This commit fixes some modules that were erroneously left out while
transitioning app.js to webpack. This commit exposes them using
expose-loader or setting them directly to window.
This commit moves all files previously under the 'app' bundle in
the Django pipeline to being compiled by webpack under the 'app'
entry point. In the process, it moves assets under the app entry
to a file called app.js that consumes all relevant css and js files.
This commit also edits the webpack config to be able to expose certain
variables for third party libraries that are currently required by
some modules. This is bad coding form and should be refactored to
requiring whatever dependencies a module may have; we're just
deferring that to the future to simplify the series of transitions we
need to do here. The variable exposure is done using expose-loader in
webpack.
The app/index.html template is edited to override the newly introduced
'commonjs' block in the base template. This is done as a temporary
measure so as not to disrupt other pages on the app during the transition.
It also fixes the value of the 'this' context that was being inferred
as window by third party libraries. This is done using imports-loader
in the webpack config. This is also messy and probably isn't how we
want things to work long term.
The only changes visible at the AST level, checked using
https://github.com/asottile/astpretty, are
zerver/lib/test_fixtures.py:
'\x1b\\[(1|0)m' ↦ '\\x1b\\[(1|0)m'
'\\[[X| ]\\] (\\d+_.+)\n' ↦ '\\[[X| ]\\] (\\d+_.+)\\n'
which is fine because re treats '\\x1b' and '\\n' the same way as
'\x1b' and '\n'.
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Adds search_pill.js to the static asset pipeline. The items
for search pill contain 2 keys, display_value and search_string.
Adding all the operator information i.e the operator, operand and
negated fields along with the search_string and description was tried out.
It was dropped because it didn't provide any advantage as one had to
always calculate the search_string and the description from the operator.
The appropriate name for the remote pointing at the repo we maintain
may be `upstream` for most of our repos... but not when we're
downstream of someone else, e.g. for react-native. So, make it easy
to configure per-repo.
This results in a significant optimization in the performance of
re-provisioning Zulip if all that you're doing is rebasing onto a
newer version of master (which just adds new migrations).
The change carries some risk of generating unpleasant-to-debug
situations, because if we merge a buggy migration and then later fix
it, some clients may not have a properly migrated database (and also,
this changes how populate_db commutes with migrations). But it seems
worth it, given how much time is currently wasted by not having this.
Fixes: #9512.
In this commit we are adding run_generate_fixtures_if_required,
a new function which is meant to de-duplicate a bit of code
between test-server and test-backend which is essentially
responsible for rebuilding the test database if that was required.
In this commit we are essentially just refactoring the function
is_template_database_current to be called template_database_status
and adjusting the return values accordingly.
This is essentially a preparatory commit for the upcoming commits
which will essentially enable us to not throw away entire DB and
rebuild from scratch if only running migrations could do the job.
This should avoid us creating duplicate webpack bundles every time we
do a deployment, even if none of the files in the bundles themselves
have changed at all.
This option (aka `--raw-output`) prints a string as itself, rather
than JSON-encoded; which makes it fit a bit better in a shell script,
saving us a layer of quoting.
This replaces ad4617c95 with a different fix for the same issue:
instead of stripping the `.git` off separately, we can just correct
the regex, using `+?` to fix our stepping in a classic regex pitfall.
This is a performance optimization: Rather than copying these files
into the `prod-static` directory and then deleting them, we just don't
copy them over in the first place.
For styles, it might have once been the case that this did something,
but we've moved them all to being managed by webpack some time ago.
For the js directory, I think it was never useful to copy and then
delete them; these files were always compiled via tools/minify-js,
and the raw JS files weren't needed, anyway.
This changes run-dev.py to ensure that we have in fact compiled
handlebars templates before running webpack, which is the right model.
Future work will likely include running the handlebars compiler from
webpack, and thus eliminating this extra process.
This improves the performance of these operations, by saving a ~50ms
Python process startup. While not a major performance improvement, it
seems worth it, given how often these commands get run.
Fixes#9571.
First, it's silly that these weren't in common.css in the first place,
since that meant these were a bunch of duplicated code, but
additionally, that meant that these weren't available on the
`/activity` page (or other pages that don't include the portico styles).
Fixes#9561.
Makes the i18n strings in this file much easier to translate by splitting
them into smaller chunks (which avoids having a lot of code in the tagged
strings), and adds a string that was missing as well.
We fix the issue of check-templates spitting out diff between
expected and found indentation of a file before mentioning the
error message and the file name. Basically stuff was being in the
wrong order despite the fact that in code stuff was happening in the
correct order ie, first print the error message along with the filename
and then the actual diff between expected and found file indentation.
