Anforderungen  |   Konzepte  |   Entwurf  |   Entwicklung  |   Qualitätssicherung  |   Lebenszyklus  |   Steuerung
 
 
 
 


Quelle  manifest.py   Sprache: Python

 
# Library for JSTest manifests.
#
# This includes classes for representing and parsing JS manifests.

import io
import os
import posixpath
import re
import sys
from subprocess import PIPE, Popen

import six

from .remote import init_device
from .tests import RefTestCase


def split_path_into_dirs(path):
    dirs = [path]

    while True:
        path, tail = os.path.split(path)
        if not tail:
            break
        dirs.append(path)
    return dirs


class XULInfo:
    def __init__(self, abi, os, isdebug):
        self.abi = abi
        self.os = os
        self.isdebug = isdebug

    def as_js(self):
        """Return JS that when executed sets up variables so that JS expression
        predicates on XUL build info evaluate properly."""

        return (
            "var winWidget = {};"
            "var gtkWidget = {};"
            "var cocoaWidget = {};"
            "var is64Bit = {};"
            "var xulRuntime = {{ shell: true }};"
            "var release_or_beta = getBuildConfiguration('release_or_beta');"
            "var isDebugBuild={}; var Android={}; ".format(
                str(self.os == "WINNT").lower(),
                str(self.os == "Darwin").lower(),
                str(self.os == "Linux").lower(),
                str("x86-" not in self.abi).lower(),
                str(self.isdebug).lower(),
                str(self.os == "Android").lower(),
            )
        )

    @classmethod
    def create(cls, jsdir):
        """Create a XULInfo based on the current platform's characteristics."""

        # Our strategy is to find the autoconf.mk generated for the build and
        # read the values from there.

        # Find config/autoconf.mk.
        dirs = split_path_into_dirs(os.getcwd()) + split_path_into_dirs(jsdir)

        path = None
        for dir in dirs:
            _path = posixpath.join(dir, "config""autoconf.mk")
            if os.path.isfile(_path):
                path = _path
                break

        if path is None:
            print(
                "Can't find config/autoconf.mk on a directory containing"
                " the JS shell (searched from {})".format(jsdir)
            )
            sys.exit(1)

        # Read the values.
        val_re = re.compile(r"(TARGET_XPCOM_ABI|OS_TARGET|MOZ_DEBUG)\s*=\s*(.*)")
        kw = {"isdebug"False}
        for line in io.open(path, encoding="utf-8"):
            m = val_re.match(line)
            if m:
                key, val = m.groups()
                val = val.rstrip()
                if key == "TARGET_XPCOM_ABI":
                    kw["abi"] = val
                if key == "OS_TARGET":
                    kw["os"] = val
                if key == "MOZ_DEBUG":
                    kw["isdebug"] = val == "1"
        return cls(**kw)


class XULInfoTester:
    def __init__(self, xulinfo, options, js_args):
        self.js_prologue = xulinfo.as_js()
        self.js_bin = options.js_shell
        self.js_args = js_args
        # options here are the command line options
        self.options = options
        # Maps JS expr to evaluation result.
        self.cache = {}

        if not self.options.remote:
            return
        self.device = init_device(options)
        self.js_bin = posixpath.join(options.remote_test_root, "bin""js")

    def test(self, cond, options=[]):
        if self.options.remote:
            return self._test_remote(cond, options=options)
        return self._test_local(cond, options=options)

    def _test_remote(self, cond, options=[]):
        from mozdevice import ADBDevice, ADBProcessError

        ans = self.cache.get(cond, None)
        if ans is not None:
            return ans

        env = {
            "LD_LIBRARY_PATH": posixpath.join(self.options.remote_test_root, "bin"),
        }

