zulip/tools/webpack

147 lines
5.9 KiB
Plaintext
Raw Normal View History

#!/usr/bin/env python3
import argparse
import json
import os
import subprocess
from typing import NoReturn
os.chdir(os.path.join(os.path.dirname(__file__), '..'))
2015-10-26 17:11:44 +01:00
STATIC_PATH = 'static/'
def build_for_prod_or_casper(quiet: bool) -> NoReturn:
"""Builds for production, writing the output to disk"""
tests: Clean up calls to tools/webpack. Before this change, the way we loaded webpack for various tools was brittle. First, I addressed test-api and test-help-documentation. These tools used to be unable to run standalone on a clean provision, because they were (indirectly) calling tools/webpack without the `--test` option. The problem was a bit obscure, since running things like `./tools/test-backend` or `./tools/test-all` in your workflow would create `./var/webpack-stats-test.json` for the broken tools (and then they would work). The tools themselves weren't broken; they were the only relying on the common `test_server_running` helper. And even that helper wasn't broken; it was just that `run-dev.py` wasn't respecting the `--test` option. So I made it so that `./tools/run-dev` passes in `--test` to `./tools/webpack`. To confuse matters even more, for some reason Casper uses `./webpack-stats-production.json` via various hacks for its webpack configuration, so when I fixed the other tests, it broke Casper. Here is the Casper-related hack in zproject/test_settings.py, which was in place before my change and remains after it: if CASPER_TESTS: WEBPACK_FILE = 'webpack-stats-production.json' else: WEBPACK_FILE = os.path.join('var', 'webpack-stats-test.json') I added similar logic in tools/webpack: if "CASPER_TESTS" in os.environ: build_for_prod_or_casper(args.quiet) I also made the helper functions in `./tools/webpack` have nicer names. So, now tools should all be able to run standalone and not rely on previous tools creating webpack stats files for them and leaving them in the file system. That's good. Things are still a bit janky, though. It's not completely clear to me why `test-js-with-casper` should work off of a different webpack configuration than the other tests. For now most of the jankiness is around Casper, and we have hacks in two different places, `zproject/test_settings.py` and `tools/webpack` to force it to use the production stats file instead of the "test" one, even though Casper uses test-like settings for other things like which database you're using.
2018-09-06 21:01:45 +02:00
webpack_args = ['node', 'node_modules/.bin/webpack-cli',
'--config', 'tools/webpack.config.ts', '-p',
'--env', 'production']
if quiet:
webpack_args += ['--display', 'errors-only']
os.execvp(webpack_args[0], webpack_args)
2015-10-26 17:11:44 +01:00
def build_for_dev_server(host: str, port: str, minify: bool, disable_host_check: bool) -> None:
"""watches and rebuilds on changes, serving files from memory via webpack-dev-server"""
tests: Clean up calls to tools/webpack. Before this change, the way we loaded webpack for various tools was brittle. First, I addressed test-api and test-help-documentation. These tools used to be unable to run standalone on a clean provision, because they were (indirectly) calling tools/webpack without the `--test` option. The problem was a bit obscure, since running things like `./tools/test-backend` or `./tools/test-all` in your workflow would create `./var/webpack-stats-test.json` for the broken tools (and then they would work). The tools themselves weren't broken; they were the only relying on the common `test_server_running` helper. And even that helper wasn't broken; it was just that `run-dev.py` wasn't respecting the `--test` option. So I made it so that `./tools/run-dev` passes in `--test` to `./tools/webpack`. To confuse matters even more, for some reason Casper uses `./webpack-stats-production.json` via various hacks for its webpack configuration, so when I fixed the other tests, it broke Casper. Here is the Casper-related hack in zproject/test_settings.py, which was in place before my change and remains after it: if CASPER_TESTS: WEBPACK_FILE = 'webpack-stats-production.json' else: WEBPACK_FILE = os.path.join('var', 'webpack-stats-test.json') I added similar logic in tools/webpack: if "CASPER_TESTS" in os.environ: build_for_prod_or_casper(args.quiet) I also made the helper functions in `./tools/webpack` have nicer names. So, now tools should all be able to run standalone and not rely on previous tools creating webpack stats files for them and leaving them in the file system. That's good. Things are still a bit janky, though. It's not completely clear to me why `test-js-with-casper` should work off of a different webpack configuration than the other tests. For now most of the jankiness is around Casper, and we have hacks in two different places, `zproject/test_settings.py` and `tools/webpack` to force it to use the production stats file instead of the "test" one, even though Casper uses test-like settings for other things like which database you're using.
