From: Yousong Zhou Date: Thu, 28 Jun 2018 10:27:27 +0000 (+0800) Subject: scripts/dl_github_archive.py: rename from download.py X-Git-Tag: v19.07.0-rc1~3250 X-Git-Url: http://git.lede-project.org./?a=commitdiff_plain;h=04b9f8587370f96366c6e53fb411473279ba7c02;p=openwrt%2Fstaging%2Fpepe2k.git scripts/dl_github_archive.py: rename from download.py - Make the code more GitHub-specific - Requires mirror hash to work with .gitattributes - Use different API depending on whether PKG_SOURCE_VERSION is a complete commit id or other ref types like tags - Fix removing symbolic link - pre-clean dir_untar for possible leftovers from previous run Signed-off-by: Yousong Zhou --- diff --git a/include/download.mk b/include/download.mk index 5cfc542042..1627b4fdfb 100644 --- a/include/download.mk +++ b/include/download.mk @@ -172,25 +172,25 @@ endef define DownloadMethod/git $(call wrap_mirror,$(1),$(2), \ - $(call DownloadMethod/git-raw) \ + $(call DownloadMethod/rawgit) \ ) endef -define DownloadMethod/github-tarball +define DownloadMethod/github_archive $(call wrap_mirror,$(1),$(2), \ - $(SCRIPT_DIR)/download.py dl \ + $(SCRIPT_DIR)/dl_github_archive.py \ --dl-dir="$(DL_DIR)" \ - --url $(foreach url,$(URL),"$(url)") \ - --proto="$(PROTO)" \ + --url="$(URL)" \ --version="$(VERSION)" \ --subdir="$(SUBDIR)" \ --source="$(FILE)" \ - || ( $(call DownloadMethod/git-raw) ); \ + --hash="$(MIRROR_HASH)" \ + || ( $(call DownloadMethod/rawgit) ); \ ) endef # Only intends to be called as a submethod from other DownloadMethod -define DownloadMethod/git-raw +define DownloadMethod/rawgit echo "Checking out files from the git repository..."; \ mkdir -p $(TMP_DIR)/dl && \ cd $(TMP_DIR)/dl && \ diff --git a/scripts/dl_github_archive.py b/scripts/dl_github_archive.py new file mode 100755 index 0000000000..5a5a016e37 --- /dev/null +++ b/scripts/dl_github_archive.py @@ -0,0 +1,426 @@ +#!/usr/bin/env python +# +# Copyright (c) 2018 Yousong Zhou +# +# This is free software, licensed under the GNU General Public License v2. +# See /LICENSE for more information. + +import argparse +import calendar +import datetime +import errno +import fcntl +import hashlib +import json +import os +import os.path +import re +import shutil +import ssl +import subprocess +import sys +import time +import urllib2 + +TMPDIR = os.environ.get('TMP_DIR') or '/tmp' +TMPDIR_DL = os.path.join(TMPDIR, 'dl') + + +class PathException(Exception): pass +class DownloadGitHubError(Exception): pass + + +class Path(object): + """Context class for preparing and cleaning up directories. + + If ```preclean` is ``False``, ``path`` will NOT be removed on context enter + + If ``path`` ``isdir``, then it will be created on context enter. + + If ``keep`` is True, then ``path`` will NOT be removed on context exit + """ + + def __init__(self, path, isdir=True, preclean=False, keep=False): + self.path = path + self.isdir = isdir + self.preclean = preclean + self.keep = keep + + def __enter__(self): + if self.preclean: + self.rm_all(self.path) + if self.isdir: + self.mkdir_all(self.path) + return self + + def __exit__(self, exc_type, exc_value, traceback): + if not self.keep: + self.rm_all(self.path) + + @staticmethod + def mkdir_all(path): + """Same as mkdir -p.""" + names = os.path.split(path) + p = '' + for name in names: + p = os.path.join(p, name) + Path._mkdir(p) + + @staticmethod + def _rmdir_dir(dir_): + names = Path._listdir(dir_) + for name in names: + p = os.path.join(dir_, name) + Path.rm_all(p) + Path._rmdir(dir_) + + @staticmethod + def _mkdir(path): + Path._os_func(os.mkdir, path, errno.EEXIST) + + @staticmethod + def _rmdir(path): + Path._os_func(os.rmdir, path, errno.ENOENT) + + @staticmethod + def _remove(path): + Path._os_func(os.remove, path, errno.ENOENT) + + @staticmethod + def _listdir(path): + return Path._os_func(os.listdir, path, errno.ENOENT, default=[]) + + @staticmethod + def _os_func(func, path, errno, default=None): + """Call func(path) in an idempotent way. + + On exception ``ex``, if the type is OSError and ``ex.errno == errno``, + return ``default``, otherwise, re-raise + """ + try: + return func(path) + except OSError as e: + if e.errno == errno: + return default + else: + raise + + @staticmethod + def rm_all(path): + """Same as rm -r.""" + if os.path.islink(path): + Path._remove(path) + elif os.path.isdir(path): + Path._rmdir_dir(path) + else: + Path._remove(path) + + @staticmethod + def untar(path, into=None): + """Extract tarball at ``path`` into subdir ``into``. + + return subdir name if and only if there exists one, otherwise raise PathException + """ + args = ('tar', '-C', into, '-xzf', path, '--no-same-permissions') + subprocess.check_call(args, preexec_fn=lambda: os.umask(0o22)) + dirs = os.listdir(into) + if len(dirs) == 1: + return dirs[0] + else: + raise PathException('untar %s: expecting a single subdir, got %s' % (path, dirs)) + + @staticmethod + def tar(path, subdir, into=None, ts=None): + """Pack ``path`` into tarball ``into``.""" + # --sort=name requires a recent build of GNU tar + args = ['tar', '--numeric-owner', '--owner=0', '--group=0', '--sort=name'] + args += ['-C', path, '-cf', into, subdir] + envs = os.environ.copy() + if ts is not None: + args.append('--mtime=@%d' % ts) + if into.endswith('.xz'): + envs['XZ_OPT'] = '-7e' + args.append('-J') + elif into.endswith('.bz2'): + args.append('-j') + elif into.endswith('.gz'): + args.append('-z') + envs['GZIP'] = '-n' + else: + raise PathException('unknown compression type %s' % into) + subprocess.check_call(args, env=envs) + + +class GitHubCommitTsCache(object): + __cachef = 'github.commit.ts.cache' + __cachen = 2048 + + def __init__(self): + Path.mkdir_all(TMPDIR_DL) + self.cachef = os.path.join(TMPDIR_DL, self.__cachef) + self.cache = {} + + def get(self, k): + """Get timestamp with key ``k``.""" + fileno = os.open(self.cachef, os.O_RDONLY | os.O_CREAT) + with os.fdopen(fileno) as fin: + try: + fcntl.lockf(fileno, fcntl.LOCK_SH) + self._cache_init(fin) + if k in self.cache: + ts = self.cache[k][0] + return ts + finally: + fcntl.lockf(fileno, fcntl.LOCK_UN) + return None + + def set(self, k, v): + """Update timestamp with ``k``.""" + fileno = os.open(self.cachef, os.O_RDWR | os.O_CREAT) + with os.fdopen(fileno, 'wb+') as f: + try: + fcntl.lockf(fileno, fcntl.LOCK_EX) + self._cache_init(f) + self.cache[k] = (v, int(time.time())) + self._cache_flush(f) + finally: + fcntl.lockf(fileno, fcntl.LOCK_UN) + + def _cache_init(self, fin): + for line in fin: + k, ts, updated = line.split() + ts = int(ts) + updated = int(updated) + self.cache[k] = (ts, updated) + + def _cache_flush(self, fout): + cache = sorted(self.cache.iteritems(), cmp=lambda a, b: b[1][1] - a[1][1]) + cache = cache[:self.