Rietveld Code Review Tool
Help | Bug tracker | Discussion group | Source code

Unified Diff: ensure_dependencies.py

Issue 29336327: Issue 2834 - Enable translations for adblockplusui (Closed)
Patch Set: Add link to Crowdin Created Feb. 15, 2016, 4:50 p.m.
Use n/p to move between diff chunks; N/P to move between comments.
Jump to:
View side-by-side diff with in-line comments
Download patch
« no previous file with comments | « dependencies ('k') | metadata.generic » ('j') | no next file with comments »
Expand Comments ('e') | Collapse Comments ('c') | Show Comments Hide Comments ('s')
Index: ensure_dependencies.py
diff --git a/ensure_dependencies.py b/ensure_dependencies.py
new file mode 100755
index 0000000000000000000000000000000000000000..707574dbbc614219b929c2354c4cf6f86790281f
--- /dev/null
+++ b/ensure_dependencies.py
@@ -0,0 +1,371 @@
+#!/usr/bin/env python
+# coding: utf-8
+
+# This Source Code Form is subject to the terms of the Mozilla Public
+# License, v. 2.0. If a copy of the MPL was not distributed with this
+# file, You can obtain one at http://mozilla.org/MPL/2.0/.
+
+import sys
+import os
+import posixpath
+import re
+import io
+import errno
+import logging
+import subprocess
+import urlparse
+import argparse
+
+from collections import OrderedDict
+from ConfigParser import RawConfigParser
+
+USAGE = """
+A dependencies file should look like this:
+
+ # VCS-specific root URLs for the repositories
+ _root = hg:https://hg.adblockplus.org/ git:https://github.com/adblockplus/
+ # File to update this script from (optional)
+ _self = buildtools/ensure_dependencies.py
+ # Clone elemhidehelper repository into extensions/elemhidehelper directory at
+ # tag "1.2".
+ extensions/elemhidehelper = elemhidehelper 1.2
+ # Clone buildtools repository into buildtools directory at VCS-specific
+ # revision IDs.
+ buildtools = buildtools hg:016d16f7137b git:f3f8692f82e5
+ # Clone the adblockplus repository into adblockplus directory, overwriting the
+ # usual source URL for Git repository and specifying VCS specific revision IDs.
+ adblockplus = adblockplus hg:893426c6a6ab git:git@github.com:user/adblockplus.git@b2ffd52b
+ # Clone the adblockpluschrome repository into the adblockpluschrome directory,
+ # from a specific Git repository, specifying the revision ID.
+ adblockpluschrome = git:git@github.com:user/adblockpluschrome.git@1fad3a7
+"""
+
+SKIP_DEPENDENCY_UPDATES = os.environ.get(
+ "SKIP_DEPENDENCY_UPDATES", ""
+).lower() not in ("", "0", "false")
+
+class Mercurial():
+ def istype(self, repodir):
+ return os.path.exists(os.path.join(repodir, ".hg"))
+
+ def clone(self, source, target):
+ if not source.endswith("/"):
+ source += "/"
+ subprocess.check_call(["hg", "clone", "--quiet", "--noupdate", source, target])
+
+ def get_revision_id(self, repo, rev=None):
+ command = ["hg", "id", "--repository", repo, "--id"]
+ if rev:
+ command.extend(["--rev", rev])
+
+ # Ignore stderr output and return code here: if revision lookup failed we
+ # should simply return an empty string.
+ result = subprocess.Popen(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE).communicate()[0]
+ return result.strip()
+
+ def pull(self, repo):
+ subprocess.check_call(["hg", "pull", "--repository", repo, "--quiet"])
+
+ def update(self, repo, rev, revname):
+ subprocess.check_call(["hg", "update", "--repository", repo, "--quiet", "--check", "--rev", rev])
+
+ def ignore(self, target, repo):
+
+ if not self.istype(target):
+
+ config_path = os.path.join(repo, ".hg", "hgrc")
+ ignore_path = os.path.abspath(os.path.join(repo, ".hg", "dependencies"))
+
+ config = RawConfigParser()
+ config.read(config_path)
+
+ if not config.has_section("ui"):
+ config.add_section("ui")
+
+ config.set("ui", "ignore.dependencies", ignore_path)
+ with open(config_path, "w") as stream:
+ config.write(stream)
+
+ module = os.path.relpath(target, repo)
+ _ensure_line_exists(ignore_path, module)
+
+ def postprocess_url(self, url):
+ return url
+
+class Git():
+ def istype(self, repodir):
+ return os.path.exists(os.path.join(repodir, ".git"))
+
+ def clone(self, source, target):
+ source = source.rstrip("/")
+ if not source.endswith(".git"):
+ source += ".git"
+ subprocess.check_call(["git", "clone", "--quiet", source, target])
+
+ def get_revision_id(self, repo, rev="HEAD"):
+ command = ["git", "rev-parse", "--revs-only", rev + '^{commit}']
+ return subprocess.check_output(command, cwd=repo).strip()
+