Fixes: #9533.
A "zform" knows how to render data that follows our
schema for widget messages with form elements like
buttons and choices.
This code won't be triggered until a subsequent
server-side commit takes widget_content from
API callers such as the trivial chat bot and
creates submessages for us.
This starts the concept of a schema checker, similar to
zerver/lib/validator.py on the server. We can use this
to validate incoming data. Our server should filter most
of our incoming data, but it's useful to have client-side
checking to defend against things like upgrade
regressions (i.e. what if we change the name of the field
on the server side without updating all client uses).
I mistakenly pushed a PR when my tests failed. I ran with
the coverage option, so I saw this brightly colored summary
report that distracted me from the failure message.
This adds a couple newlines and some all caps.
The timezone environment variable was set to UTC initially. It was
changed to something other than UTC so that any local vs UTC
conversion issues will manifest in the tests.
Fixes: #5105.
We essentially stop running create_realm_internal_bots during
every provisioing and move its operations to run from populate db.
In fact to speed things up a bit we actually make populate db call the
funcs which create_realm_internal_bots calls behind the scenes.
Fixes: #9467.
This is required because the --settings=zproject.test_settings param
doesn't work with migrate or the dumpdata management commands. Thus
untill now if one ran just this tool ended up with test database not
properly setup. We never noticed this because test-backend ran this
tool again (after exporting DJANGO_SETTINGS_MODULE) thus making the
tool work this time.
I've often done this by hand -- basically typed out the last line,
with the variables found from looking at the PR page in a browser.
Seems nicer for both us maintainers and the contributor, in particular
because the PR gets marked as merged, instead of closed. But it's a
bit of a pain, and I do it maybe half the time or less; plus it's kind
of a subtle GitHub feature, and as a result I think other maintainers
of Zulip repos do this approximately never.
I've always figured this couldn't be hard to automate; today I decided
to take the 45 minutes to look up how, write out the script, QA it,
write up a nice usage message and some comments, and commit it. :)
This commit improves the output that blueslip produces while
showing error stack traces on the front-end. This is done by
using a library called error-stack-parser to format the stack
traces.
This commit also edits the webpack config to use a different
devtool setting since the previous one did not support sourcemaps
within stack traces. It also removes a plugin that was obviated
by this change.
This improves test coverage for a lot of our webhooks that relied
on ad-hoc methods to handle unexpected event types.
Note that I have deliberately skipped github_legacy, it isn't
advertised and is officially deprecated.
Also, I have refrained from making further changes to Trello, I
believe further improvements to test coverage should be covered
in separate per-webhook commits/PRs.
UnexpectedWebhookEventType is a generic exception that we may
now raise when we encounter a webhook event that is new or one
that we simply aren't aware of.
Now that we have tsearch_extras packages uploaded, this mostly works.
There's a few issues being debugged in #9460; they should be fixed
soon, and regardless, merging this will make it easier to develop.
This makes it possible to again use the *.zulipdev.com domains in the
development environment.
Ideally, we'd also read REALM_HOSTS to make this more flexible.
This adds a tour of Zulip to the bottom of the homepage.
In order to get the carousel nave, we use Bootstrap 2 from a CDN on
this page; this isn't ideal in the medium term, but upgrading
Bootstrap across the project is too much work for now.
Apparently, we were incorrectly appending each new hash onto the end
of the file, basically resulting in every run of provision being
treated as a miss for this cache.
Fixing this saves about 4s (over 1/3) of the no-op provision time.
Fixes#9233.
Uses nargs='*' instead of nargs='argparse.REMAINDER'.
nargs='argparse.REMAINDER' gathers remaining terms as arguments
even if it is an option e.g --coverage, while '*' gathers all the
command-line arguments until the next option is encountered.
The only slash command implemented in this initial
version is an extremely crippled version of a
"/stats" slash command that reports that you are
running 1 server.
This has a cool structure, but it's written against the long-dead
South API, and we can always pull it out of the Git history if we want
to use this approach in the future.
This leaves the wrapper script with very little left to do!
The main thing left is finding scripts by searching for shebang lines;
mypy itself would happily do the search for importable Python files.
Cleaned up add_user_list_args(). The "help" and
"all_users_help" have all default values. As noted in
an earlier commit, "all_users_help" is always passed in,
so we can get rid of "all_users_arg". We keep the default
for "all_users_help" so we don't have to change variable order
in function definition.