__cachen] + self.cache = {} + os.ftruncate(fout.fileno(), 0) + fout.seek(0, os.SEEK_SET) + for k, ent in cache: + ts = ent[0] + updated = ent[1] + line = '{0} {1} {2}\n'.format(k, ts, updated) + fout.write(line) + + +class DownloadGitHubTarball(object): + """Download and repack archive tarabll from GitHub. + + Compared with the method of packing after cloning the whole repo, this + method is more friendly to users with fragile internet connection. + + However, there are limitations with this method + + - GitHub imposes a 60 reqs/hour limit for unauthenticated API access. + This affects fetching commit date for reproducible tarballs. Download + through the archive link is not affected. + + - GitHub archives do not contain source codes for submodules. + + - GitHub archives seem to respect .gitattributes and ignore pathes with + export-ignore attributes. + + For the first two issues, the method will fail loudly to allow fallback to + clone-then-pack method. + + As for the 3rd issue, to make sure that this method only produces identical + tarballs as the fallback method, we require the expected hash value to be + supplied. That means the first tarball will need to be prepared by the + clone-then-pack method + """ + + __repo_url_regex = re.compile(r'^(?:https|git)://github.com/(?P[^/]+)/(?P[^/]+)') + + def __init__(self, args): + self.dl_dir = args.dl_dir + self.version = args.version + self.subdir = args.subdir + self.source = args.source + self.url = args.url + self._init_owner_repo() + self.xhash = args.hash + self._init_hasher() + self.commit_ts = None # lazy load commit timestamp + self.commit_ts_cache = GitHubCommitTsCache() + self.name = 'github-tarball' + + def download(self): + """Download and repack GitHub archive tarball.""" + self._init_commit_ts() + with Path(TMPDIR_DL, keep=True) as dir_dl: + # fetch tarball from GitHub + tarball_path = os.path.join(dir_dl.path, self.subdir + '.tar.gz.dl') + with Path(tarball_path, isdir=False): + self._fetch(tarball_path) + # unpack + d = os.path.join(dir_dl.path, self.subdir + '.untar') + with Path(d, preclean=True) as dir_untar: + tarball_prefix = Path.untar(tarball_path, into=dir_untar.path) + dir0 = os.path.join(dir_untar.path, tarball_prefix) + dir1 = os.path.join(dir_untar.path, self.subdir) + # submodules check + if self._has_submodule(dir0): + raise self._error('Fetching submodules is not yet supported') + # rename subdir + os.rename(dir0, dir1) + # repack + into=os.path.join(TMPDIR_DL, self.source) + Path.tar(dir_untar.path, self.subdir, into=into, ts=self.commit_ts) + try: + self._hash_check(into) + except Exception: + Path.rm_all(into) + raise + # move to target location + file1 = os.path.join(self.dl_dir, self.source) + if into != file1: + shutil.move(into, file1) + + def _has_submodule(self, dir_): + m = os.path.join(dir_, '.gitmodules') + try: + st = os.stat(m) + return st.st_size > 0 + except OSError as e: + return e.errno != errno.ENOENT + + def _init_owner_repo(self): + m = self.__repo_url_regex.search(self.url) + if m is None: + raise self._error('Invalid github url: {}'.format(self.url)) + owner = m.group('owner') + repo = m.group('repo') + if repo.endswith('.git'): + repo = repo[:-4] + self.owner = owner + self.repo = repo + + def _init_hasher(self): + xhash = self.xhash + if len(xhash) == 64: + self.hasher = hashlib.sha256() + elif len(xhash) == 32: + self.hasher = hashlib.md5() + else: + raise self._error('Requires sha256sum for verification') + self.xhash = xhash + + def _hash_check(self, f): + with open(f, 'rb') as fin: + while True: + d = fin.read(4096) + if not d: + break + self.hasher.update(d) + xhash = self.hasher.hexdigest() + if xhash != self.xhash: + raise self._error('Wrong hash (probably caused by .gitattributes), expecting {}, got {}'.format(self.xhash, xhash)) + + def _init_commit_ts(self): + if self.commit_ts is not None: + return + # GitHub provides 2 APIs[1,2] for fetching commit data. API[1] is more + # terse while API[2] provides more verbose info such as commit diff + # etc. That's the main