+ def pull(self, repo):
+ # Fetch tracked branches, new tags and the list of available remote branches
+ subprocess.check_call(["git", "fetch", "--quiet", "--all", "--tags"], cwd=repo)
+ # Next we need to ensure all remote branches are tracked
+ newly_tracked = False
+ remotes = subprocess.check_output(["git", "branch", "--remotes"], cwd=repo)
+ for match in re.finditer(r"^\s*(origin/(\S+))$", remotes, re.M):
+ remote, local = match.groups()
+ with open(os.devnull, "wb") as devnull:
+ if subprocess.call(["git", "branch", "--track", local, remote],
+ cwd=repo, stdout=devnull, stderr=devnull) == 0:
+ newly_tracked = True
+ # Finally fetch any newly tracked remote branches
+ if newly_tracked:
+ subprocess.check_call(["git", "fetch", "--quiet", "origin"], cwd=repo)
+
+ def update(self, repo, rev, revname):
+ subprocess.check_call(["git", "checkout", "--quiet", revname], cwd=repo)
+
+ def ignore(self, target, repo):
+ module = os.path.sep + os.path.relpath(target, repo)
+ exclude_file = os.path.join(repo, ".git", "info", "exclude")
+ _ensure_line_exists(exclude_file, module)
+
+ def postprocess_url(self, url):
+ # Handle alternative syntax of SSH URLS
+ if "@" in url and ":" in url and not urlparse.urlsplit(url).scheme:
+ return "ssh://" + url.replace(":", "/", 1)
+ return url
+
+repo_types = OrderedDict((
+ ("hg", Mercurial()),
+ ("git", Git()),
+))
+
+# [vcs:]value
+item_regexp = re.compile(
+ "^(?:(" + "|".join(map(re.escape, repo_types.keys())) +"):)?"
+ "(.+)$"
+)
+
+# [url@]rev
+source_regexp = re.compile(
+ "^(?:(.*)@)?"
+ "(.+)$"
+)
+
+def merge_seqs(seq1, seq2):
+ """Return a list of any truthy values from the suplied sequences
+
+ (None, 2), (1,) => [1, 2]
+ None, (1, 2) => [1, 2]
+ (1, 2), (3, 4) => [3, 4]
+ """
+ return map(lambda item1, item2: item2 or item1, seq1 or (), seq2 or ())
+
+def parse_spec(path, line):
+ if "=" not in line:
+ logging.warning("Invalid line in file %s: %s" % (path, line))
+ return None, None
+
+ key, value = line.split("=", 1)
+ key = key.strip()
+ items = value.split()
+ if not len(items):
+ logging.warning("No value specified for key %s in file %s" % (key, path))
+ return key, None
+
+ result = OrderedDict()
+ is_dependency_field = not key.startswith("_")
+
+ for i, item in enumerate(items):
+ try:
+ vcs, value = re.search(item_regexp, item).groups()
+ vcs = vcs or "*"
+ if is_dependency_field:
+ if i == 0 and vcs == "*":
+ # In order to be backwards compatible we have to assume that the first
+ # source contains only a URL/path for the repo if it does not contain
+ # the VCS part
+ url_rev = (value, None)
+ else:
+ url_rev = re.search(source_regexp, value).groups()
+ result[vcs] = merge_seqs(result.get(vcs), url_rev)
+ else:
+ if vcs in result:
+ logging.warning("Ignoring duplicate value for type %r "
+ "(key %r in file %r)" % (vcs, key, path))
+ result[vcs] = value
+ except AttributeError:
+ logging.warning("Ignoring invalid item %r for type %r "
+ "(key %r in file %r)" % (item, vcs, key, path))
+ continue
+ return key, result
+
+def read_deps(repodir):
+ result = {}
+ deps_path = os.path.join(repodir, "dependencies")
+ try:
+ with io.open(deps_path, "rt", encoding="utf-8") as handle:
+ for line in handle:
+ # Remove comments and whitespace
+ line = re.sub(r"#.*", "", line).strip()
+ if not line:
+ continue
+
+ key, spec = parse_spec(deps_path, line)
+ if spec:
+ result[key] = spec
+ return result
+ except IOError, e:
+ if e.errno != errno.ENOENT:
+ raise
+ return None
+
+def safe_join(path, subpath):
+ # This has been inspired by Flask's safe_join() function
+ forbidden = {os.sep, os.altsep} - {posixpath.sep, None}
+ if any(sep in subpath for sep in forbidden):
+ raise Exception("Illegal directory separator in dependency path %s" % subpath)
+
+ normpath = posixpath.normpath(subpath)
+ if posixpath.isabs(normpath):
+ raise Exception("Dependency path %s cannot be absolute" % subpath)
+ if normpath == posixpath.pardir or normpath.startswith(posixpath.pardir + posixpath.sep):
+ raise Exception("Dependency path %s has to be inside the repository" % subpath)
+ return os.path.join(path, *normpath.split(posixpath.sep))
+
+def get_repo_type(repo):
+ for name, repotype in repo_types.iteritems():
+ if repotype.istype(repo):
+ return name
+ return None
+
+def ensure_repo(parentrepo, parenttype, target, type, root, sourcename):
+ if os.path.exists(target):
+ return
+
+ if SKIP_DEPENDENCY_UPDATES:
+ logging.warning("SKIP_DEPENDENCY_UPDATES environment variable set, "