2018-09-06 21:01:45 +02:00
# This is our most dynamic configuration, which we use for our
# dev server. The key piece here is that we identify changes to
# files as devs make edits and serve new assets on the fly.
webpack_args = ['node', 'node_modules/.bin/webpack-dev-server']
webpack_args += [
'--config',
'tools/webpack.config.ts',
# webpack-cli has a bug where it ignores --watch-poll with
# multi-config, and we don't need the katex-cli part anyway.
'--config-name', 'frontend',
'--allowed-hosts', ','.join([host, '.zulipdev.com', '.zulipdev.org']),
'--host', host,
'--port', port,
# We add the hot flag using the cli because it takes care
# of addition to entry points and adding the plugin
# automatically
'--hot',
]
if minify:
webpack_args.append('--optimize-minimize')
if disable_host_check:
webpack_args.append('--disable-host-check')
# Tell webpack-dev-server to fall back to periodic polling on
# filesystems where inotify is known to be broken.
cwd = os.getcwd()
inotify_broken = False
with open('/proc/self/mounts') as mounts:
for line in mounts:
fields = line.split()
if fields[1] == cwd:
inotify_broken = fields[2] in ["nfs", "vboxsf"]
if inotify_broken:
webpack_args.append('--watch-poll=1000')
# TODO: This process should also fall back to periodic polling if
# inotify_broken.
import pyinotify
try:
webpack_process = subprocess.Popen(webpack_args)
class WebpackConfigFileChangeHandler(pyinotify.ProcessEvent):
def process_default(self, event: pyinotify.Event) -> None:
nonlocal webpack_process
print('Restarting webpack-dev-server due to config changes...')
webpack_process.terminate()
webpack_process.wait()
webpack_process = subprocess.Popen(webpack_args)
watch_manager = pyinotify.WatchManager()
event_notifier = pyinotify.Notifier(watch_manager, WebpackConfigFileChangeHandler())
for file in ['webpack.config.ts', 'webpack.assets.json']:
filepath = os.path.join(os.path.dirname(__file__), file)
watch_manager.add_watch(filepath, pyinotify.IN_MODIFY)
event_notifier.loop()
finally:
webpack_process.terminate()
webpack_process.wait()
2015-10-26 17:11:44 +01:00
def build_for_most_tests() -> None:
"""Generates a stub asset stat file for django so backend test can render a page"""
tests: Clean up calls to tools/webpack. Before this change, the way we loaded webpack for various tools was brittle. First, I addressed test-api and test-help-documentation. These tools used to be unable to run standalone on a clean provision, because they were (indirectly) calling tools/webpack without the `--test` option. The problem was a bit obscure, since running things like `./tools/test-backend` or `./tools/test-all` in your workflow would create `./var/webpack-stats-test.json` for the broken tools (and then they would work). The tools themselves weren't broken; they were the only relying on the common `test_server_running` helper. And even that helper wasn't broken; it was just that `run-dev.py` wasn't respecting the `--test` option. So I made it so that `./tools/run-dev` passes in `--test` to `./tools/webpack`. To confuse matters even more, for some reason Casper uses `./webpack-stats-production.json` via various hacks for its webpack configuration, so when I fixed the other tests, it broke Casper. Here is the Casper-related hack in zproject/test_settings.py, which was in place before my change and remains after it: if CASPER_TESTS: WEBPACK_FILE = 'webpack-stats-production.json' else: WEBPACK_FILE = os.path.join('var', 'webpack-stats-test.json') I added similar logic in tools/webpack: if "CASPER_TESTS" in os.environ: build_for_prod_or_casper(args.quiet) I also made the helper functions in `./tools/webpack` have nicer names. So, now tools should all be able to run standalone and not rely on previous tools creating webpack stats files for them and leaving them in the file system. That's good. Things are still a bit janky, though. It's not completely clear to me why `test-js-with-casper` should work off of a different webpack configuration than the other tests. For now most of the jankiness is around Casper, and we have hacks in two different places, `zproject/test_settings.py` and `tools/webpack` to force it to use the production stats file instead of the "test" one, even though Casper uses test-like settings for other things like which database you're using.