Since the region of base.zulipdev.org is
NYC3 we had to add SFO1 as an additional
region each time a snapshot of base droplet
is created. This is required as droplets
can be created in SFO1 only if there is
an image present in that region. Adding
of droplet image to SFO1 takes a lot of
time as well as cost 2X as we are storing
2 images. It's better to just create new
droplets in NYC3 instead. Alternatively we
can create a new base droplet in SFO1 if
we want all the droplets to be created in
SFO1.
This puts all of this config in one place, and also needs a lot fewer
lines to describe it; which, combined, makes it a lot clearer what our
normal config actually is. (I'd been looking at this script for a few
minutes without realizing that we have `--disallow-untyped-defs` *on*
by default, not off.)
Experimenting with different values is still easy; just comment the
line in the config.
This commit moves the stylesheets under the archive bundle in
the Django pipeline to being compiled by webpack instead. It
also removes a remaining call to a portico stylesheet that no
longer exists.
This commit transitions landing-page.css from the Django pipeline
to being compiled by webpack as landing-page.scss under the
'landing-page' and 'integration' bundles.
This commit transitions all styles in app.css in the Django pipeline
to being compiled by webpack in an app-styles bundle, and renames the
various files to now be processed as SCSS.
To implement this transition, we move the old CSS file refernces in
settings.py and replace them with a bundle declared in
`webpack.assets.json` and includedn in the index.html template
Tweaked by tabbott to keep the list of files in `app.css` in
`webpack.assets.json`, and to preserve the ordering from the old
`settings.py`.
This commit removes the flash on unstyled content while in dev
mode that was caused by the use of style-loader. Instead it
enables mini-css-extract-plugin in dev in combination with
css-hot-loader which enables HMR for development.
This is because mini-css-extract-plugin does not currently support
HMR out of the box. It also adds a SourceMapDevtoolPlugin to enable
sourcemaps with css since mini-css breaks sourcemaps when used in
combination with the cheap-module-evel-source-map setting.
Related issues:
https://github.com/webpack-contrib/mini-css-extract-plugin/issues/34https://github.com/webpack-contrib/mini-css-extract-plugin/issues/29
We modify check-templates to check for duplicate id's in archive
templates and app templates separately. This means we are allowing
app and archive templates to potentially use same id's. This is
needed because we intend to re use some js from the main app and
having same id's help achieve that.
Note: We haven't up until this point actually added archive
templates. This commit is more of a preparatory commit for merging
the basic archive infra.
Adds support for importing png files using file-loader in webpack.
Changes the name of the output directory to be files instead of
fonts for better readability.
This commit removes the need for portico.css to be generated
by the Django pipeline and makes the error page use the css
file compiled by webpack instead.
Combines, both portico js and css into one bundle. This for now solve
the issue of an empty js bundle being generated by webpack for the
portico-styles stylesheet.
This was a bit more than moving code. I extracted the
following things:
$widget (and three helper methods)
$input
text()
empty()
expand_column
close_widget
activity.clear_highlight
There was a minor bug before this commit, where we were inconsistent
about trimming spaces. The introduction of text() and empty() should
prevent bugs where users type the space bar into search.
This commit adds a --quiet argument to tools/webpack which removes
the verbose output from webpack and replaces it with showing only
errors. It also makes tools/run-dev --tests use this argument while
running webpack for testing.
Tweaked by tabbott to clean up the code a bit.
static/styles/scss/portico.scss is now compiled by webpack
and supports SCSS syntax.
Changed the server-side templates to render the portico-styles
bundle instead of directly requiring the portico stylesheet. This
allows webpack to handle stylesheet compilation and minification.
We use the mini-css-extract-plugin to extract out css from the
includes in webpack and let webpacks production mode handle
minification. Currently we're not able to use it for dev mode
because it does not support HMR so we use style-loader instead.
Once the plugin supports HMR we can go on to use it for both
dev and prod.
The downside of this is that when reloading pages in the development
environment, there's an annoying flash of unstyled content :(.
It is now possible to make a change in any of the styles included
by static/styles/scss/portico.scss and see the code reload live
in the browser. This is because style-loader which we currently
use has the module.accept code built-in.
This commit fixes hot module replacement in webpack. To do this
we open port 9994 used by webpack to communicate between browser
and devserver. The attempts to forward the proxy from 9991 failed
so the last resort was to open up the webpack port.
It also removes an uncessary plugin in the webpack config and moves
the --hot flag to tools/webpack.
We've already got a bunch of other comments on work we need to do for
this decorator and an open issue that will ensure we at some point
rework this and add tests for it. In the meantime, I'd like to lock
down the rest of decorator.py at 100% coverage.
