function>". Don't ever trace these executions, since we
# can't do anything with the data later anyway.
return nope(disp, "not a real file name")
# Jython reports the .class file to the tracer, use the source file.
if filename.endswith("$py.class"):
filename = filename[:-9] + ".py"
canonical = canonical_filename(filename)
disp.canonical_filename = canonical
# Try the plugins, see if they have an opinion about the file.
plugin = None
for plugin in self.plugins.file_tracers:
if not plugin._coverage_enabled:
continue
try:
file_tracer = plugin.file_tracer(canonical)
if file_tracer is not None:
file_tracer._coverage_plugin = plugin
disp.trace = True
disp.file_tracer = file_tracer
if file_tracer.has_dynamic_source_filename():
disp.has_dynamic_filename = True
else:
disp.source_filename = canonical_filename(
file_tracer.source_filename())
break
except Exception:
plugin_name = plugin._coverage_plugin_name
tb = traceback.format_exc()
self.warn(
f"Disabling plug-in {plugin_name!r} due to an exception:\n{tb}"
)
plugin._coverage_enabled = False
continue
else:
# No plugin wanted it: it's Python.
disp.trace = True
disp.source_filename = canonical
if not disp.has_dynamic_filename:
if not disp.source_filename:
raise PluginError(
f"Plugin {plugin!r} didn't set source_filename for '{disp.original_filename}'"
)
reason = self.check_include_omit_etc(disp.source_filename, frame)
if reason:
nope(disp, reason)
return disp
def check_include_omit_etc(self, filename, frame):
"""Check a file name against the include, omit, etc, rules.
Returns a string or None. String means, don't trace, and is the reason
why. None means no reason found to not trace.
"""
modulename = name_for_module(filename, frame)
# If the user specified source or include, then that's authoritative
# about the outer bound of what to measure and we don't have to apply
# any canned exclusions. If they didn't, then we have to exclude the
# stdlib and coverage.py directories.
if self.source_match or self.source_pkgs_match:
extra = ""
ok = False
if self.source_pkgs_match:
if self.source_pkgs_match.match(modulename):
ok = True
if modulename in self.source_pkgs_unmatched:
self.source_pkgs_unmatched.remove(modulename)
else:
extra = f"module {modulename!r} "
if not ok and self.source_match:
if self.source_match.match(filename):
ok = True
if not ok:
return extra + "falls outside the --source spec"
if not self.source_in_third:
if self.third_match.match(filename):
return "inside --source, but is third-party"
elif self.include_match:
if not self.include_match.match(filename):
return "falls outside the --include trees"
else:
# We exclude the coverage.py code itself, since a little of it
# will be measured otherwise.
if self.cover_match.match(filename):
return "is part of coverage.py"
# If we aren't supposed to trace installed code, then check if this
# is near the Python standard library and skip it if so.
if self.pylib_match and self.pylib_match.match(filename):
return "is in the stdlib"
# Exclude anything in the third-party installation areas.
if self.third_match.match(filename):
return "is a third-party module"
# Check the file against the omit pattern.
if self.omit_match and self.omit_match.match(filename):
return "is inside an --omit pattern"
# No point tracing a file we can't later write to SQLite.
try:
filename.encode("utf-8")
except UnicodeEncodeError:
return "non-encodable filename"
# No reason found to skip this file.
return None
def warn_conflicting_settings(self):
"""Warn if there are settings that conflict."""
if self.include:
if self.source or self.source_pkgs:
self.warn("--include is ignored because --source is set",
slug="include-ignored")
def warn_already_imported_files(self):
"""Warn if files have already been imported that we will be measuring."""
if self.include or self.source or self.source_pkgs:
warned = set()
for mod in list(sys.modules.values()):
filename = getattr(mod, "__file__", None)
if filename is None:
continue
if filename in warned:
continue
if len(getattr(mod, "__path__", ())) > 1:
# A namespace package, which confuses this code, so ignore it.
continue
disp = self.should_trace(filename)
if disp.has_dynamic_filename:
# A plugin with dynamic filenames: the Python file
# shouldn't cause a warning, since it won't be the subject
# of tracing anyway.
continue
if disp.trace:
msg = f"Already imported a file that will be measured: {filename}"
self.warn(msg, slug="already-imported")
warned.add(filename)
elif self.debug and self.debug.should('trace'):
self.debug.write(
"Didn't trace already imported file {!r}: {}".format(
disp.original_filename, disp.reason))
def warn_unimported_source(self):
"""Warn about source packages that were of interest, but never traced."""
for pkg in self.source_pkgs_unmatched:
self._warn_about_unmeasured_code(pkg)
def _warn_about_unmeasured_code(self, pkg):
"""Warn about a package or module that we never traced.