reason why API[1] is preferred: the response + # size is predictable. + # + # However, API[1] only accepts complete commit sha1sum as the parameter + # while API[2] is more liberal accepting also partial commit id and + # tags, etc. + # + # [1] Get a single commit, Repositories, https://developer.github.com/v3/repos/commits/#get-a-single-commit + # [2] Git Commits, Git Data, https://developer.github.com/v3/git/commits/#get-a-commit + apis = [ + { + 'url': self._make_repo_url_path('git', 'commits', self.version), + 'attr_path': ('committer', 'date'), + }, { + 'url': self._make_repo_url_path('commits', self.version), + 'attr_path': ('commit', 'committer', 'date'), + }, + ] + version_is_sha1sum = len(self.version) == 40 + if not version_is_sha1sum: + apis.insert(0, apis.pop()) + for api in apis: + url = api['url'] + attr_path = api['attr_path'] + try: + ct = self.commit_ts_cache.get(url) + if ct is not None: + self.commit_ts = ct + return + ct = self._init_commit_ts_remote_get(url, attr_path) + self.commit_ts = ct + self.commit_ts_cache.set(url, ct) + return + except Exception: + pass + raise self._error('Cannot fetch commit ts: {}'.format(url)) + + def _init_commit_ts_remote_get(self, url, attrpath): + resp = self._make_request(url) + data = resp.read() + date = json.loads(data) + for attr in attrpath: + date = date[attr] + date = datetime.datetime.strptime(date, '%Y-%m-%dT%H:%M:%SZ') + date = date.timetuple() + ct = calendar.timegm(date) + return ct + + def _fetch(self, path): + """Fetch tarball of the specified version ref.""" + ref = self.version + url = self._make_repo_url_path('tarball', ref) + resp = self._make_request(url) + with open(path, 'wb') as fout: + while True: + d = resp.read(4096) + if not d: + break + fout.write(d) + + def _make_repo_url_path(self, *args): + url = '/repos/{0}/{1}'.format(self.owner, self.repo) + if args: + url += '/' + '/'.join(args) + return url + + def _make_request(self, path): + """Request GitHub API endpoint on ``path``.""" + url = 'https://api.github.com' + path + headers = { + 'Accept': 'application/vnd.github.v3+json', + 'User-Agent': 'OpenWrt', + } + req = urllib2.Request(url, headers=headers) + sslcontext = ssl._create_unverified_context() + fileobj = urllib2.urlopen(req, context=sslcontext) + return fileobj + + def _error(self, msg): + return DownloadGitHubError('{}: {}'.format(self.source, msg)) + + +def main(): + parser = argparse.ArgumentParser() + parser.add_argument('--dl-dir', default=os.getcwd(), help='Download dir') + parser.add_argument('--url', help='Download URL') + parser.add_argument('--subdir', help='Source code subdir name') + parser.add_argument('--version', help='Source code version') + parser.add_argument('--source', help='Source tarball filename') + parser.add_argument('--hash', help='Source tarball\'s expected sha256sum') + args = parser.parse_args() + method = DownloadGitHubTarball(args) + try: + method.download() + except Exception: + sys.stderr.write('download {} from {} failed\n'.format(args.source, args.url)) + raise + +if __name__ == '__main__': + main() diff --git a/scripts/download.py b/scripts/download.py deleted file mode 100755 index 779d7b3de2..0000000000 --- a/scripts/download.py +++ /dev/null @@ -1,421 +0,0 @@ -#!