        cmd = (
            [self.js_bin]
            + self.js_args
            + options
            + [
                # run in safe configuration, since it is hard to debug
                # crashes when running code here. In particular, msan will
                # error out if the jit is active.
                "--no-baseline",
                "--no-blinterp",
                "-e",
                self.js_prologue,
                "-e",
                "print(!!({}))".format(cond),
            ]
        )
        cmd = ADBDevice._escape_command_line(cmd)
        try:
            # Allow ADBError or ADBTimeoutError to terminate the test run,
            # but handle ADBProcessError in order to support the use of
            # non-zero exit codes in the JavaScript shell tests.
            out = self.device.shell_output(
                cmd, env=env, cwd=self.options.remote_test_root, timeout=None
            )
            err = ""
        except ADBProcessError as e:
            out = ""
            err = str(e.adb_process.stdout)

        if out == "true":
            ans = True
        elif out == "false":
            ans = False
        else:
            raise Exception(
                "Failed to test XUL condition {!r};"
                " output was {!r}, stderr was {!r}".format(cond, out, err)
            )
        self.cache[cond] = ans
        return ans

    def _test_local(self, cond, options=[]):
        """Test a XUL predicate condition against this local info."""
        ans = self.cache.get(cond, None)
        if ans is None:
            cmd = (
                [self.js_bin]
                + self.js_args
                + options
                + [
                    # run in safe configuration, since it is hard to debug
                    # crashes when running code here. In particular, msan will
                    # error out if the jit is active.
                    "--no-baseline",
                    "--no-blinterp",
                    "-e",
                    self.js_prologue,
                    "-e",
                    "print(!!({}))".format(cond),
                ]
            )
            p = Popen(
                cmd, stdin=PIPE, stdout=PIPE, stderr=PIPE, universal_newlines=True
            )
            out, err = p.communicate()
            if out in ("true\n""true\r\n"):
                ans = True
            elif out in ("false\n""false\r\n"):
                ans = False
            else:
                raise Exception(
                    "Failed to test XUL condition {!r};"
                    " output was {!r}, stderr was {!r}".format(cond, out, err)
                )
            self.cache[cond] = ans
        return ans


class NullXULInfoTester:
    """Can be used to parse manifests without a JS shell."""

    def test(self, cond, options=[]):
        return False


def _parse_one(testcase, terms, xul_tester):
    pos = 0
    parts = terms.split()
    while pos < len(parts):
        if parts[pos] == "fails":
            testcase.expect = False
            pos += 1
        elif parts[pos] == "skip":
            testcase.expect = testcase.enable = False
            pos += 1
        elif parts[pos] == "random":
            testcase.random = True
            pos += 1
        elif parts[pos].startswith("shell-option("):
            # This directive adds an extra option to pass to the shell.
            option = parts[pos][len("shell-option(") : -1]
            testcase.options.append(option)
            pos += 1
        elif parts[pos].startswith("fails-if"):
            cond = parts[pos][len("fails-if(") : -1]
            if xul_tester.test(cond, testcase.options):
                testcase.expect = False
            pos += 1
        elif parts[pos].startswith("asserts-if"):
            # This directive means we may flunk some number of
            # NS_ASSERTIONs in the browser. For the shell, ignore it.
            pos += 1
        elif parts[pos].startswith("skip-if"):
            cond = parts[pos][len("skip-if(") : -1]
            if xul_tester.test(cond, testcase.options):
                testcase.expect = testcase.enable = False
            pos += 1
        elif parts[pos].startswith("ignore-flag"):
            flag = parts[pos][len("ignore-flag(") : -1]
            testcase.ignoredflags.append(flag)
            pos += 1
        elif parts[pos].startswith("random-if"):
            cond = parts[pos][len("random-if(") : -1]
            if xul_tester.test(cond, testcase.options):
                testcase.random = True
            pos += 1
        elif parts[pos] == "slow":
            testcase.slow = True
            pos += 1
        elif parts[pos].startswith("slow-if"):
            cond = parts[pos][len("slow-if(") : -1]
            if xul_tester.test(cond, testcase.options):
                testcase.slow = True
            pos += 1
        elif parts[pos] == "silentfail":
            # silentfails use tons of memory, and Darwin doesn't support ulimit.
            if xul_tester.test("cocoaWidget", testcase.options):
                testcase.expect = testcase.enable = False
            pos += 1
        elif parts[pos].startswith("error:"):
            # This directive allows to specify an error type.
            (_, _, errortype) = parts[pos].partition(":")
            testcase.error = errortype
            pos += 1
        elif parts[pos] == "module":
            # This directive marks the test as module code.
            testcase.is_module = True
            pos += 1
        elif parts[pos] == "test262-raw":
            testcase.is_test262_raw = True
            pos += 1
        elif parts[pos] == "async":
            # This directive marks the test as async.
            testcase.is_async = True
            pos += 1
        else:
            print('warning: invalid manifest line element "{}"'.format(parts[pos]))
            pos += 1


