911 lines
31 KiB
Python
911 lines
31 KiB
Python
from __future__ import absolute_import
|
|
|
|
import cgi
|
|
import email.utils
|
|
import hashlib
|
|
import getpass
|
|
import json
|
|
import logging
|
|
import mimetypes
|
|
import os
|
|
import platform
|
|
import re
|
|
import shutil
|
|
import sys
|
|
import tempfile
|
|
|
|
try:
|
|
import ssl # noqa
|
|
HAS_TLS = True
|
|
except ImportError:
|
|
HAS_TLS = False
|
|
|
|
from pip._vendor.six.moves.urllib import parse as urllib_parse
|
|
from pip._vendor.six.moves.urllib import request as urllib_request
|
|
|
|
import pip
|
|
|
|
from pip.exceptions import InstallationError, HashMismatch
|
|
from pip.models import PyPI
|
|
from pip.utils import (splitext, rmtree, format_size, display_path,
|
|
backup_dir, ask_path_exists, unpack_file,
|
|
call_subprocess, ARCHIVE_EXTENSIONS)
|
|
from pip.utils.filesystem import check_path_owner
|
|
from pip.utils.logging import indent_log
|
|
from pip.utils.ui import DownloadProgressBar, DownloadProgressSpinner
|
|
from pip.locations import write_delete_marker_file
|
|
from pip.vcs import vcs
|
|
from pip._vendor import requests, six
|
|
from pip._vendor.requests.adapters import BaseAdapter, HTTPAdapter
|
|
from pip._vendor.requests.auth import AuthBase, HTTPBasicAuth
|
|
from pip._vendor.requests.models import Response
|
|
from pip._vendor.requests.structures import CaseInsensitiveDict
|
|
from pip._vendor.requests.packages import urllib3
|
|
from pip._vendor.cachecontrol import CacheControlAdapter
|
|
from pip._vendor.cachecontrol.caches import FileCache
|
|
from pip._vendor.lockfile import LockError
|
|
from pip._vendor.six.moves import xmlrpc_client
|
|
|
|
|
|
__all__ = ['get_file_content',
|
|
'is_url', 'url_to_path', 'path_to_url',
|
|
'is_archive_file', 'unpack_vcs_link',
|
|
'unpack_file_url', 'is_vcs_url', 'is_file_url',
|
|
'unpack_http_url', 'unpack_url']
|
|
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
|
|
def user_agent():
|
|
"""
|
|
Return a string representing the user agent.
|
|
"""
|
|
data = {
|
|
"installer": {"name": "pip", "version": pip.__version__},
|
|
"python": platform.python_version(),
|
|
"implementation": {
|
|
"name": platform.python_implementation(),
|
|
},
|
|
}
|
|
|
|
if data["implementation"]["name"] == 'CPython':
|
|
data["implementation"]["version"] = platform.python_version()
|
|
elif data["implementation"]["name"] == 'PyPy':
|
|
if sys.pypy_version_info.releaselevel == 'final':
|
|
pypy_version_info = sys.pypy_version_info[:3]
|
|
else:
|
|
pypy_version_info = sys.pypy_version_info
|
|
data["implementation"]["version"] = ".".join(
|
|
[str(x) for x in pypy_version_info]
|
|
)
|
|
elif data["implementation"]["name"] == 'Jython':
|
|
# Complete Guess
|
|
data["implementation"]["version"] = platform.python_version()
|
|
elif data["implementation"]["name"] == 'IronPython':
|
|
# Complete Guess
|
|
data["implementation"]["version"] = platform.python_version()
|
|
|
|
if sys.platform.startswith("linux"):
|
|
distro = dict(filter(
|
|
lambda x: x[1],
|
|
zip(["name", "version", "id"], platform.linux_distribution()),
|
|
))
|
|
libc = dict(filter(
|
|
lambda x: x[1],
|
|
zip(["lib", "version"], platform.libc_ver()),
|
|
))
|
|
if libc:
|
|
distro["libc"] = libc
|
|
if distro:
|
|
data["distro"] = distro
|
|
|
|
if sys.platform.startswith("darwin") and platform.mac_ver()[0]:
|
|
data["distro"] = {"name": "OS X", "version": platform.mac_ver()[0]}
|
|
|
|
if platform.system():
|
|
data.setdefault("system", {})["name"] = platform.system()
|
|
|
|
if platform.release():
|
|
data.setdefault("system", {})["release"] = platform.release()
|
|
|
|
if platform.machine():