/usr/bin/env python -# -# Copyright (c) 2018 Yousong Zhou -# -# This is free software, licensed under the GNU General Public License v2. -# See /LICENSE for more information. - -import argparse -import calendar -import datetime -import errno -import fcntl -import json -import os -import os.path -import re -import shutil -import ssl -import subprocess -import sys -import time -import urllib2 - -TMPDIR = os.environ.get('TMP_DIR') or '/tmp' -TMPDIR_DL = os.path.join(TMPDIR, 'dl') -DOWNLOAD_METHODS = [] - -class PathException(Exception): pass -class DownloadException(Exception): pass - - -class Path(object): - """Context class for preparing and cleaning up directories. - - If ``path`` ``isdir``, then it will be created on context enter. - - If ``keep`` is True, then ``path`` will NOT be removed on context exit - """ - - def __init__(self, path, isdir=True, keep=False): - self.path = path - self.isdir = isdir - self.keep = keep - - def __enter__(self): - if self.isdir: - self.mkdir_all(self.path) - return self - - def __exit__(self, exc_type, exc_value, traceback): - if not self.keep: - self.rm_all(self.path) - - @staticmethod - def mkdir_all(path): - """Same as mkdir -p.""" - names = os.path.split(path) - p = '' - for name in names: - p = os.path.join(p, name) - Path._mkdir(p) - - @staticmethod - def _rmdir_all(dir_): - names = Path._listdir(dir_) - for name in names: - p = os.path.join(dir_, name) - if os.path.isdir(p): - Path._rmdir_all(p) - else: - Path._remove(p) - Path._rmdir(dir_) - - @staticmethod - def _mkdir(path): - Path._os_func(os.mkdir, path, errno.EEXIST) - - @staticmethod - def _rmdir(path): - Path._os_func(os.rmdir, path, errno.ENOENT) - - @staticmethod - def _remove(path): - Path._os_func(os.remove, path, errno.ENOENT) - - @staticmethod - def _listdir(path): - return Path._os_func(os.listdir, path, errno.ENOENT, default=[]) - - @staticmethod - def _os_func(func, path, errno, default=None): - """Call func(path) in an idempotent way. - - On exception ``ex``, if the type is OSError and ``ex.errno == errno``, - return ``default``, otherwise, re-raise - """ - try: - return func(path) - except OSError as e: - if e.errno == errno: - return default - else: - raise - - @staticmethod - def rm_all(path): - """Same as rm -r.""" - if os.path.isdir(path): - Path._rmdir_all(path) - else: - Path._remove(path) - - @staticmethod - def untar(path, into=None): - """Extract tarball at ``path`` into subdir ``into``. - - return subdir name if and only if there exists one, otherwise raise PathException - """ - args = ('tar', '-C', into, '-xzf', path, '--no-same-permissions') - subprocess.check_call(args, preexec_fn=lambda: os.umask(0o22)) - dirs = os.listdir(into) - if len(dirs) == 1: - return dirs[0] - else: - raise PathException('untar %s: expecting a single subdir, got %s' % (path, dirs)) - - @staticmethod - def tar(path, subdir, into=None, ts=None): - """Pack ``path`` into tarball ``into``.""" - # --sort=name requires a recent build of GNU tar - args = ['tar', '--numeric-owner', '--owner=0', '--group=0', '--sort=name'] - args += ['-C', path, '-cf', into, subdir] - envs = os.environ.copy() - if ts is not None: - args.append('--mtime=@%d' % ts) - if into.endswith('.xz'): - envs['XZ_OPT'] = '-7e' - args.append('-J') - elif into.endswith('.bz2'): - args.append('-j') - elif into.endswith('.gz'): - args.append('-z') - envs['GZIP'] = '-n' - else: - raise PathException('unknown compression type %s' % into) - subprocess.check_call(args, env=envs) - - -class GitHubCommitTsCache(object): - __cachef = 'github.commit.ts.cache' - __cachen = 2048 - - def __init__(self): - Path.mkdir_all(TMPDIR_DL) - self.cachef = os.path.join(TMPDIR_DL, self.