Fixes#1000.
Webpack dev server by default does host checking for requests. so
in dev enviorment if the the request came for zulipdev.com it would not
send js files which caused dev envoirment to not work.
This should make it easier to find the templates that are actually
part of the core webapp, instead of having them all mixed together
with the portico pages.
Update the config file to show slightly more information while
compiling webpack. Also decreased the time webpack waits before
recompiling in order to speed up HMR.
Changed the devtoool setting for development from 'eval' to
'cheap-module-eval-source-map' as it has better support for
breakpoints in Google Chrome and the difference is time is
negligible at the number and size of files currently being
consumed by webpack. This stragtegy can be reviewed in the
future as the size of files grows or Chrome adds better
support.
Upgrade webpack to latest version at the time of authoring. This
involves upgrading webpack version and its loaders to compatible
versions. It also involved editing tools/webpack to use the
executable for webpack-cli instead because of a change in how the
webpack package wants you to handle shell execution.
It also fixes the confugration for TypeScript in the webpack config
as that was previously broken. Including TypeScript files in JS
files compiled by webpack now works.
This prevents us from using const in our JS code, with exceptions
for test code and the portico. Hopefully this is just a temporary
rule until we make our pipelines with work with ES6.
I tried to prevent "let", but that was too noisy.
This adjusts the one false-negative case of using const in a comment.
We now allow you to run --coverage on individual files. This helps
when you want to make sure a file is being covered directly and not
just getting incidental coverage from higher level tests.
Before this commit, we were conflating wanting coverage reports with
wanting coverage checks. For individual files, we now solve that by
simply eliminating the coverage checks. This required some minor
refactoring to extract some functions.
This fixes a bug where provision was failing since our most recent
upgrade to yarn/nvm/node.
It turns out my original fix was the correct fix, but to the wrong
third-party tool: nvm, not yarn, was the offender.
String.prototype.endsWith is not supported in ie11.
Adds string.prototype.endswith package to dependencies and places
it at `common` entry point in webpack.assets.json.
Fixes#8944.
Adds string.prototype.startswith package to dependencies and places
it at `common` entry point in webpack.assets.json. As common.js is
loaded on all code paths first, there is no need to place this package
into other entry points.
We make some specific cases of tags use 2 space indents.
The case description:
* A tag with opening tag spread over multiple lines and closing tag
on the same line as of the closing angle bracket of the opening tag.
* A tag with opening tag spread over multiple lines and closing tag
not on the same line as of the closing angle bracket of the opening
tag.
Example:
Case 1:
Not linted:
<button type="button"
class="btn btn-primary btn-small">{{t "Yes" }}</button>
After linting:
<button type="button"
class="btn btn-primary btn-small">{{t "Yes" }}</button>
Case 2:
Before linting:
<div class = "foo"
id = "bar"
role = "whatever">
{{ bla }}
</div>
After linting:
<div class = "foo"
id = "bar"
role = "whatever">
{{ bla }}
</div>
After some thinking, I don't think there's any actual value to doing
the ../ style relative links here, whereas there is actual harm from
the links being slightly broken in the current model. We fix this by
just using /#settings as the URL.
Fixes#8978.
Added support for passing a filename without `.js` suffix.
This then fixed the issue of no complaints for invalid test
files. Now, throws an error for invalid test files.
Fixes#8579.
Update perfect-scrollbar to fix stutter space-scrolling in #8544. Also
reworked deprecated `element.perfectScrollbar` to `new
PerfectScrollbar(element)`. Lastly, updated provision version and
changed node module path to new path.
This also refactors perfect-scrollbar in help.js to work with updated
version of perfect-scrollbar. Because the update also changed
perfect-scrollbar's css selectors for all scrollbars in zulip, we
update those too.
Fixes#8544.
Namely, annotate as best as possible, and add notes to indicate preference,
if QuerySet develops generic typing.
Note that the return values of functions with annotations changed in this
commit are used elsewhere as QuerySets, so the Sequence[T] approach used
for some functions in models.py is not applicable.
Other functions took the form of returning Sequence[T] when the QuerySet
functionality is unused beyond the function, with T being the objects
filtered for in the function body; this commit follows that practice for the
one remaining python2 comment-annotated function, completing the transition
of models.py to py3.5 function annotations.
A note is also added to another function regarding a need to return a
QuerySet, and ideally a QuerySet[T] in line with the other functions, as and
when QuerySet becomes annotated as a generic.
