Clean up and fix tools

* Make sync_third_party work in general
* Un-break build.py and run_hooks.py on windows
* Partially fix format.py on windows
* Reduce code duplication between run_hooks and sync_third_party

6 files changed, 219 insertions, 80 deletions

diff --git a/tools/build.py b/tools/build.py
index fb061eae3..8b33cef89 100755
--- a/tools/build.py
+++ b/tools/build.py
@@ -4,7 +4,8 @@ import argparse
 import os
 import sys
 from os.path import join
-from util import run
+from third_party import depot_tools_path, third_party_path, fix_symlinks, google_env
+from util import root_path, run
 import distutils.spawn
 
 parser = argparse.ArgumentParser(description='')
@@ -16,19 +17,13 @@ parser.add_argument(
     '--mode', default='debug', help='Build configuration: debug, release.')
 options, targets = parser.parse_known_args()
 
-root_path = os.path.dirname(os.path.dirname(os.path.realpath(__file__)))
-third_party_path = join(root_path, "third_party")
-depot_tools_path = join(third_party_path, "depot_tools")
-gn_path = join(depot_tools_path, "gn")
-ninja_path = join(depot_tools_path, "ninja")
-
-# Add third_party/depot_tools to PATH  because some google tools (e.g.
-# tool_wrapper, download_from_google_storage) use some google specific python
-# wrapper.
-os.environ["PATH"] = depot_tools_path + os.pathsep + os.environ["PATH"]
+fix_symlinks()
 
 os.chdir(root_path)
 
+gn_path = join(depot_tools_path, "gn")
+ninja_path = join(depot_tools_path, "ninja")
+
 if options.build_path:
     build_path = options.build_path
 else:
@@ -63,7 +58,7 @@ if not os.path.exists(args_filename) or options.args:
     with open(args_filename, "w+") as f:
         f.write("\n".join(gn_args) + "\n")
 
-run([gn_path, "gen", build_path])
+run([gn_path, "gen", build_path], env=google_env())
 
 target = " ".join(targets) if targets else ":all"
-run([ninja_path, "-C", build_path, target])
+run([ninja_path, "-C", build_path, target], env=google_env())
diff --git a/tools/format.py b/tools/format.py
index b994535f7..2bbc95bed 100755
--- a/tools/format.py
+++ b/tools/format.py
@@ -1,9 +1,10 @@
 #!/usr/bin/env python
 import os
-from util import run, find_exts
+from third_party import third_party_path, fix_symlinks, google_env
+from util import root_path, run, find_exts
+
+fix_symlinks()
 
-root_path = os.path.dirname(os.path.dirname(os.path.realpath(__file__)))
-third_party_path = os.path.join(root_path, "third_party")
 prettier = os.path.join(third_party_path, "node_modules", "prettier",
                         "bin-prettier.js")
 tools_path = os.path.join(root_path, "tools")
@@ -15,7 +16,7 @@ os.chdir(root_path)
 run(["clang-format", "-i", "-style", "Google"] + find_exts("src", ".cc", ".h"))
 
 for fn in ["BUILD.gn", ".gn"] + find_exts("build_extra", ".gn", ".gni"):
-    run(["third_party/depot_tools/gn", "format", fn])
+    run(["third_party/depot_tools/gn", "format", fn], env=google_env())
 
 # TODO(ry) Install yapf in third_party.
 run(["yapf", "-i"] + find_exts("tools/", ".py") +
diff --git a/tools/run_hooks.py b/tools/run_hooks.py
index 58178a69c..59161f5f2 100755
--- a/tools/run_hooks.py
+++ b/tools/run_hooks.py
@@ -1,29 +1,7 @@
 #!/usr/bin/env python
-import os
-import sys
-from util import run
+import third_party
 
-root_path = os.path.dirname(os.path.dirname(os.path.realpath(__file__)))
-third_party_path = os.path.join(root_path, "third_party")
-depot_tools_path = os.path.join(third_party_path, "depot_tools")
-os.chdir(root_path)
+third_party.fix_symlinks()
 