__cachef) - self.cache = {} - - def get(self, k): - """Get timestamp with key ``k``.""" - fileno = os.open(self.cachef, os.O_RDONLY | os.O_CREAT) - with os.fdopen(fileno) as fin: - try: - fcntl.lockf(fileno, fcntl.LOCK_SH) - self._cache_init(fin) - if k in self.cache: - ts = self.cache[k][0] - return ts - finally: - fcntl.lockf(fileno, fcntl.LOCK_UN) - return None - - def set(self, k, v): - """Update timestamp with ``k``.""" - fileno = os.open(self.cachef, os.O_RDWR | os.O_CREAT) - with os.fdopen(fileno, 'wb+') as f: - try: - fcntl.lockf(fileno, fcntl.LOCK_EX) - self._cache_init(f) - self.cache[k] = (v, int(time.time())) - self._cache_flush(f) - finally: - fcntl.lockf(fileno, fcntl.LOCK_UN) - - def _cache_init(self, fin): - for line in fin: - k, ts, updated = line.split() - ts = int(ts) - updated = int(updated) - self.cache[k] = (ts, updated) - - def _cache_flush(self, fout): - cache = sorted(self.cache.iteritems(), cmp=lambda a, b: b[1][1] - a[1][1]) - cache = cache[:self.__cachen] - self.cache = {} - os.ftruncate(fout.fileno(), 0) - fout.seek(0, os.SEEK_SET) - for k, ent in cache: - ts = ent[0] - updated = ent[1] - line = '{0} {1} {2}\n'.format(k, ts, updated) - fout.write(line) - - -class DownloadMethod(object): - """Base class of all download method.""" - - def __init__(self, args): - self.args = args - self.urls = args.urls - self.url = self.urls[0] - self.dl_dir = args.dl_dir - - @classmethod - def resolve(cls, args): - """Resolve download method to use. - - return instance of subclass of DownloadMethod - """ - for c in DOWNLOAD_METHODS: - if c.match(args): - return c(args) - - @staticmethod - def match(args): - """Return True if it can do the download.""" - return NotImplemented - - def download(self): - """Do the download and put it into the download dir.""" - return NotImplemented - - -class DownloadMethodGitHubTarball(DownloadMethod): - """Download and repack archive tarabll from GitHub.""" - - __repo_url_regex = re.compile(r'^(?:https|git)://github.com/(?P[^/]+)/(?P[^/]+)') - - def __init__(self, args): - super(DownloadMethodGitHubTarball, self).__init__(args) - self._init_owner_repo() - self.version = args.version - self.subdir = args.subdir - self.source = args.source - self.commit_ts = None # lazy load commit timestamp - self.commit_ts_cache = GitHubCommitTsCache() - self.name = 'github-tarball' - - @staticmethod - def match(args): - """Match if it's a GitHub clone url.""" - url = args.urls[0] - proto = args.proto - if proto == 'git' and isinstance(url, basestring) \ - and (url.startswith('https://github.com/') or url.startswith('git://github.com/')): - return True - return False - - def download(self): - """Download and repack GitHub archive tarball.""" - self._init_commit_ts() - with Path(TMPDIR_DL, keep=True) as dir_dl: - # fetch tarball from GitHub - tarball_path = os.path.join(dir_dl.path, self.subdir + '.tar.gz.dl') - with Path(tarball_path, isdir=False): - self._fetch(tarball_path) - # unpack - d = os.path.join(dir_dl.path, self.subdir + '.untar') - with Path(d) as dir_untar: - tarball_prefix = Path.untar(tarball_path, into=dir_untar.path) - dir0 = os.path.join(dir_untar.path, tarball_prefix) - dir1 = os.path.join(dir_untar.path, self.subdir) - # submodules check - if self._has_submodule(dir0): - raise DownloadException('unable to fetch submodules\' source code') - # rename subdir - os.rename(dir0, dir1) - # repack - into=os.path.join(TMPDIR_DL, self.source) - Path.tar(dir_untar.path, self.subdir, into=into, ts=self.commit_ts) - # move to target location - file1 = os.path.join(self.dl_dir, self.source) - if into != file1: - shutil.move(into, file1) - - def _has_submodule(self, dir_): - m = os.path.join(dir_, '.gitmodules') - try: - st = os.stat(m) - return st.st_size > 0 - except OSError as e: - return e.errno != errno.ENOENT - - def _init_owner_repo(self): - url = self.url - m = self.