`pkg` is a string, the name of the package or module.
"""
mod = sys.modules.get(pkg)
if mod is None:
self.warn(f"Module {pkg} was never imported.",
slug="module-not-imported")
return
if module_is_namespace(mod):
# A namespace package. It's OK for this not to have been traced,
# since there is no code directly in it.
return
if not module_has_file(mod):
self.warn(f"Module {pkg} has no Python source.",
slug="module-not-python")
return
# The module was in sys.modules, and seems like a module with code, but
# we never measured it. I guess that means it was imported before
# coverage even started.
msg = f"Module {pkg} was previously imported, but not measured"
self.warn(msg, slug="module-not-measured")
def find_possibly_unexecuted_files(self):
"""Find files in the areas of interest that might be untraced.
Yields pairs: file path, and responsible plug-in name.
"""
for pkg in self.source_pkgs:
if (not pkg in sys.modules
or not module_has_file(sys.modules[pkg])):
continue
pkg_file = source_for_file(sys.modules[pkg].__file__)
yield from self._find_executable_files(canonical_path(pkg_file))
for src in self.source:
yield from self._find_executable_files(src)
def _find_plugin_files(self, src_dir):
"""Get executable files from the plugins."""
for plugin in self.plugins.file_tracers:
for x_file in plugin.find_executable_files(src_dir):
yield x_file, plugin._coverage_plugin_name
def _find_executable_files(self, src_dir):
"""Find executable files in `src_dir`.
Search for files in `src_dir` that can be executed because they
are probably importable. Don't include ones that have been omitted
by the configuration.
Yield the file path, and the plugin name that handles the file.
"""
py_files = ((py_file, None) for py_file in find_python_files(src_dir))
plugin_files = self._find_plugin_files(src_dir)
for file_path, plugin_name in itertools.chain(py_files, plugin_files):
file_path = canonical_filename(file_path)
if self.omit_match and self.omit_match.match(file_path):
# Turns out this file was omitted, so don't pull it back
# in as unexecuted.
continue
yield file_path, plugin_name
def sys_info(self):
"""Our information for Coverage.sys_info.
Returns a list of (key, value) pairs.
"""
info = [
("coverage_paths", self.cover_paths),
("stdlib_paths", self.pylib_paths),
("third_party_paths", self.third_paths),
]
matcher_names = [
'source_match',
'source_pkgs_match',
'include_match',
'omit_match',
'cover_match',
'pylib_match',
'third_match',
]
for matcher_name in matcher_names:
matcher = getattr(self, matcher_name)
if matcher:
matcher_info = matcher.info()
else:
matcher_info = '-none-'
info.append((matcher_name, matcher_info))
return info
class InOrOut(object):
"""Machinery for determining what files to measure."""
def __init__(self, warn):
self.warn = warn
# The matchers for should_trace.
self.source_match = None
self.source_pkgs_match = None
self.pylib_paths = self.cover_paths = None
self.pylib_match = self.cover_match = None
self.include_match = self.omit_match = None
self.plugins = []
self.disp_class = FileDisposition
# The source argument can be directories or package names.
self.source = []
self.source_pkgs = []
self.source_pkgs_unmatched = []
self.omit = self.include = None
def configure(self, config):
"""Apply the configuration to get ready for decision-time."""
for src in config.source or []:
if os.path.isdir(src):
self.source.append(canonical_filename(src))
else:
self.source_pkgs.append(src)
self.source_pkgs_unmatched = self.source_pkgs[:]
self.omit = prep_patterns(config.run_omit)
self.include = prep_patterns(config.run_include)
# The directories for files considered "installed with the interpreter".
self.pylib_paths = set()
if not config.cover_pylib:
# Look at where some standard modules are located. That's the
# indication for "installed with the interpreter". In some
# environments (virtualenv, for example), these modules may be
# spread across a few locations. Look at all the candidate modules
# we've imported, and take all the different ones.
for m in (atexit, inspect, os, platform, _pypy_irc_topic, re, _structseq, traceback):
if m is not None and hasattr(m, "__file__"):
self.pylib_paths.add(canonical_path(m, directory=True))
if _structseq and not hasattr(_structseq, '__file__'):