|
|
data["cpu"] = platform.machine()
|
|
|
|
return "{data[installer][name]}/{data[installer][version]} {json}".format(
|
|
data=data,
|
|
json=json.dumps(data, separators=(",", ":"), sort_keys=True),
|
|
)
|
|
|
|
|
|
class MultiDomainBasicAuth(AuthBase):
|
|
|
|
def __init__(self, prompting=True):
|
|
self.prompting = prompting
|
|
self.passwords = {}
|
|
|
|
def __call__(self, req):
|
|
parsed = urllib_parse.urlparse(req.url)
|
|
|
|
# Get the netloc without any embedded credentials
|
|
netloc = parsed.netloc.rsplit("@", 1)[-1]
|
|
|
|
# Set the url of the request to the url without any credentials
|
|
req.url = urllib_parse.urlunparse(parsed[:1] + (netloc,) + parsed[2:])
|
|
|
|
# Use any stored credentials that we have for this netloc
|
|
username, password = self.passwords.get(netloc, (None, None))
|
|
|
|
# Extract credentials embedded in the url if we have none stored
|
|
if username is None:
|
|
username, password = self.parse_credentials(parsed.netloc)
|
|
|
|
if username or password:
|
|
# Store the username and password
|
|
self.passwords[netloc] = (username, password)
|
|
|
|
# Send the basic auth with this request
|
|
req = HTTPBasicAuth(username or "", password or "")(req)
|
|
|
|
# Attach a hook to handle 401 responses
|
|
req.register_hook("response", self.handle_401)
|
|
|
|
return req
|
|
|
|
def handle_401(self, resp, **kwargs):
|
|
# We only care about 401 responses, anything else we want to just
|
|
# pass through the actual response
|
|
if resp.status_code != 401:
|
|
return resp
|
|
|
|
# We are not able to prompt the user so simple return the response
|
|
if not self.prompting:
|
|
return resp
|
|
|
|
parsed = urllib_parse.urlparse(resp.url)
|
|
|
|
# Prompt the user for a new username and password
|
|
username = six.moves.input("User for %s: " % parsed.netloc)
|
|
password = getpass.getpass("Password: ")
|
|
|
|
# Store the new username and password to use for future requests
|
|
if username or password:
|
|
self.passwords[parsed.netloc] = (username, password)
|
|
|
|
# Consume content and release the original connection to allow our new
|
|
# request to reuse the same one.
|
|
resp.content
|
|
resp.raw.release_conn()
|
|
|
|
# Add our new username and password to the request
|
|
req = HTTPBasicAuth(username or "", password or "")(resp.request)
|
|
|
|
# Send our new request
|
|
new_resp = resp.connection.send(req, **kwargs)
|
|
new_resp.history.append(resp)
|
|
|
|
return new_resp
|
|
|
|
def parse_credentials(self, netloc):
|
|
if "@" in netloc:
|
|
userinfo = netloc.rsplit("@", 1)[0]
|
|
if ":" in userinfo:
|
|
return userinfo.split(":", 1)
|
|
return userinfo, None
|
|
return None, None
|
|
|
|
|
|
class LocalFSAdapter(BaseAdapter):
|
|
|
|
def send(self, request, stream=None, timeout=None, verify=None, cert=None,
|
|
proxies=None):
|
|
pathname = url_to_path(request.url)
|
|
|
|
resp = Response()
|
|
resp.status_code = 200
|
|
resp.url = request.url
|
|
|
|
try:
|
|
stats = os.stat(pathname)
|
|
except OSError as exc:
|
|
resp.status_code = 404
|
|
resp.raw = exc
|
|
else:
|
|
modified = email.utils.formatdate(stats.st_mtime, usegmt=True)
|
|
content_type = mimetypes.guess_type(pathname)[0] or "text/plain"
|
|
resp.headers = CaseInsensitiveDict({
|
|
"Content-Type": content_type,
|
|
"Content-Length": stats.st_size,
|
|
"Last-Modified": modified,
|
|
})
|
|
|
|
resp.raw = open(pathname, "rb")
|
|
resp.close = resp.raw.close
|
|
|
|
return resp
|
|
|
|
def close(self):
|
|
pass
|
|
|
|
|
|
class SafeFileCache(FileCache):
|
|
"""
|
|
A file based cache which is safe to use even when the target directory may
|
|
not be accessible or writable.