This commit switches our emoji infrastructure to use 256 color indexed
64px spritesheets. Earlier we were using non-indexed 32px spritesheets
which were blurry on high dpi displays. These indexed spritesheets not
only provide a crispier display but are also smaller in size.
This commit also removes the `emoji-datasource` package as a dependency
as all the data is now sourced from individual datasource packages.
Fixes: #7862.
Main exceptions are scripts/tools/puppet & tests.
Other current exclusions:
- api_test_helpers.py (avoid changing test code);
- cache.py, due to comments in file;
- models.py, due to failure on QuerySet[Message];
- stream_subscription.py, due to failure on QuerySet[Subscription];
- tornado/descriptors.py;
- views/streams.py, due to remaining FuncKwargPair issue;
- zthumbor/, since thumbor is in python2.
Tweaked by tabbott to partially document the stauts in comments.
This flips the experimental `--express` option to be the default.
We retain the old behavior, where the script exits before
`initialize-database`, as an option `--no-init-db`; it might be useful
in e.g. a migration scenario (from a Zulip install elsewhere, or
another chat system) where the admin wants to set up the database
separately.
The install instructions are adjusted to match, getting shorter by two
steps and a bunch of words. I think this opens up opportunities to
refactor the text to simplify things further, too, but leaving that
for another commit.
Also tweak the "production" test suite to match.
This replaces the cumbersome system we had for giving users feedback
on settings state changes in the display settings UI.
We expect this new system to be what we will attempt to migrate other
settings widgets to match over the coming weeks and months. It also
provides the opportunity to significant refactor away a lot of the
code duplication in settings_display.js.
Thanks to Brock Whittaker for redoing the styling and improving the
code simplicity.
Fixes#7622.
These two classes are tricky to test, and nocoverage-ing them
allows us to mark queue_processors.py as fully covered. We
still want to cover these two workers at some point, but for
now, it's nice to enforce full coverage for any future changes
to queue_processors.py.
Fixes (sort of) #6542.
This should help prevent against bugs where we accidentally introduce
use of sudo somewhere in the production installer or upgrade code path
(these used to happen all the time), which doesn't work on production
systems that don't have sudo setup.
Revert c8f034e9a "queue: Remove missedmessage_email_senders code."
As the comment in the code says, it ensures a smooth upgrade path
from 1.7.x; we can delete it in master after 1.8.0 is released.
The removal commit was merged early due to a communication failure.
This commit adds tests for the fixture for when a user is not
authorized (perhaps because the query requires the use of admin
privileges) for a particular query.
In templates/zerver/api/update-message.md, we have a sample fixture
for when a zulip.Client does not have the permission to update/edit
a particular message. This commit adds a test for that fixture.
Also, tools/test-api now also uses a non-admin client for this test,
which might come in handy in the future.
We now isolate the code to transmit messages into transmit.js.
It is stable code that most folks doing UI work in compose.js don't
care about the details of, so it's just clutter there. Also, we may
soon have other widgets than the compose box that send messages.
This change mostly preserves test coverage, although in some cases
we stub at a higher level for the compose path (this is a good thing).
Extracting out transmit.js allows us to lock down 100% coverage on that
file.
In this commit we add support for some tags which are also called
void-elements according to
http://w3c.github.io/html/syntax.html#void-elements to be parsed by
our template parser and get tagged as singleton_html_tags.
Fixes: #8387.
Now executable! Just run `tools/tagmessages`.
Also, get the username and password from a `.transifexrc` file.
And hardcode the project slug to `zulip-test` rather than to `zulip`;
the Transifex API is bad at namespacing, so this makes it possible to
run this script on a test project (the only way we're currently using
it) even for people like me who can also upload to the real Zulip
project on Transifex.
We now have a separate page for common error payloads, for example,
the payload for when the client's API key is invalid. All error
payloads that are presented on this page will be tested similarly
to our other non-error sample fixtures.
Otherwise prepare-base is likely to fail when first run (but then
succeed when rerun, because the container is left running), because
the container isn't up yet when we try to operate in it.
Also clean up the placement of `set -e` vs `set -x`.
Apparently, we've now had the first time one of our contributors had
their account deleted (at least, the author page for the contributor
who has 21 commits in python-zulip-api now 404s).
See 625939 for more information. In short, the purpose of this delay is
to give autoreload code enough time to touch every watched file at least
once before the change is made.
This script and our CI scripts tools/travis/{backend,frontend} have
stayed pretty well in sync in the 6 months since 360c27ded made that
relationship explicit and easy to check!
Just one small exception; so fix that.