-
-def download(filename):
-    run([
-        "python",
-        os.path.join(depot_tools_path + '/download_from_google_storage.py'),
-        '--platform=' + sys.platform, '--no_auth', '--bucket=chromium-gn',
-        '--sha1_file',
-        os.path.join(root_path, filename)
-    ])
-
-
-if sys.platform == 'win32':
-    download("third_party/v8/buildtools/win/gn.exe.sha1")
-elif sys.platform == 'darwin':
-    download("third_party/v8/buildtools/mac/gn.sha1")
-elif sys.platform.startswith('linux'):
-    download("third_party/v8/buildtools/linux64/gn.sha1")
-
-run(['python', 'third_party/v8/tools/clang/scripts/update.py', '--if-needed'])
+third_party.download_gn()
+third_party.download_clang()
diff --git a/tools/sync_third_party.py b/tools/sync_third_party.py
index f32d669b2..cf0de30f3 100755
--- a/tools/sync_third_party.py
+++ b/tools/sync_third_party.py
@@ -1,35 +1,13 @@
 #!/usr/bin/env python
-# Only run this script if you are changing Deno's dependencies.
-
-import os
-from os.path import join
-from util import run, remove_and_symlink
-
-root_path = os.path.dirname(os.path.dirname(os.path.realpath(__file__)))
-third_party_path = join(root_path, "third_party")
-
-try:
-    os.makedirs(third_party_path)
-except:
-    pass
-os.chdir(third_party_path)
-
-# Run yarn to install JavaScript dependencies.
-remove_and_symlink("../package.json", "package.json")
-remove_and_symlink("../yarn.lock", "yarn.lock")
-run(["yarn"])
-# Run cargo to install Rust dependencies.
-run(["cargo", "fetch", "--manifest-path=" + root_path + "/Cargo.toml"],
-    envs={'CARGO_HOME': third_party_path + '/rust_crates'})
-# Run gclient to install other dependencies.
-run(["gclient", "sync", "--reset", "--shallow", "--no-history", "--nohooks"],
-    envs={'GCLIENT_FILE': root_path + "/gclient_config.py"})
-# TODO(ry) Is it possible to remove these symlinks?
-remove_and_symlink("v8/third_party/googletest", "googletest", True)
-remove_and_symlink("v8/third_party/jinja2", "jinja2", True)
-remove_and_symlink("v8/third_party/llvm-build", "llvm-build", True)
-remove_and_symlink("v8/third_party/markupsafe", "markupsafe", True)
-
+# Run this script if you are changing Deno's dependencies.
 # To update the deno_third_party git repo after running this, try the following:
 # cd third_party
 # find . -type f | grep -v "\.git" | xargs -I% git add -f --no-warn-embedded-repo "%"
+
+import third_party
+
+third_party.fix_symlinks()
+
+third_party.run_yarn()
+third_party.run_cargo()
+third_party.run_gclient_sync()
diff --git a/tools/third_party.py b/tools/third_party.py
new file mode 100644
index 000000000..ef8ed466f
--- /dev/null
+++ b/tools/third_party.py
@@ -0,0 +1,166 @@
+#!/usr/bin/env python
+# This script contains helper functions to work with the third_party subrepo.
+
+import os
+import sys
+from os import path
+from util import find_exts, make_env, remove_and_symlink, rmtree, root_path, run
+
+
+# Helper function that returns the full path to a subpath of the repo root.
+def root(*subpath_parts):
+    return path.normpath(path.join(root_path, *subpath_parts))
+
+
+# Helper function that returns the full path to a file/dir in third_party.
+def tp(*subpath_parts):
+    return root("third_party", *subpath_parts)
+
+
+third_party_path = tp()
+depot_tools_path = tp("depot_tools")
+rust_crates_path = tp("rust_crates")
+
+
+# This function creates or modifies an environment so that it matches the
+# expectations of various google tools (gn, gclient, etc).
+def google_env(env=None, merge_env={}, depot_tools_path=depot_tools_path):
+    env = make_env(env=env, merge_env=merge_env)
+    # Depot_tools to be in the PATH, before Python.
+    path_prefix = depot_tools_path + os.path.pathsep
+    if not env['PATH'].startswith(path_prefix):
+        env['PATH'] = path_prefix + env['PATH']
+    # We're not using Google's internal infrastructure.
+    if os.name == 'nt' and not 'DEPOT_TOOLS_WIN_TOOLCHAIN' in env:
+        env['DEPOT_TOOLS_WIN_TOOLCHAIN'] = "0"
+    return env
+
+
+def fix_symlinks():
+    # Ensure the third_party directory exists.
+    try:
+        os.makedirs(third_party_path)
+    except:
+        pass
+
+    # Make symlinks to Yarn metadata living in the root repo.
+    remove_and_symlink("../package.json", tp("package.json"))
+    remove_and_symlink("../yarn.lock", tp("yarn.lock"))
+
+    # TODO(ry) Is it possible to remove these symlinks?
+    remove_and_symlink("v8/third_party/googletest", tp("googletest"), True)
+    remove_and_symlink("v8/third_party/jinja2", tp("jinja2"), True)
+    remove_and_symlink("v8/third_party/llvm-build", tp("llvm-build"), True)
+    remove_and_symlink("v8/third_party/markupsafe", tp("markupsafe"), True)
+
+    # On Windows, git doesn't create the right type of symlink if the symlink
+    # and it's target are in different repos. Here we fix the symlinks that exist