2018-09-06 21:01:45 +02:00
# Tests like test-backend, test-api, and test-home-documentation use
# our "test" configuration. The one exception is Casper, which uses
# our production configuration.
#
# It's not completely clear why we don't also use the same
# configuration for ALL tests, but figuring out the full history here
# was out of the scope of the effort here to add some comments and
# clean up names.
entries = {}
with open('tools/webpack.assets.json') as json_data:
for entry in json.load(json_data).keys():
entries[entry] = [{
"name": f"{entry}.js",
"publicPath": f"http://localhost:3000/webpack-stub/{entry}-stubentry.js",
"path": f"/stubfolder/{entry}-stubfile.js",
}]
stat_data = {
"status": "done",
"chunks": entries,
"entryPoints": {name: [chunk] for name, chunk in entries.items()},
}
directory = 'var'
if not os.path.exists(directory):
os.makedirs(directory)
with open(os.path.join(directory, 'webpack-stats-test.json'), 'w') as outfile:
json.dump(stat_data, outfile)
parser = argparse.ArgumentParser()
parser.add_argument('--test',
action='store_true', dest='test', default=False,
help='generate a stub webpack-stats.json file (for backend testing)')
parser.add_argument('--quiet',
action='store_true', dest='quiet', default=False,
help='Minimizes webpack output while running')
parser.add_argument('--watch',
action='store_true', dest='watch', default=False,
help='watch for changes to source files (for development)')
parser.add_argument('--host',
action='store', dest='host',
default='127.0.0.1', help='set the host for the webpack server to run on')
parser.add_argument('--port',
action='store', dest='port',
default='9994', help='set the port for the webpack server to run on')
parser.add_argument('--minify',
action='store_true', dest='minify', default=False,
help='Minify and optimize the assets (for development)')
parser.add_argument('--disable-host-check',
action='store_true', dest='disable_host_check', default=None,
help='Disable host check for webpack-dev-server')
args = parser.parse_args()
tests: Clean up calls to tools/webpack. Before this change, the way we loaded webpack for various tools was brittle. First, I addressed test-api and test-help-documentation. These tools used to be unable to run standalone on a clean provision, because they were (indirectly) calling tools/webpack without the `--test` option. The problem was a bit obscure, since running things like `./tools/test-backend` or `./tools/test-all` in your workflow would create `./var/webpack-stats-test.json` for the broken tools (and then they would work). The tools themselves weren't broken; they were the only relying on the common `test_server_running` helper. And even that helper wasn't broken; it was just that `run-dev.py` wasn't respecting the `--test` option. So I made it so that `./tools/run-dev` passes in `--test` to `./tools/webpack`. To confuse matters even more, for some reason Casper uses `./webpack-stats-production.json` via various hacks for its webpack configuration, so when I fixed the other tests, it broke Casper. Here is the Casper-related hack in zproject/test_settings.py, which was in place before my change and remains after it: if CASPER_TESTS: WEBPACK_FILE = 'webpack-stats-production.json' else: WEBPACK_FILE = os.path.join('var', 'webpack-stats-test.json') I added similar logic in tools/webpack: if "CASPER_TESTS" in os.environ: build_for_prod_or_casper(args.quiet) I also made the helper functions in `./tools/webpack` have nicer names. So, now tools should all be able to run standalone and not rely on previous tools creating webpack stats files for them and leaving them in the file system. That's good. Things are still a bit janky, though. It's not completely clear to me why `test-js-with-casper` should work off of a different webpack configuration than the other tests. For now most of the jankiness is around Casper, and we have hacks in two different places, `zproject/test_settings.py` and `tools/webpack` to force it to use the production stats file instead of the "test" one, even though Casper uses test-like settings for other things like which database you're using.