+ "%s not cloned", target)
+ return
+
+ postprocess_url = repo_types[type].postprocess_url
+ root = postprocess_url(root)
+ sourcename = postprocess_url(sourcename)
+
+ if os.path.exists(root):
+ url = os.path.join(root, sourcename)
+ else:
+ url = urlparse.urljoin(root, sourcename)
+
+ logging.info("Cloning repository %s into %s" % (url, target))
+ repo_types[type].clone(url, target)
+ repo_types[parenttype].ignore(target, parentrepo)
+
+def update_repo(target, type, revision):
+ resolved_revision = repo_types[type].get_revision_id(target, revision)
+ current_revision = repo_types[type].get_revision_id(target)
+
+ if resolved_revision != current_revision:
+ if SKIP_DEPENDENCY_UPDATES:
+ logging.warning("SKIP_DEPENDENCY_UPDATES environment variable set, "
+ "%s not checked out to %s", target, revision)
+ return
+
+ if not resolved_revision:
+ logging.info("Revision %s is unknown, downloading remote changes" % revision)
+ repo_types[type].pull(target)
+ resolved_revision = repo_types[type].get_revision_id(target, revision)
+ if not resolved_revision:
+ raise Exception("Failed to resolve revision %s" % revision)
+
+ logging.info("Updating repository %s to revision %s" % (target, resolved_revision))
+ repo_types[type].update(target, resolved_revision, revision)
+
+def resolve_deps(repodir, level=0, self_update=True, overrideroots=None, skipdependencies=set()):
+ config = read_deps(repodir)
+ if config is None:
+ if level == 0:
+ logging.warning("No dependencies file in directory %s, nothing to do...\n%s" % (repodir, USAGE))
+ return
+ if level >= 10:
+ logging.warning("Too much subrepository nesting, ignoring %s" % repo)
+ return
+
+ if overrideroots is not None:
+ config["_root"] = overrideroots
+
+ for dir, sources in config.iteritems():
+ if (dir.startswith("_") or
+ skipdependencies.intersection([s[0] for s in sources if s[0]])):
+ continue
+
+ target = safe_join(repodir, dir)
+ parenttype = get_repo_type(repodir)
+ _root = config.get("_root", {})
+
+ for key in sources.keys() + _root.keys():
+ if key == parenttype or key is None and vcs != "*":
+ vcs = key
+ source, rev = merge_seqs(sources.get("*"), sources.get(vcs))
+
+ if not (vcs and source and rev):
+ logging.warning("No valid source / revision found to create %s" % target)
+ continue
+
+ ensure_repo(repodir, parenttype, target, vcs, _root.get(vcs, ""), source)
+ update_repo(target, vcs, rev)
+ resolve_deps(target, level + 1, self_update=False,
+ overrideroots=overrideroots, skipdependencies=skipdependencies)
+
+ if self_update and "_self" in config and "*" in config["_self"]:
+ source = safe_join(repodir, config["_self"]["*"])
+ try:
+ with io.open(source, "rb") as handle:
+ sourcedata = handle.read()
+ except IOError, e:
+ if e.errno != errno.ENOENT:
+ raise
+ logging.warning("File %s doesn't exist, skipping self-update" % source)
+ return
+
+ target = __file__
+ with io.open(target, "rb") as handle:
+ targetdata = handle.read()
+
+ if sourcedata != targetdata:
+ logging.info("Updating %s from %s, don't forget to commit" % (source, target))
+ with io.open(target, "wb") as handle:
+ handle.write(sourcedata)
+ if __name__ == "__main__":
+ logging.info("Restarting %s" % target)
+ os.execv(sys.executable, [sys.executable, target] + sys.argv[1:])
+ else:
+ logging.warning("Cannot restart %s automatically, please rerun" % target)
+
+def _ensure_line_exists(path, pattern):
+ with open(path, 'a+') as f:
+ file_content = [l.strip() for l in f.readlines()]
+ if not pattern in file_content:
+ file_content.append(pattern)
+ f.seek(0, os.SEEK_SET)
+ f.truncate()
+ for l in file_content:
+ print >>f, l
+
+if __name__ == "__main__":
+ logging.basicConfig(format='%(levelname)s: %(message)s', level=logging.INFO)
+
+ parser = argparse.ArgumentParser(description="Verify dependencies for a set of repositories, by default the repository of this script.")
+ parser.add_argument("repos", metavar="repository", type=str, nargs="*", help="Repository path")
+ parser.add_argument("-q", "--quiet", action="store_true", help="Suppress informational output")
+ args = parser.parse_args()
+
+ if args.quiet:
+ logging.disable(logging.INFO)
+
+ repos = args.repos
+ if not len(repos):
+ repos = [os.path.dirname(__file__)]
+ for repo in repos:
+ resolve_deps(repo)
« no previous file with comments | « dependencies ('k') | metadata.generic » ('j') | no next file with comments »

Powered by Google App Engine
This is Rietveld