This may or may not be temporary, but either way, the other code is
there in source control, and the "why" of disabling gitlint is the
helpful bit for a comment.
Injecting the generated-file warning into the settings dict felt a
little unnecessarily magical. A warning like this is always going
to be at the top; the way it might differ between files is mainly
if the syntax for a comment varies, and in that case a simple
substitution like we're doing in this template wouldn't be enough
to express the difference anyway. So, embrace the hardcoding.
Now, the template and the images.yml entry have a very simple
relationship: the keys in one are exactly the keys in the other.
That's good for people quickly and confidently understanding it.
Now that the Markdown extension defined in
zerver/lib/bugdown/api_generate_examples depended on code in the
tools/lib/* directory, it caused the production tests to fail since
the tools/ directory wouldn't exist in a production environment.
This commit uses the Markdown extension defined in
zerver/lib/bugdown/api_generate_example to generate the example
fixture and code example, so that both are tested in
tools/lib/api_tests.
We don't disable code by commenting it out -- that leaves a mess.
We delete it. Remembering what the code was is what source control
is for.
This fixes cd849bc3f "test-run-dev: Disable Nagios check."
from a few weeks ago.
For now, this does nothing in a production environment, but it should
simplify the process of doing testing on the Thumbor implementation,
by integrating a lot of dependency management logic.
This commit also adds a tool to push translation sources to Transifex.
This tool makes sure that we don't push mobile source file. Mobile
source file is supposed to be handled from Zulip-Mobile repo.
(This is a small fixup to the main change, which was accidentally
included in a previous commit:
08bbd7e61 "settings: Slightly simplify EMAIL_BACKEND logic."
Oops. See there for most of the changes described here.)
The installer works out of a release-tarball tree. We typically want
to share this tree between successive test-install runs (with an rsync
or similar command to update source files of interest) because
rebuilding a release tree from scratch is slow. But the installer
will munge the tree; so instead of directly bind-mounting the tree
into the container, we need to give it an overlay over the tree, as a
sandbox to play in.
Previously we used lxc-copy's `-m overlay=...` feature to do this,
mounting an overlay in the container. But then sometimes in
development we want to reach in and edit some code in the tree,
e.g. before rerunning the installer after something failed. Reaching
inside the container for this is a pain (`ssh` would add latency, and
I haven't installed sshd in the containers; and getting rsync to work
with `lxc-attach` was beyond what I could figure out in a few minutes
of fiddling); and editing the base tree often doesn't work.
So, create the overlay with our own `mount -t overlay`, and have
`lxc-copy` just bind-mount that in. Now the host has direct access to
the same overlay which the guest is working from.
Also this makes it past time to help the user out in finding the fresh
names we've created: first the container, now this shared tree. Print
those at the end, rather than make the user scroll to the top and find
the right `set -x` line to copy-paste from.
This is the last commit in the series of commits for completing the
project of cleaning up our html templates to have 4 space and
valid indentation.
Fixes: #1236.
In this commit we also fix a test which would fail as a result of
doing this cleanup since the test wasn't designed to take into
account the space chars which might occur in the beginning of a
html line.
This test randomly fails far too often in Travis -- I think more than
all our other tests combined. It needs to be fixed before we can ask
everyone to look at build failures it causes.
Now that we have `eval_rst` and can explicitly exclude pages from the
toctree completely, we no longer need to set `includehidden`, and we
can return to using upstream's template.
(Meanwhile, our feature request upstream was successful! See
rtfd/sphinx_rtd_theme#485, which upstream implemented just a week
after we requested it. So that would have been another option.)
This reverts commit 11b8b8f48 "docs: Add rtd layout template."
It runs in kind of a peculiar environment -- in particular with the
`tags` identifier injected into the namespace -- and it contains
very little code more complex than `foo = "bar"`, so there's not
much to check anyway.
Before this fix, the installer has an extremely annoying bug where
when run inside a container with `lxc-attach`, when the installer
finishes, the `lxc-attach` just hangs and doesn't respond even to
C-c or C-z. The only way to get the terminal back is to root around
from some other terminal to find the PID and kill it; then run
something like `stty sane` to fix the messed-up terminal settings
left behind.
After bisecting pieces of the install script to locate which step
was causing the issue, it comes down to the `service camo restart`.
The comment here indicates that we knew about an annoying bug here
years ago, and just swept it under the rug by skipping this step
when in Travis. >_<
The issue can be reproduced by running simply `service camo restart`
under `lxc-attach` instead of the installer; or `service camo start`,
following a `service camo stop`. If `lxc-attach` is used to get an
interactive shell, these commands appear to work fine; but then when
that shell exits, the same hang appears. So, when we start camo
we're evidently leaving some kind of mess that entangles the daemon
with our shell.