|
|
"""
|
|
|
|
def __init__(self, *args, **kwargs):
|
|
super(SafeFileCache, self).__init__(*args, **kwargs)
|
|
|
|
# Check to ensure that the directory containing our cache directory
|
|
# is owned by the user current executing pip. If it does not exist
|
|
# we will check the parent directory until we find one that does exist.
|
|
# If it is not owned by the user executing pip then we will disable
|
|
# the cache and log a warning.
|
|
if not check_path_owner(self.directory):
|
|
logger.warning(
|
|
"The directory '%s' or its parent directory is not owned by "
|
|
"the current user and the cache has been disabled. Please "
|
|
"check the permissions and owner of that directory. If "
|
|
"executing pip with sudo, you may want sudo's -H flag.",
|
|
self.directory,
|
|
)
|
|
|
|
# Set our directory to None to disable the Cache
|
|
self.directory = None
|
|
|
|
def get(self, *args, **kwargs):
|
|
# If we don't have a directory, then the cache should be a no-op.
|
|
if self.directory is None:
|
|
return
|
|
|
|
try:
|
|
return super(SafeFileCache, self).get(*args, **kwargs)
|
|
except (LockError, OSError, IOError):
|
|
# We intentionally silence this error, if we can't access the cache
|
|
# then we can just skip caching and process the request as if
|
|
# caching wasn't enabled.
|
|
pass
|
|
|
|
def set(self, *args, **kwargs):
|
|
# If we don't have a directory, then the cache should be a no-op.
|
|
if self.directory is None:
|
|
return
|
|
|
|
try:
|
|
return super(SafeFileCache, self).set(*args, **kwargs)
|
|
except (LockError, OSError, IOError):
|
|
# We intentionally silence this error, if we can't access the cache
|
|
# then we can just skip caching and process the request as if
|
|
# caching wasn't enabled.
|
|
pass
|
|
|
|
def delete(self, *args, **kwargs):
|
|
# If we don't have a directory, then the cache should be a no-op.
|
|
if self.directory is None:
|
|
return
|
|
|
|
try:
|
|
return super(SafeFileCache, self).delete(*args, **kwargs)
|
|
except (LockError, OSError, IOError):
|
|
# We intentionally silence this error, if we can't access the cache
|
|
# then we can just skip caching and process the request as if
|
|
# caching wasn't enabled.
|
|
pass
|
|
|
|
|
|
class InsecureHTTPAdapter(HTTPAdapter):
|
|
|
|
def cert_verify(self, conn, url, verify, cert):
|
|
conn.cert_reqs = 'CERT_NONE'
|
|
conn.ca_certs = None
|
|
|
|
|
|
class PipSession(requests.Session):
|
|
|
|
timeout = None
|
|
|
|
def __init__(self, *args, **kwargs):
|
|
retries = kwargs.pop("retries", 0)
|
|
cache = kwargs.pop("cache", None)
|
|
insecure_hosts = kwargs.pop("insecure_hosts", [])
|
|
|
|
super(PipSession, self).__init__(*args, **kwargs)
|
|
|
|
# Attach our User Agent to the request
|
|
self.headers["User-Agent"] = user_agent()
|
|
|
|
# Attach our Authentication handler to the session
|
|
self.auth = MultiDomainBasicAuth()
|
|
|
|
# Create our urllib3.Retry instance which will allow us to customize
|
|
# how we handle retries.
|
|
retries = urllib3.Retry(
|
|
# Set the total number of retries that a particular request can
|
|
# have.
|
|
total=retries,
|
|
|
|
# A 503 error from PyPI typically means that the Fastly -> Origin
|
|
# connection got interupted in some way. A 503 error in general
|
|
# is typically considered a transient error so we'll go ahead and
|
|
# retry it.
|
|
status_forcelist=[503],
|
|
|
|
# Add a small amount of back off between failed requests in
|
|
# order to prevent hammering the service.
|
|
backoff_factor=0.25,
|
|
)