def _build_manifest_script_entry(script_name, test):
    line = []
    properties = []
    if test.terms:
        # Remove jsreftest internal terms.
        terms = " ".join(
            [
                term
                for term in test.terms.split()
                if not (
                    term == "module"
                    or term == "async"
                    or term.startswith("error:")
                    or term.startswith("ignore-flag(")
                    or term.startswith("shell-option(")
                    or term == "test262-raw"
                )
            ]
        )
        if terms:
            line.append(terms)
    if test.error:
        properties.append("error=" + test.error)
    if test.is_module:
        properties.append("module")
    if test.is_async:
        properties.append("async")
    if test.is_test262_raw:
        properties.append("test262_raw")
    line.append("script")
    script = script_name
    if properties:
        script = ";".join([script] + properties)
    line.append(script)
    if test.comment:
        line.append("#")
        line.append(test.comment)
    return " ".join(line)


def _map_prefixes_left(test_gen):
    """
    Splits tests into a dictionary keyed on the first component of the test
    path, aggregating tests with a common base path into a list.
    """
    byprefix = {}
    for t in test_gen:
        left, sep, remainder = t.path.partition(os.sep)
        if left not in byprefix:
            byprefix[left] = []
        if remainder:
            t.path = remainder
        byprefix[left].append(t)
    return byprefix


def _emit_manifest_at(location, relative, test_gen, depth):
    """
    location  - str: absolute path where we want to write the manifest
    relative  - str: relative path from topmost manifest directory to current
    test_gen  - (str): generator of all test paths and directorys
    depth     - int: number of dirs we are below the topmost manifest dir
    """
    manifests = _map_prefixes_left(test_gen)

    filename = os.path.join(location, "jstests.list")
    manifest = []
    numTestFiles = 0
    for k, test_list in manifests.items():
        fullpath = os.path.join(location, k)
        if os.path.isdir(fullpath):
            manifest.append("include " + k + "/jstests.list")
            relpath = os.path.join(relative, k)
            _emit_manifest_at(fullpath, relpath, test_list, depth + 1)
        else:
            numTestFiles += 1
            assert len(test_list) == 1, test_list
            line = _build_manifest_script_entry(k, test_list[0])
            manifest.append(line)

    # Always present our manifest in sorted order.
    manifest.sort()

    # If we have tests, we have to set the url-prefix so reftest can find them.
    if numTestFiles > 0:
        manifest = [
            "url-prefix {}jsreftest.html?test={}/".format("../" * depth, relative)
        ] + manifest

    fp = io.open(filename, "w", encoding="utf-8", newline="\n")
    try:
        fp.write("\n".join(manifest) + "\n")
    finally:
        fp.close()


def make_manifests(location, test_gen):
    _emit_manifest_at(location, "", test_gen, 0)


def _find_all_js_files(location):
    for root, dirs, files in os.walk(location):
        root = root[len(location) + 1 :]
        for fn in files:
            if fn.endswith(".js"):
                yield root, fn