# PyPy 2.4 has no __file__ in the builtin modules, but the code
# objects still have the file names. So dig into one to find
# the path to exclude. The "filename" might be synthetic,
# don't be fooled by those.
structseq_new = _structseq.structseq_new
try:
structseq_file = structseq_new.func_code.co_filename
except AttributeError:
structseq_file = structseq_new.__code__.co_filename
if not structseq_file.startswith("<"):
self.pylib_paths.add(canonical_path(structseq_file))
# To avoid tracing the coverage.py code itself, we skip anything
# located where we are.
self.cover_paths = [canonical_path(__file__, directory=True)]
if env.TESTING:
# Don't include our own test code.
self.cover_paths.append(os.path.join(self.cover_paths[0], "tests"))
# When testing, we use PyContracts, which should be considered
# part of coverage.py, and it uses six. Exclude those directories
# just as we exclude ourselves.
import contracts
import six
for mod in [contracts, six]:
self.cover_paths.append(canonical_path(mod))
# Create the matchers we need for should_trace
if self.source or self.source_pkgs:
self.source_match = TreeMatcher(self.source)
self.source_pkgs_match = ModuleMatcher(self.source_pkgs)
else:
if self.cover_paths:
self.cover_match = TreeMatcher(self.cover_paths)
if self.pylib_paths:
self.pylib_match = TreeMatcher(self.pylib_paths)
if self.include:
self.include_match = FnmatchMatcher(self.include)
if self.omit:
self.omit_match = FnmatchMatcher(self.omit)
def should_trace(self, filename, frame=None):
"""Decide whether to trace execution in `filename`, with a reason.
This function is called from the trace function. As each new file name
is encountered, this function determines whether it is traced or not.
Returns a FileDisposition object.
"""
original_filename = filename
disp = disposition_init(self.disp_class, filename)
def nope(disp, reason):
"""Simple helper to make it easy to return NO."""
disp.trace = False
disp.reason = reason
return disp
if frame is not None:
# Compiled Python files have two file names: frame.f_code.co_filename is
# the file name at the time the .pyc was compiled. The second name is
# __file__, which is where the .pyc was actually loaded from. Since
# .pyc files can be moved after compilation (for example, by being
# installed), we look for __file__ in the frame and prefer it to the
# co_filename value.
dunder_file = frame.f_globals and frame.f_globals.get('__file__')
if dunder_file:
filename = source_for_file(dunder_file)
if original_filename and not original_filename.startswith('<'):
orig = os.path.basename(original_filename)
if orig != os.path.basename(filename):
# Files shouldn't be renamed when moved. This happens when
# exec'ing code. If it seems like something is wrong with
# the frame's file name, then just use the original.
filename = original_filename
if not filename:
# Empty string is pretty useless.
return nope(disp, "empty string isn't a file name")
if filename.startswith('memory:'):
return nope(disp, "memory isn't traceable")
if filename.startswith('<'):
# Lots of non-file execution is represented with artificial
# file names like "<string>", "<doctest readme.txt[0]>", or
# "<exec_function>". Don't ever trace these executions, since we
# can't do anything with the data later anyway.
return nope(disp, "not a real file name")
# pyexpat does a dumb thing, calling the trace function explicitly from
# C code with a C file name.
if re.search(r"[/\\]Modules[/\\]pyexpat.c", filename):
return nope(disp, "pyexpat lies about itself")
# Jython reports the .class file to the tracer, use the source file.
if filename.endswith("$py.class"):
filename = filename[:-9] + ".py"
canonical = canonical_filename(filename)
disp.canonical_filename = canonical
# Try the plugins, see if they have an opinion about the file.
plugin = None
for plugin in self.plugins.file_tracers:
if not plugin._coverage_enabled:
continue
try:
file_tracer = plugin.file_tracer(canonical)
if file_tracer is not None:
file_tracer._coverage_plugin = plugin
disp.trace = True
disp.file_tracer = file_tracer
if file_tracer.has_dynamic_source_filename():
disp.has_dynamic_filename = True
else:
disp.source_filename = canonical_filename(
file_tracer.source_filename()
)
break
except Exception:
self.warn(
"Disabling plug-in %r due to an exception:" % (plugin._coverage_plugin_name)
)
traceback.print_exc()
plugin._coverage_enabled = False
continue
else:
# No plugin wanted it: it's Python.
disp.trace = True
disp.source_filename = canonical
if not disp.has_dynamic_filename:
if not disp.source_filename:
raise CoverageException(
"Plugin %r didn't set source_filename for %r" %
(plugin, disp.original_filename)
)
reason = self.check_include_omit_etc(disp.source_filename, frame)
if reason:
nope(disp, reason)
return disp
def check_include_omit_etc(self, filename, frame):
"""Check a file name against the include, omit, etc, rules.
Returns a string or None. String means, don't trace, and is the reason
why. None means no reason found to not trace.