|
|
|
|
# We want to _only_ cache responses on securely fetched origins. We do
|
|
# this because we can't validate the response of an insecurely fetched
|
|
# origin, and we don't want someone to be able to poison the cache and
|
|
# require manual evication from the cache to fix it.
|
|
if cache:
|
|
secure_adapter = CacheControlAdapter(
|
|
cache=SafeFileCache(cache, use_dir_lock=True),
|
|
max_retries=retries,
|
|
)
|
|
else:
|
|
secure_adapter = HTTPAdapter(max_retries=retries)
|
|
|
|
# Our Insecure HTTPAdapter disables HTTPS validation. It does not
|
|
# support caching (see above) so we'll use it for all http:// URLs as
|
|
# well as any https:// host that we've marked as ignoring TLS errors
|
|
# for.
|
|
insecure_adapter = InsecureHTTPAdapter(max_retries=retries)
|
|
|
|
self.mount("https://", secure_adapter)
|
|
self.mount("http://", insecure_adapter)
|
|
|
|
# Enable file:// urls
|
|
self.mount("file://", LocalFSAdapter())
|
|
|
|
# We want to use a non-validating adapter for any requests which are
|
|
# deemed insecure.
|
|
for host in insecure_hosts:
|
|
self.mount("https://{0}/".format(host), insecure_adapter)
|
|
|
|
def request(self, method, url, *args, **kwargs):
|
|
# Allow setting a default timeout on a session
|
|
kwargs.setdefault("timeout", self.timeout)
|
|
|
|
# Dispatch the actual request
|
|
return super(PipSession, self).request(method, url, *args, **kwargs)
|
|
|
|
|
|
def get_file_content(url, comes_from=None, session=None):
|
|
"""Gets the content of a file; it may be a filename, file: URL, or
|
|
http: URL. Returns (location, content). Content is unicode."""
|
|
if session is None:
|
|
raise TypeError(
|
|
"get_file_content() missing 1 required keyword argument: 'session'"
|
|
)
|
|
|
|
match = _scheme_re.search(url)
|
|
if match:
|
|
scheme = match.group(1).lower()
|
|
if (scheme == 'file' and comes_from and
|
|
comes_from.startswith('http')):
|
|
raise InstallationError(
|
|
'Requirements file %s references URL %s, which is local'
|
|
% (comes_from, url))
|
|
if scheme == 'file':
|
|
path = url.split(':', 1)[1]
|
|
path = path.replace('\\', '/')
|
|
match = _url_slash_drive_re.match(path)
|
|
if match:
|
|
path = match.group(1) + ':' + path.split('|', 1)[1]
|
|
path = urllib_parse.unquote(path)
|
|
if path.startswith('/'):
|
|
path = '/' + path.lstrip('/')
|
|
url = path
|
|
else:
|
|
# FIXME: catch some errors
|
|
resp = session.get(url)
|
|
resp.raise_for_status()
|
|
|
|
if six.PY3:
|
|
return resp.url, resp.text
|
|
else:
|
|
return resp.url, resp.content
|
|
try:
|
|
with open(url) as f:
|
|
content = f.read()
|
|
except IOError as exc:
|
|
raise InstallationError(
|
|
'Could not open requirements file: %s' % str(exc)
|
|
)
|
|
return url, content
|
|
|
|
|
|
_scheme_re = re.compile(r'^(http|https|file):', re.I)
|
|
_url_slash_drive_re = re.compile(r'/*([a-z])\|', re.I)
|
|
|
|
|
|
def is_url(name):
|
|
"""Returns true if the name looks like a URL"""
|
|
if ':' not in name:
|
|
return False
|
|
scheme = name.split(':', 1)[0].lower()
|
|
return scheme in ['http', 'https', 'file', 'ftp'] + vcs.all_schemes
|
|
|
|
|
|
def url_to_path(url):
|
|
"""
|
|
Convert a file: URL to a path.
|
|
"""
|
|
assert url.startswith('file:'), (
|
|
"You can only turn file: urls into filenames (not %r)" % url)
|
|
|
|
_, netloc, path, _, _ = urllib_parse.urlsplit(url)
|
|
|
|
# if we have a UNC path, prepend UNC share notation
|
|
if netloc:
|
|
netloc = '\\\\' + netloc
|
|
|
|
path = urllib_request.url2pathname(netloc + path)
|
|
return path
|
|
|
|
|
|
def path_to_url(path):
|
|
"""
|
|
Convert a path to a file: URL. The path will be made absolute and have
|
|
quoted path parts.