# The pattern for test header lines.
TEST_HEADER_PATTERN = r"""
# Ignore any space before the tag.
\s*

# The reftest tag is enclosed in pipes.
\|(?P<tag>.*?)\|

# Ignore any space before the options.
\s*

# Accept some options.
(?P<options>.*?)

# Ignore space before the comments.
\s*

# Accept an optional comment starting with "--".
(?:
  # Unless "--" is directly preceded by "(".
  (?<!\()
  --

  # Ignore more space.
  \s*

  # The actual comment.
  (?P<comment>.*)
)?
"""


TEST_HEADER_PATTERN_INLINE = re.compile(
    r"""
# Start a single line comment
//
"""
    + TEST_HEADER_PATTERN
    + r"""
# Match the end of line.
$
""",
    re.VERBOSE,
)
TEST_HEADER_PATTERN_MULTI = re.compile(
    r"""
# Start a multi line comment
/\*
"""
    + TEST_HEADER_PATTERN
    + r"""
# Match the end of comment.
\*/
""",
    re.VERBOSE,
)


def _append_terms_and_comment(testcase, terms, comment):
    if testcase.terms is None:
        testcase.terms = terms
    else:
        testcase.terms += " " + terms

    if testcase.comment is None:
        testcase.comment = comment
    elif comment:
        testcase.comment += "; " + comment


def _parse_test_header(fullpath, testcase, xul_tester):
    """
    This looks a bit weird.  The reason is that it needs to be efficient, since
    it has to be done on every test
    """
    if six.PY3:
        fp = open(fullpath, encoding="utf-8")
    else:
        fp = open(fullpath)
    try:
        buf = fp.read(512)
    finally:
        fp.close()

    # Bail early if we do not start with a single comment.
    if not buf.startswith("//"):
        return

    # Extract the token.
    buf, _, _ = buf.partition("\n")
    matches = TEST_HEADER_PATTERN_INLINE.match(buf)

    if not matches:
        matches = TEST_HEADER_PATTERN_MULTI.match(buf)
        if not matches:
            return

    testcase.tag = matches.group("tag")
    _append_terms_and_comment(
        testcase, matches.group("options"), matches.group("comment")
    )
    _parse_one(testcase, matches.group("options"), xul_tester)


def _parse_external_manifest(filename, relpath):
    """
    Reads an external manifest file for test suites whose individual test cases
    can't be decorated with reftest comments.
    filename - str: name of the manifest file
    relpath - str: relative path of the directory containing the manifest
                   within the test suite
    """
    if not os.path.exists(filename):
        return []

    entries = []

    with io.open(filename, "r", encoding="utf-8"as fp:
        manifest_re = re.compile(
            r"^\s*(?P.*)\s+(?Pinclude|script)\s+(?P\S+)$"
        )
        include_re = re.compile(r"^\s*include\s+(?P\S+)$")
        for line in fp:
            line, _, comment = line.partition("#")
            line = line.strip()
            if not line:
                continue
            matches = manifest_re.match(line)
            if not matches:
                matches = include_re.match(line)
                if not matches:
                    print(
                        "warning: unrecognized line in jstests.list:"
                        " {0}".format(line)
                    )
                    continue

                include_file = matches.group("path")
                include_filename = os.path.join(os.path.dirname(filename), include_file)
                include_relpath = os.path.join(relpath, os.path.dirname(include_file))
                include_entries = _parse_external_manifest(
                    include_filename, include_relpath
                )
                entries.extend(include_entries)
                continue

            path = os.path.normpath(os.path.join(relpath, matches.group("path")))
            if matches.group("type") == "include":
                # The manifest spec wants a reference to another manifest here,
                # but we need just the directory. We do need the trailing
                # separator so we don't accidentally match other paths of which
                # this one is a prefix.
                assert path.endswith("jstests.list")
                path = path[: -len("jstests.list")]

            entries.append(
                {
                    "path": path,
                    "terms": matches.group("terms"),
                    "comment": comment.strip(),
                }
            )