+    # in the root repo while their target is in the third_party repo.
+    remove_and_symlink("third_party/node_modules", root("node_modules"), True)
+    remove_and_symlink("third_party/v8/build", root("build"), True)
+    remove_and_symlink("third_party/v8/buildtools", root("buildtools"), True)
+    remove_and_symlink("third_party/v8/build_overrides",
+                       root("build_overrides"), True)
+    remove_and_symlink("third_party/v8/testing", root("testing"), True)
+
+
+# Run Yarn to install JavaScript dependencies.
+def run_yarn():
+    run(["yarn"], cwd=third_party_path)
+
+
+# Run Cargo to install Rust dependencies.
+def run_cargo():
+    # Deletes the cargo index lockfile; it appears that cargo itself doesn't do it.
+    # If the lockfile ends up in the git repo, it'll make cargo hang for everyone
+    # else who tries to run sync_third_party.
+    def delete_lockfile():
+        lockfiles = find_exts(
+            path.join(rust_crates_path, "registry/index"), '.cargo-index-lock')
+        for lockfile in lockfiles:
+            os.remove(lockfile)
+
+    # Delete the index lockfile in case someone accidentally checked it in.
+    delete_lockfile()
+
+    run(["cargo", "fetch", "--manifest-path=" + root("Cargo.toml")],
+        cwd=third_party_path,
+        merge_env={'CARGO_HOME': rust_crates_path})
+
+    # Delete the lockfile again so it doesn't end up in the git repo.
+    delete_lockfile()
+
+
+# Run gclient to install other dependencies.
+def run_gclient_sync():
+    # Depot_tools will normally try to self-update, which will fail because
+    # it's not checked out from it's own git repository; gclient will then try
+    # to fix things up and not succeed, and and we'll end up with a huge mess.
+    # To work around this, we rename the `depot_tools` directory to
+    # `{root_path}/depot_tools_temp` first, and we set DEPOT_TOOLS_UPDATE=0 in
+    # the environment so depot_tools doesn't attempt to self-update.
+    # Since depot_tools is listed in .gclient_entries, gclient will install a
+    # fresh copy in `third_party/depot_tools`.
+    # If it all works out, we remove the depot_tools_temp directory afterwards.
+    depot_tools_temp_path = root("depot_tools_temp")
+
+    # Rename depot_tools to depot_tools_temp.
+    try:
+        os.rename(depot_tools_path, depot_tools_temp_path)
+    except:
+        # If renaming failed, and the depot_tools_temp directory already exists,
+        # assume that it's still there because a prior run_gclient_sync() call
+        # failed half-way, before we got the chance to remove the temp dir.
+        # We'll use whatever is in the temp dir that was already there.
+        # If not, the user can recover by removing the temp directory manually.
+        if path.isdir(depot_tools_temp_path):
+            pass
+        else:
+            raise
+
+    args = [
+        "gclient", "sync", "--reset", "--shallow", "--no-history", "--nohooks"
+    ]
+    envs = {
+        'DEPOT_TOOLS_UPDATE': "0",
+        'GCLIENT_FILE': root("gclient_config.py")
+    }
+    env = google_env(depot_tools_path=depot_tools_temp_path, merge_env=envs)
+    run(args, cwd=third_party_path, env=env)
+
+    # Delete the depot_tools_temp directory, but not before verifying that
+    # gclient did indeed install a fresh copy.
+    # Also check that `{depot_tools_temp_path}/gclient.py` exists, so a typo in
+    # this script won't accidentally blow out someone's home dir.
+    if (path.isdir(path.join(depot_tools_path, ".git"))
+            and path.isfile(path.join(depot_tools_path, "gclient.py"))
+            and path.isfile(path.join(depot_tools_temp_path, "gclient.py"))):
+        rmtree(depot_tools_temp_path)
+
+
+# Download gn from Google storage.
+def download_gn():
+    if sys.platform == 'win32':
+        sha1_file = "v8/buildtools/win/gn.exe.sha1"
+    elif sys.platform == 'darwin':
+        sha1_file = "v8/buildtools/mac/gn.sha1"
+    elif sys.platform.startswith('linux'):
+        sha1_file = "v8/buildtools/linux64/gn.sha1"
+
+    run([
+        "python",
+        tp('depot_tools/download_from_google_storage.py'),
+        '--platform=' + sys.platform,
+        '--no_auth',
+        '--bucket=chromium-gn',
+        '--sha1_file',
+        tp(sha1_file),
+    ],
+        env=google_env())
+
+
+# Download clang by calling the clang update script.
+def download_clang():
+    run(['python',
+         tp('v8/tools/clang/scripts/update.py'), '--if-needed'],
+        env=google_env())
diff --git a/tools/util.py b/tools/util.py
index 4938c5157..436630401 100644
--- a/tools/util.py
+++ b/tools/util.py
@@ -1,20 +1,30 @@
 # Copyright 2018 Ryan Dahl <ry@tinyclouds.org>
 # All rights reserved. MIT License.
 import os
+import shutil
+import stat
 import subprocess
 