|
|
"""
|
|
path = os.path.normpath(os.path.abspath(path))
|
|
url = urllib_parse.urljoin('file:', urllib_request.pathname2url(path))
|
|
return url
|
|
|
|
|
|
def is_archive_file(name):
|
|
"""Return True if `name` is a considered as an archive file."""
|
|
ext = splitext(name)[1].lower()
|
|
if ext in ARCHIVE_EXTENSIONS:
|
|
return True
|
|
return False
|
|
|
|
|
|
def unpack_vcs_link(link, location):
|
|
vcs_backend = _get_used_vcs_backend(link)
|
|
vcs_backend.unpack(location)
|
|
|
|
|
|
def _get_used_vcs_backend(link):
|
|
for backend in vcs.backends:
|
|
if link.scheme in backend.schemes:
|
|
vcs_backend = backend(link.url)
|
|
return vcs_backend
|
|
|
|
|
|
def is_vcs_url(link):
|
|
return bool(_get_used_vcs_backend(link))
|
|
|
|
|
|
def is_file_url(link):
|
|
return link.url.lower().startswith('file:')
|
|
|
|
|
|
def _check_hash(download_hash, link):
|
|
if download_hash.digest_size != hashlib.new(link.hash_name).digest_size:
|
|
logger.critical(
|
|
"Hash digest size of the package %d (%s) doesn't match the "
|
|
"expected hash name %s!",
|
|
download_hash.digest_size, link, link.hash_name,
|
|
)
|
|
raise HashMismatch('Hash name mismatch for package %s' % link)
|
|
if download_hash.hexdigest() != link.hash:
|
|
logger.critical(
|
|
"Hash of the package %s (%s) doesn't match the expected hash %s!",
|
|
link, download_hash.hexdigest(), link.hash,
|
|
)
|
|
raise HashMismatch(
|
|
'Bad %s hash for package %s' % (link.hash_name, link)
|
|
)
|
|
|
|
|
|
def _get_hash_from_file(target_file, link):
|
|
try:
|
|
download_hash = hashlib.new(link.hash_name)
|
|
except (ValueError, TypeError):
|
|
logger.warning(
|
|
"Unsupported hash name %s for package %s", link.hash_name, link,
|
|
)
|
|
return None
|
|
|
|
with open(target_file, 'rb') as fp:
|
|
while True:
|
|
chunk = fp.read(4096)
|
|
if not chunk:
|
|
break
|
|
download_hash.update(chunk)
|
|
return download_hash
|
|
|
|
|
|
def _progress_indicator(iterable, *args, **kwargs):
|
|
return iterable
|
|
|
|
|
|
def _download_url(resp, link, content_file):
|
|
download_hash = None
|
|
if link.hash and link.hash_name:
|
|
try:
|
|
download_hash = hashlib.new(link.hash_name)
|
|
except ValueError:
|
|
logger.warning(
|
|
"Unsupported hash name %s for package %s",
|
|
link.hash_name, link,
|
|
)
|
|
|
|
try:
|
|
total_length = int(resp.headers['content-length'])
|
|
except (ValueError, KeyError, TypeError):
|
|
total_length = 0
|
|
|
|
cached_resp = getattr(resp, "from_cache", False)
|
|
|
|
if logger.getEffectiveLevel() > logging.INFO:
|
|
show_progress = False
|
|
elif cached_resp:
|
|
show_progress = False
|
|
elif total_length > (40 * 1000):
|
|
show_progress = True
|
|
elif not total_length:
|
|
show_progress = True
|
|
else:
|
|
show_progress = False
|
|
|
|
show_url = link.show_url
|
|
|
|
def resp_read(chunk_size):
|
|
try:
|
|
# Special case for urllib3.
|
|
for chunk in resp.raw.stream(
|
|
chunk_size,
|
|
# We use decode_content=False here because we do
|
|
# want urllib3 to mess with the raw bytes we get
|
|
# from the server. If we decompress inside of
|
|
# urllib3 then we cannot verify the checksum
|
|
# because the checksum will be of the compressed
|
|
# file. This breakage will only occur if the
|
|
# server adds a Content-Encoding header, which
|
|
# depends on how the server was configured:
|
|
# - Some servers will notice that the file isn't a
|
|
# compressible file and will leave the file alone
|
|
# and with an empty Content-Encoding
|
|
# - Some servers will notice that the file is
|
|
# already compressed and will leave the file
|
|
# alone and will add a Content-Encoding: gzip
|
|
# header
|
|
# - Some servers won't notice anything at all and
|
|
# will take a file that's already been compressed
|
|
# and compress it again and set the
|
|
# Content-Encoding: gzip header
|
|
#
|
|
# By setting this not to decode automatically we
|
|
# hope to eliminate problems with the second case.
|
|
decode_content=False):
|
|
yield chunk
|
|
except AttributeError:
|
|
# Standard file-like object.
|
|
while True:
|
|
chunk = resp.raw.read(chunk_size)
|
|
if not chunk:
|
|
break
|
|
yield chunk
|
|
|
|
progress_indicator = _progress_indicator
|
|
|
|
if link.netloc == PyPI.netloc:
|
|
url = show_url
|
|
else:
|
|
url = link.url_without_fragment
|
|
|
|
if show_progress: # We don't show progress on cached responses
|
|
if total_length:
|
|
logger.info(
|
|
"Downloading %s (%s)", url, format_size(total_length),
|
|
)
|
|
progress_indicator = DownloadProgressBar(
|
|
max=total_length,
|
|
).iter
|
|
else:
|
|
logger.info("Downloading %s", url)
|
|
progress_indicator = DownloadProgressSpinner().iter
|
|
elif cached_resp:
|
|
logger.info("Using cached %s", url)
|
|
else:
|
|
logger.info("Downloading %s", url)
|
|
|
|
logger.debug('Downloading from URL %s', link)
|
|
|
|
for chunk in progress_indicator(resp_read(4096), 4096):
|
|
if download_hash is not None:
|
|
download_hash.update(chunk)
|
|
content_file.write(chunk)
|
|
if link.hash and link.hash_name:
|
|
_check_hash(download_hash, link)
|
|
return download_hash
|
|
|
|
|
|
def _copy_file(filename, location, content_type, link):
|
|
copy = True
|
|
download_location = os.path.join(location, link.filename)
|
|
if os.path.exists(download_location):
|
|
response = ask_path_exists(
|
|
'The file %s exists. (i)gnore, (w)ipe, (b)ackup ' %
|
|
display_path(download_location), ('i', 'w', 'b'))
|
|
if response == 'i':
|
|
copy = False
|
|
elif response == 'w':
|
|
logger.warning('Deleting %s', display_path(download_location))
|
|
os.remove(download_location)
|
|
elif response == 'b':
|
|
dest_file = backup_dir(download_location)
|
|
logger.warning(
|
|
'Backing up %s to %s',
|
|
display_path(download_location),
|
|
display_path(dest_file),
|
|
)
|
|
shutil.move(download_location, dest_file)
|
|
if copy:
|
|
shutil.copy(filename, download_location)
|
|
logger.info('Saved %s', display_path(download_location))
|
|
|
|
|
|
def unpack_http_url(link, location, download_dir=None, session=None):
|
|
if session is None:
|
|
raise TypeError(
|
|
"unpack_http_url() missing 1 required keyword argument: 'session'"
|
|
)
|
|
|
|
temp_dir = tempfile.mkdtemp('-unpack', 'pip-')
|
|
|
|
# If a download dir is specified, is the file already downloaded there?
|
|
already_downloaded_path = None
|
|
if download_dir:
|
|
already_downloaded_path = _check_download_dir(link, download_dir)
|
|
|
|
if already_downloaded_path:
|
|
from_path = already_downloaded_path
|
|
content_type = mimetypes.guess_type(from_path)[0]
|
|
else:
|
|
# let's download to a tmp dir
|
|
from_path, content_type = _download_http_url(link, session, temp_dir)
|
|
|
|
# unpack the archive to the build dir location. even when only downloading
|
|
# archives, they have to be unpacked to parse dependencies
|
|
unpack_file(from_path, location, content_type, link)
|
|
|
|
# a download dir is specified; let's copy the archive there
|
|
if download_dir and not already_downloaded_path:
|
|
_copy_file(from_path, download_dir, content_type, link)
|
|
|
|
if not already_downloaded_path:
|
|
os.unlink(from_path)
|
|
rmtree(temp_dir)
|
|
|
|
|
|
def unpack_file_url(link, location, download_dir=None):
|
|
"""Unpack link into location.
|
|
If download_dir is provided and link points to a file, make a copy
|
|
of the link file inside download_dir."""