    # if one directory name is a prefix of another, we want the shorter one
    # first
    entries.sort(key=lambda x: x["path"])
    return entries


def _apply_external_manifests(filename, testcase, entries, xul_tester):
    for entry in entries:
        if filename.startswith(entry["path"]):
            # The reftest spec would require combining the terms (failure types)
            # that may already be defined in the test case with the terms
            # specified in entry; for example, a skip overrides a random, which
            # overrides a fails. Since we don't necessarily know yet in which
            # environment the test cases will be run, we'd also have to
            # consider skip-if, random-if, and fails-if with as-yet unresolved
            # conditions.
            # At this point, we use external manifests only for test cases
            # that can't have their own failure type comments, so we simply
            # use the terms for the most specific path.
            _append_terms_and_comment(testcase, entry["terms"], entry["comment"])
            _parse_one(testcase, entry["terms"], xul_tester)


def _is_test_file(path_from_root, basename, filename, path_options):
    # Any file whose basename matches something in this set is ignored.
    EXCLUDED = set(
        (
            "browser.js",
            "shell.js",
            "template.js",
            "user.js",
            "js-test-driver-begin.js",
            "js-test-driver-end.js",
        )
    )

    # Skip js files in the root test directory.
    if not path_from_root:
        return False

    # Skip files that we know are not tests.
    if basename in EXCLUDED:
        return False

    if not path_options.should_run(filename):
        return False

    return True


def count_tests(location, path_options):
    count = 0
    for root, basename in _find_all_js_files(location):
        filename = os.path.join(root, basename)
        if _is_test_file(root, basename, filename, path_options):
            count += 1
    return count


def load_reftests(location, path_options, xul_tester):
    """
    Locates all tests by walking the filesystem starting at |location|.
    Uses xul_tester to evaluate any test conditions in the test header.
    Failure type and comment for a test case can come from
    - an external manifest entry for the test case,
    - an external manifest entry for a containing directory,
    - most commonly: the header of the test case itself.
    """
    manifestFile = os.path.join(location, "jstests.list")
    externalManifestEntries = _parse_external_manifest(manifestFile, "")

    for root, basename in _find_all_js_files(location):
        # Get the full path and relative location of the file.
        filename = os.path.join(root, basename)
        if not _is_test_file(root, basename, filename, path_options):
            continue

        # Skip empty files.
        fullpath = os.path.join(location, filename)

        testcase = RefTestCase(location, filename)
        _apply_external_manifests(
            filename, testcase, externalManifestEntries, xul_tester
        )
        _parse_test_header(fullpath, testcase, xul_tester)
        yield testcase

Messung V0.5
C=90 H=87 G=88

¤ Dauer der Verarbeitung: 0.29 Sekunden  (vorverarbeitet)  ¤

*© Formatika GbR, Deutschland






Wurzel

Suchen

Beweissystem der NASA

Beweissystem Isabelle

NIST Cobol Testsuite

Cephes Mathematical Library

Wiener Entwicklungsmethode

Haftungshinweis

Die Informationen auf dieser Webseite wurden nach bestem Wissen sorgfältig zusammengestellt. Es wird jedoch weder Vollständigkeit, noch Richtigkeit, noch Qualität der bereit gestellten Informationen zugesichert.

Bemerkung:

Die farbliche Syntaxdarstellung und die Messung sind noch experimentell.






                                                                                                                                                                                                                                                                                                                                                                                                     


Neuigkeiten

     Aktuelles
     Motto des Tages

Software

     Produkte
     Quellcodebibliothek

Aktivitäten

     Artikel über Sicherheit
     Anleitung zur Aktivierung von SSL

Muße

     Gedichte
     Musik
     Bilder

Jenseits des Üblichen ....
    

Besucherstatistik

Besucherstatistik

Monitoring

Montastic status badge