Looking at the camo initscript where it starts the daemon, there's
not much code, and one flag jumps out as suspicious:
start-stop-daemon --start --quiet --pidfile $PIDFILE -bm \
--exec $DAEMON --no-close -c nobody --test > /dev/null 2>&1 \
|| return 1
start-stop-daemon --start --quiet --pidfile $PIDFILE -bm \
--no-close -c nobody --exec $DAEMON -- \
$DAEMON_ARGS >> /var/log/camo/camo.log 2>&1 \
|| return 2
What does `--no-close` do?
-C, --no-close
Do not close any file descriptor when forcing the daemon
into the background (since version 1.16.5). Used for
debugging purposes to see the process output, or to
redirect file descriptors to log the process output.
And in fact, looking in /proc/PID/fd while a hang is happening finds
that fd 0 on the camo daemon process, aka stdin, is connected to our
terminal.
So, stop that by denying the initscript our stdin in the first place.
This fixes the problem.
The Debian maintainer turns out to be "Zulip Debian Packaging Team",
at debian@zulip.com; so this package and its bugs are basically ours.
This is a tool that throws away `fsync` calls and other requests for
the system to sync files to disk. It may make the install faster; for
example, if it has to install a number of system packages, `dpkg` is
known to make a lot of `fsync` calls which slow things down
significantly. Conversely, if there's a power failure in the middle
of running a test install, we really don't mind if the test install's
data becomes corrupt.
When the install script is successful, one of the final things it
wants to do is to move the tree that Zulip was installed from into the
deployments directory. It can't do that, at least not in a naive way
with `mv`, if the tree is actually a mount point. So, stick the tree
inside some other directory that we create just for the purpose of
being the mount point and containing the install tree.
This saves several minutes off the install time. Sadly pip still
clones Git repos for dependencies that point to them, but for many
others (not all? not sure) it just gets a wheel from the cache.
The third-party `install-yarn.sh` script uses `curl`, and we invoke it
in `install-node`. So we need to install it as a dependency.
We've mostly gotten away with this because it's common for `curl` to
already be installed; but it isn't always.
This greatly simplifies iterating on changes to the installer and
associated code: just edit in the shared directory (or edit in your
worktree and rsync to the directory), and rerun.
With this change, the form with a directory is now really the main
way to run the script; the form accepting a tarball is really just
a convenience feature, unpacking the tarball and then proceeding with
that directory.
This will facilitate testing interesting installer features
using its own CLI.
On my laptop, with a recent base image (updated a few days ago with
`prepare-base`), it takes just 7 or 8 seconds to get to the installer
running, as timed by passing `--help` so that the installer promptly
exits.
In order to do development on the installer itself in a sane way,
we need a reasonably fast and automatic way to get a fresh environment
to try to run it in.
This calls for some form of virtualization. Choices include
* A public cloud, like EC2 or Digital Ocean. These could work, if we
wrote some suitable scripts against their APIs, to manage
appropriate base images (as AMIs or snapshots respectively) and to
start fresh instances/droplets from a base image. There'd be some
latency on starting a new VM, and this would also require the user
to have an account on the relevant cloud with API access to create
images and VMs.
* A local whole-machine VM system (hypervisor) like VirtualBox or
VMware, perhaps managing the configuration through Vagrant. These
hypervisors can be unstable and painfully slow. They're often the
only way to get development work done on a Mac or Windows machine,
which is why we use them there for the normal Zulip development
environment; but I don't really want to find out how their
instability scales when constantly spawning fresh VMs from an image.
* Containers. The new hotness, the name on everyone's lips, is Docker.
But Docker is not designed for virtualizing a traditional Unix server,
complete with its own init system and a fleet of processes with a
shared filesystem -- in other words, the platform Zulip's installer
and deployment system are for. Docker brings its own quite
different model of deployment, and someday we may port Zulip from
the traditional Unix server to the Docker-style deployment model,
but for testing our traditional-Unix-server deployment we need a
(virtualized) traditional Unix server.
* Containers, with LXC. LXC provides containers that function as
traditional Unix servers; because of the magic of containers, the
overhead is quite low, and LXC offers handy snapshotting features
so that we can quickly start up a fresh environment from a base
image. Running LXC does require a Linux base system. For
contributors whose local development machine isn't already Linux,
the same solutions are available as for our normal development
environment: the base system for running LXC could be e.g. a
Vagrant-managed VirtualBox VM, or a machine in a public cloud.