|
|
|
|
link_path = url_to_path(link.url_without_fragment)
|
|
|
|
# If it's a url to a local directory
|
|
if os.path.isdir(link_path):
|
|
if os.path.isdir(location):
|
|
rmtree(location)
|
|
shutil.copytree(link_path, location, symlinks=True)
|
|
if download_dir:
|
|
logger.info('Link is a directory, ignoring download_dir')
|
|
return
|
|
|
|
# if link has a hash, let's confirm it matches
|
|
if link.hash:
|
|
link_path_hash = _get_hash_from_file(link_path, link)
|
|
_check_hash(link_path_hash, link)
|
|
|
|
# If a download dir is specified, is the file already there and valid?
|
|
already_downloaded_path = None
|
|
if download_dir:
|
|
already_downloaded_path = _check_download_dir(link, download_dir)
|
|
|
|
if already_downloaded_path:
|
|
from_path = already_downloaded_path
|
|
else:
|
|
from_path = link_path
|
|
|
|
content_type = mimetypes.guess_type(from_path)[0]
|
|
|
|
# unpack the archive to the build dir location. even when only downloading
|
|
# archives, they have to be unpacked to parse dependencies
|
|
unpack_file(from_path, location, content_type, link)
|
|
|
|
# a download dir is specified and not already downloaded
|
|
if download_dir and not already_downloaded_path:
|
|
_copy_file(from_path, download_dir, content_type, link)
|
|
|
|
|
|
def _copy_dist_from_dir(link_path, location):
|
|
"""Copy distribution files in `link_path` to `location`.
|
|
|
|
Invoked when user requests to install a local directory. E.g.:
|
|
|
|
pip install .
|
|
pip install ~/dev/git-repos/python-prompt-toolkit
|
|
|
|
"""
|
|
|
|
# Note: This is currently VERY SLOW if you have a lot of data in the
|
|
# directory, because it copies everything with `shutil.copytree`.
|
|
# What it should really do is build an sdist and install that.
|
|
# See https://github.com/pypa/pip/issues/2195
|
|
|
|
if os.path.isdir(location):
|
|
rmtree(location)
|
|
|
|
# build an sdist
|
|
setup_py = 'setup.py'
|
|
sdist_args = [sys.executable]
|
|
sdist_args.append('-c')
|
|
sdist_args.append(
|
|
"import setuptools, tokenize;__file__=%r;"
|
|
"exec(compile(getattr(tokenize, 'open', open)(__file__).read()"
|
|
".replace('\\r\\n', '\\n'), __file__, 'exec'))" % setup_py)
|
|
sdist_args.append('sdist')
|
|
sdist_args += ['--dist-dir', location]
|
|
logger.info('Running setup.py sdist for %s', link_path)
|
|
|
|
with indent_log():
|
|
call_subprocess(sdist_args, cwd=link_path, show_stdout=False)
|
|
|
|
# unpack sdist into `location`
|
|
sdist = os.path.join(location, os.listdir(location)[0])
|
|
logger.info('Unpacking sdist %s into %s', sdist, location)
|
|
unpack_file(sdist, location, content_type=None, link=None)
|
|
|
|
|
|
class PipXmlrpcTransport(xmlrpc_client.Transport):
|
|
"""Provide a `xmlrpclib.Transport` implementation via a `PipSession`
|
|
object.
|
|
"""
|
|
def __init__(self, index_url, session, use_datetime=False):
|
|
xmlrpc_client.Transport.__init__(self, use_datetime)
|
|
index_parts = urllib_parse.urlparse(index_url)
|
|
self._scheme = index_parts.scheme
|
|
self._session = session
|
|
|
|
def request(self, host, handler, request_body, verbose=False):
|
|
parts = (self._scheme, host, handler, None, None, None)
|
|
url = urllib_parse.urlunparse(parts)
|
|
try:
|
|
headers = {'Content-Type': 'text/xml'}
|
|
response = self._session.post(url, data=request_body,
|
|
headers=headers, stream=True)
|
|
response.raise_for_status()
|
|
self.verbose = verbose
|
|
return self.parse_response(response.raw)
|
|
except requests.HTTPError as exc:
|
|
logger.critical(
|
|
"HTTP error %s while getting %s",
|
|
exc.response.status_code, url,
|
|
)
|
|
raise
|
|
|
|
|
|
def unpack_url(link, location, download_dir=None,
|
|
only_download=False, session=None):
|
|
"""Unpack link.