__repo_url_regex.search(url) - if m is None: - raise DownloadException('invalid github url: %s' % url) - owner = m.group('owner') - repo = m.group('repo') - if repo.endswith('.git'): - repo = repo[:-4] - self.owner = owner - self.repo = repo - - def _init_commit_ts(self): - if self.commit_ts is not None: - return - url = self._make_repo_url_path('git', 'commits', self.version) - ct = self.commit_ts_cache.get(url) - if ct is not None: - self.commit_ts = ct - return - resp = self._make_request(url) - data = resp.read() - data = json.loads(data) - date = data['committer']['date'] - date = datetime.datetime.strptime(date, '%Y-%m-%dT%H:%M:%SZ') - date = date.timetuple() - ct = calendar.timegm(date) - self.commit_ts = ct - self.commit_ts_cache.set(url, ct) - - def _fetch(self, path): - """Fetch tarball of the specified version ref.""" - ref = self.version - url = self._make_repo_url_path('tarball', ref) - resp = self._make_request(url) - with open(path, 'wb') as fout: - while True: - d = resp.read(4096) - if not d: - break - fout.write(d) - - def _make_repo_url_path(self, *args): - url = '/repos/{0}/{1}'.format(self.owner, self.repo) - if args: - url += '/' + '/'.join(args) - return url - - def _make_request(self, path): - """Request GitHub API endpoint on ``path``.""" - url = 'https://api.github.com' + path - headers = { - 'Accept': 'application/vnd.github.v3+json', - 'User-Agent': 'OpenWrt', - } - req = urllib2.Request(url, headers=headers) - sslcontext = ssl._create_unverified_context() - fileobj = urllib2.urlopen(req, context=sslcontext) - return fileobj - - -class DownloadMethodCatchall(DownloadMethod): - """Dummy method that knows names but not ways of download.""" - - def __init__(self, args): - super(DownloadMethodCatchall, self).__init__(args) - self.args = args - self.proto = args.proto - self.name = self._resolve_name() - - def _resolve_name(self): - if self.proto: - return self.proto - methods_map = ( - ('default', ('@APACHE/', '@GITHUB/', '@GNOME/', '@GNU/', - '@KERNEL/', '@SF/', '@SAVANNAH/', 'ftp://', 'http://', - 'https://', 'file://')), - ('git', ('git://', )), - ('svn', ('svn://', )), - ('cvs', ('cvs://', )), - ('bzr', ('sftp://', )), - ('bzr', ('sftp://', )), - ('unknown', ('', )), - ) - for name, prefixes in methods_map: - if any(url.startswith(prefix) for prefix in prefixes for url in self.urls): - return name - - @staticmethod - def match(args): - """Return True.""" - return True - - def download(self): - """Not implemented. - - raise DownloadException - """ - raise DownloadException('download method for %s is not yet implemented' % self.name) - -# order matters -DOWNLOAD_METHODS = [ - DownloadMethodGitHubTarball, - DownloadMethodCatchall, -] - - -def main(): - parser = argparse.ArgumentParser() - parser.add_argument('action', choices=('dl_method', 'dl'), help='Action to take') - parser.add_argument('--urls', nargs='+', metavar='URL', help='Download URLs') - parser.add_argument('--proto', help='Download proto') - parser.add_argument('--subdir', help='Source code subdir name') - parser.add_argument('--version', help='Source code version') - parser.add_argument('--source', help='Source tarball filename') - parser.add_argument('--dl-dir', default=os.getcwd(), help='Download dir') - args = parser.parse_args() - if args.action == 'dl_method': - method = DownloadMethod.resolve(args) - sys.stdout.write(method.name + '\n') - elif args.action == 'dl': - method = DownloadMethod.resolve(args) - try: - method.download() - except Exception: - raise - -if __name__ == '__main__': - main()