2018-09-06 21:01:45 +02:00
if "CASPER_TESTS" in os.environ:
build_for_prod_or_casper(args.quiet)
elif args.test:
build_for_most_tests()
elif args.watch:
tests: Clean up calls to tools/webpack. Before this change, the way we loaded webpack for various tools was brittle. First, I addressed test-api and test-help-documentation. These tools used to be unable to run standalone on a clean provision, because they were (indirectly) calling tools/webpack without the `--test` option. The problem was a bit obscure, since running things like `./tools/test-backend` or `./tools/test-all` in your workflow would create `./var/webpack-stats-test.json` for the broken tools (and then they would work). The tools themselves weren't broken; they were the only relying on the common `test_server_running` helper. And even that helper wasn't broken; it was just that `run-dev.py` wasn't respecting the `--test` option. So I made it so that `./tools/run-dev` passes in `--test` to `./tools/webpack`. To confuse matters even more, for some reason Casper uses `./webpack-stats-production.json` via various hacks for its webpack configuration, so when I fixed the other tests, it broke Casper. Here is the Casper-related hack in zproject/test_settings.py, which was in place before my change and remains after it: if CASPER_TESTS: WEBPACK_FILE = 'webpack-stats-production.json' else: WEBPACK_FILE = os.path.join('var', 'webpack-stats-test.json') I added similar logic in tools/webpack: if "CASPER_TESTS" in os.environ: build_for_prod_or_casper(args.quiet) I also made the helper functions in `./tools/webpack` have nicer names. So, now tools should all be able to run standalone and not rely on previous tools creating webpack stats files for them and leaving them in the file system. That's good. Things are still a bit janky, though. It's not completely clear to me why `test-js-with-casper` should work off of a different webpack configuration than the other tests. For now most of the jankiness is around Casper, and we have hacks in two different places, `zproject/test_settings.py` and `tools/webpack` to force it to use the production stats file instead of the "test" one, even though Casper uses test-like settings for other things like which database you're using.
2018-09-06 21:01:45 +02:00
build_for_dev_server(args.host, args.port, args.minify, args.disable_host_check)
else:
tests: Clean up calls to tools/webpack. Before this change, the way we loaded webpack for various tools was brittle. First, I addressed test-api and test-help-documentation. These tools used to be unable to run standalone on a clean provision, because they were (indirectly) calling tools/webpack without the `--test` option. The problem was a bit obscure, since running things like `./tools/test-backend` or `./tools/test-all` in your workflow would create `./var/webpack-stats-test.json` for the broken tools (and then they would work). The tools themselves weren't broken; they were the only relying on the common `test_server_running` helper. And even that helper wasn't broken; it was just that `run-dev.py` wasn't respecting the `--test` option. So I made it so that `./tools/run-dev` passes in `--test` to `./tools/webpack`. To confuse matters even more, for some reason Casper uses `./webpack-stats-production.json` via various hacks for its webpack configuration, so when I fixed the other tests, it broke Casper. Here is the Casper-related hack in zproject/test_settings.py, which was in place before my change and remains after it: if CASPER_TESTS: WEBPACK_FILE = 'webpack-stats-production.json' else: WEBPACK_FILE = os.path.join('var', 'webpack-stats-test.json') I added similar logic in tools/webpack: if "CASPER_TESTS" in os.environ: build_for_prod_or_casper(args.quiet) I also made the helper functions in `./tools/webpack` have nicer names. So, now tools should all be able to run standalone and not rely on previous tools creating webpack stats files for them and leaving them in the file system. That's good. Things are still a bit janky, though. It's not completely clear to me why `test-js-with-casper` should work off of a different webpack configuration than the other tests. For now most of the jankiness is around Casper, and we have hacks in two different places, `zproject/test_settings.py` and `tools/webpack` to force it to use the production stats file instead of the "test" one, even though Casper uses test-like settings for other things like which database you're using.
2018-09-06 21:01:45 +02:00
build_for_prod_or_casper(args.quiet)