|
|
If link is a VCS link:
|
|
if only_download, export into download_dir and ignore location
|
|
else unpack into location
|
|
for other types of link:
|
|
- unpack into location
|
|
- if download_dir, copy the file into download_dir
|
|
- if only_download, mark location for deletion
|
|
"""
|
|
# non-editable vcs urls
|
|
if is_vcs_url(link):
|
|
unpack_vcs_link(link, location)
|
|
|
|
# file urls
|
|
elif is_file_url(link):
|
|
unpack_file_url(link, location, download_dir)
|
|
|
|
# http urls
|
|
else:
|
|
if session is None:
|
|
session = PipSession()
|
|
|
|
unpack_http_url(
|
|
link,
|
|
location,
|
|
download_dir,
|
|
session,
|
|
)
|
|
if only_download:
|
|
write_delete_marker_file(location)
|
|
|
|
|
|
def _download_http_url(link, session, temp_dir):
|
|
"""Download link url into temp_dir using provided session"""
|
|
target_url = link.url.split('#', 1)[0]
|
|
try:
|
|
resp = session.get(
|
|
target_url,
|
|
# We use Accept-Encoding: identity here because requests
|
|
# defaults to accepting compressed responses. This breaks in
|
|
# a variety of ways depending on how the server is configured.
|
|
# - Some servers will notice that the file isn't a compressible
|
|
# file and will leave the file alone and with an empty
|
|
# Content-Encoding
|
|
# - Some servers will notice that the file is already
|
|
# compressed and will leave the file alone and will add a
|
|
# Content-Encoding: gzip header
|
|
# - Some servers won't notice anything at all and will take
|
|
# a file that's already been compressed and compress it again
|
|
# and set the Content-Encoding: gzip header
|
|
# By setting this to request only the identity encoding We're
|
|
# hoping to eliminate the third case. Hopefully there does not
|
|
# exist a server which when given a file will notice it is
|
|
# already compressed and that you're not asking for a
|
|
# compressed file and will then decompress it before sending
|
|
# because if that's the case I don't think it'll ever be
|
|
# possible to make this work.
|
|
headers={"Accept-Encoding": "identity"},
|
|
stream=True,
|
|
)
|
|
resp.raise_for_status()
|
|
except requests.HTTPError as exc:
|
|
logger.critical(
|
|
"HTTP error %s while getting %s", exc.response.status_code, link,
|
|
)
|
|
raise
|
|
|
|
content_type = resp.headers.get('content-type', '')
|
|
filename = link.filename # fallback
|
|
# Have a look at the Content-Disposition header for a better guess
|
|
content_disposition = resp.headers.get('content-disposition')
|
|
if content_disposition:
|
|
type, params = cgi.parse_header(content_disposition)
|
|
# We use ``or`` here because we don't want to use an "empty" value
|
|
# from the filename param.
|
|
filename = params.get('filename') or filename
|
|
ext = splitext(filename)[1]
|
|
if not ext:
|
|
ext = mimetypes.guess_extension(content_type)
|
|
if ext:
|
|
filename += ext
|
|
if not ext and link.url != resp.url:
|
|
ext = os.path.splitext(resp.url)[1]
|
|
if ext:
|
|
filename += ext
|
|
file_path = os.path.join(temp_dir, filename)
|
|
with open(file_path, 'wb') as content_file:
|
|
_download_url(resp, link, content_file)
|
|
return file_path, content_type
|
|
|
|
|
|
def _check_download_dir(link, download_dir):
|
|
""" Check download_dir for previously downloaded file with correct hash
|
|
If a correct file is found return its path else None
|
|
"""
|
|
download_path = os.path.join(download_dir, link.filename)
|
|
if os.path.exists(download_path):
|
|
# If already downloaded, does its hash match?
|
|
logger.info('File was already downloaded %s', download_path)
|
|
if link.hash:
|
|
download_hash = _get_hash_from_file(download_path, link)
|
|
try:
|
|
_check_hash(download_hash, link)
|
|
except HashMismatch:
|
|
logger.warning(
|
|
'Previously-downloaded file %s has bad hash, '
|
|
're-downloading.',
|
|
download_path
|
|
)
|
|
os.unlink(download_path)
|
|
return None
|
|
return download_path
|
|
return None
|