 executable_suffix = ".exe" if os.name == "nt" else ""
+root_path = os.path.dirname(os.path.dirname(os.path.realpath(__file__)))
 
 
-def run(args, quiet=False, envs={}):
+def make_env(merge_env={}, env=None):
+    if env is None:
+        env = os.environ
+    env = env.copy()
+    for key in merge_env.keys():
+        env[key] = merge_env[key]
+    return env
+
+
+def run(args, quiet=False, cwd=None, env=None, merge_env={}):
+    args[0] = os.path.normpath(args[0])
     if not quiet:
         print " ".join(args)
-    env = os.environ.copy()
-    for key in envs.keys():
-        env[key] = envs[key]
-    args[0] = os.path.normpath(args[0])
+    env = make_env(env=env, merge_env=merge_env)
     shell = os.name == "nt"  # Run through shell to make .bat/.cmd files work.
-    subprocess.check_call(args, env=env, shell=shell)
+    subprocess.check_call(args, cwd=cwd, env=env, shell=shell)
 
 
 def remove_and_symlink(target, name, target_is_dir=False):
@@ -68,3 +78,14 @@ def find_exts(directory, *extensions):
                     matches.append(os.path.join(root, filename))
                     break
     return matches
+
+
+# The Python equivalent of `rm -rf`.
+def rmtree(directory):
+    # On Windows, shutil.rmtree() won't delete files that have a readonly bit.
+    # Git creates some files that do. The 'onerror' callback deals with those.
+    def rm_readonly(func, path, _):
+        os.chmod(path, stat.S_IWRITE)
+        func(path)
+
+    shutil.rmtree(directory, onerror=rm_readonly)