"""
modulename = name_for_module(filename, frame)
# If the user specified source or include, then that's authoritative
# about the outer bound of what to measure and we don't have to apply
# any canned exclusions. If they didn't, then we have to exclude the
# stdlib and coverage.py directories.
if self.source_match:
if self.source_pkgs_match.match(modulename):
if modulename in self.source_pkgs_unmatched:
self.source_pkgs_unmatched.remove(modulename)
elif not self.source_match.match(filename):
return "falls outside the --source trees"
elif self.include_match:
if not self.include_match.match(filename):
return "falls outside the --include trees"
else:
# If we aren't supposed to trace installed code, then check if this
# is near the Python standard library and skip it if so.
if self.pylib_match and self.pylib_match.match(filename):
return "is in the stdlib"
# We exclude the coverage.py code itself, since a little of it
# will be measured otherwise.
if self.cover_match and self.cover_match.match(filename):
return "is part of coverage.py"
# Check the file against the omit pattern.
if self.omit_match and self.omit_match.match(filename):
return "is inside an --omit pattern"
# No reason found to skip this file.
return None
def warn_conflicting_settings(self):
"""Warn if there are settings that conflict."""
if self.include:
if self.source or self.source_pkgs:
self.warn("--include is ignored because --source is set", slug="include-ignored")
def warn_already_imported_files(self):
"""Warn if files have already been imported that we will be measuring."""
if self.include or self.source or self.source_pkgs:
warned = set()
for mod in list(sys.modules.values()):
filename = getattr(mod, "__file__", None)
if filename is None:
continue
if filename in warned:
continue
disp = self.should_trace(filename)
if disp.trace:
msg = "Already imported a file that will be measured: {0}".format(filename)
self.warn(msg, slug="already-imported")
warned.add(filename)
def warn_unimported_source(self):
"""Warn about source packages that were of interest, but never traced."""
for pkg in self.source_pkgs_unmatched:
self._warn_about_unmeasured_code(pkg)
def _warn_about_unmeasured_code(self, pkg):
"""Warn about a package or module that we never traced.
`pkg` is a string, the name of the package or module.
"""
mod = sys.modules.get(pkg)
if mod is None:
self.warn("Module %s was never imported." % pkg, slug="module-not-imported")
return
if module_is_namespace(mod):
# A namespace package. It's OK for this not to have been traced,
# since there is no code directly in it.
return
if not module_has_file(mod):
self.warn("Module %s has no Python source." % pkg, slug="module-not-python")
return
# The module was in sys.modules, and seems like a module with code, but
# we never measured it. I guess that means it was imported before
# coverage even started.
self.warn(
"Module %s was previously imported, but not measured" % pkg,
slug="module-not-measured",
)
def find_unexecuted_files(self):
"""Find files in the areas of interest that weren't traced.
Yields pairs: file path, and responsible plug-in name.
"""
for pkg in self.source_pkgs:
if (not pkg in sys.modules or
not module_has_file(sys.modules[pkg])):
continue
pkg_file = source_for_file(sys.modules[pkg].__file__)
for ret in self._find_unexecuted_files(canonical_path(pkg_file)):
yield ret
for src in self.source:
for ret in self._find_unexecuted_files(src):
yield ret
def _find_plugin_files(self, src_dir):
"""Get executable files from the plugins."""
for plugin in self.plugins.file_tracers:
for x_file in plugin.find_executable_files(src_dir):
yield x_file, plugin._coverage_plugin_name
def _find_unexecuted_files(self, src_dir):
"""Find unexecuted files in `src_dir`.
Search for files in `src_dir` that are probably importable,
and add them as unexecuted files in `self.data`.
"""
py_files = ((py_file, None) for py_file in find_python_files(src_dir))
plugin_files = self._find_plugin_files(src_dir)
for file_path, plugin_name in itertools.chain(py_files, plugin_files):
file_path = canonical_filename(file_path)
if self.omit_match and self.omit_match.match(file_path):
# Turns out this file was omitted, so don't pull it back
# in as unexecuted.
continue
yield file_path, plugin_name
def sys_info(self):
"""Our information for Coverage.sys_info.
Returns a list of (key, value) pairs.
"""
info = [
('cover_paths', self.cover_paths),
('pylib_paths', self.pylib_paths),
]
matcher_names = [
'source_match', 'source_pkgs_match',
'include_match', 'omit_match',
'cover_match', 'pylib_match',
]
for matcher_name in matcher_names:
matcher = getattr(self, matcher_name)
if matcher:
matcher_info = matcher.info()
else:
matcher_info = '-none-'
info.append((matcher_name, matcher_info))
return info