This commit adds a first version of such a thing, using LXC to manage
a base image plus a fresh container for each test run. The test
containers function as VMs: once installed, all the Zulip services run
normally in them and can be managed in the normal production ways.
This initial version has a shortage of usage messages or docs, and
likely has some sharp edges. It also requires familiarity with the
basics of LXC commands in order to make good use of the resulting
containers: `lxc-ls -f`, `lxc-attach`, `lxc-stop`, and `lxc-start`,
in particular.
Stripe Checkout means using JS code provided by Stripe to handle
almost all of the UI, which is great for us.
There are more features we should add to this page and changes we
should make, but this gives us an MVP.
[greg: expanded commit message; fixed import ordering and some types.]
This is fairly often -- though not always! -- failing, with a nasty
failure mode where it takes like 6 minutes to time out. See
discussion on #7748 (search for "bad link").
Actually, after seeing it happen just now when running
test-documentation on my laptop, on some other link, it occurs to me
that I've seen this before -- it's fairly common in Travis, too. It's
just that it doesn't actually cause the build to fail :-/, and on
Travis we haven't been paying as close attention to slow builds as we
are on Circle right now.
Generally stderr is the conventional place for this sort of running
commentary, and it's better set up for it: by default stdout may have
a buffer inside the process so that things written to it don't reach
the outside until later, while stderr is always by default unbuffered,
so messages are printed immediately.
Here, until the previous commit, because our color-reset sequence was
being printed without a following newline (with `echo -n`), it was
getting buffered; and then error messages from `scrapy` to stderr were
being erroneously painted with the color intended for the message
"Testing links in documentation...".
The autoreload code of Django works by looping over the files associated
with all the loaded modules. This loop is run after every 1 second. If
the file is found for the first time by the loop, it is assumed that the
file is new and is not modified between the time it is loaded and is
checked by the loop. This assumption is the source of a race condition.
We can either implement a more sensitive version of the loop or we can
just allow enough time to the Django loop to touch every file at least
once.
For the time being, we are going with the second option.
Previously, there were following problems with the implmentation:
* Same file handle was being used to read and write. We used to do
`seek(0)` and then `read()`. This had a chance to overwrite
file data. Now we use different file handles to read and write data.
* We were using text streams. Text streams cannot be used with
`bufferring=0`. Now we use binary streams without buffering so that
data is available for reading without any delay.
This commit also updates the key(s) that we search in the logfile.
Previously, launch of all queues was announced in the log, now we only
anounce the number of threads that were launched.
This commit also makes sure that we always exit after gracefull shutting
down the development server.
Previously, there were following problems with the implmentation:
* Same file handle was being used to read and write. We used to do
`seek(0)` and then `read()`. This had a chance to overwrite file
data. Now we use different file handles to read and write data.
* We were using text streams. Text streams cannot be used with
`bufferring=0`. Now we use binary streams without buffering so that
data is available for reading without any delay.
This commit just copies all the code from MissedMessageSendingWorker
class to a new EmailSendingWorker class. All the logic to send an email
through a queue was already there. This commit only makes the logic
generic. It does so by creating a special purpose queue called
'email_senders' to send any type of email. To make
MissedMessageSendingWorker still work we derive it from
EmailSendingWorker. All the tests that were testing
MissedMessageSendingWorker now run against EmailSendingWorker.
We get the following error (edited slightly):
Dec 19 06:13:27 commit_messages| An error occurred while executing
'/usr/bin/git rev-list --max-count=-1 upstream/master..HEAD':
fatal: ambiguous argument 'upstream/master..HEAD':
unknown revision or path not in the working tree.
Use '--' to separate paths from revisions, like this:
'git <command> [<revision>...] -- [<file>...]'
We'll need to adjust the remotes to make `upstream` mean what we expect.
This isn't really the right way to do this -- commit
dcd80e665 "travis/setup-backend: Remove the '--travis' flag"
took us in the wrong direction by introducing more magic, deeper in
the stack.
But it's the same way we do it for Travis. For now, just copy that.
[Thanks to hackerkid for cleaning up my original crude hack.]
At this point if we were accidentally using `/srv/zulip-venv` for
anything, we'd have run into it by now. So just drop the bit of
historical logic that we had to ensure that.
This reverts commit 66261f1cc. See parent commit for reason; here,
provision worked but `tools/run-dev.py` would give errors.
We need to figure out a test that reproduces these issues, then make a
version of these changes that keeps that test working, before we
re-merge them.
We should omit these for mypy. For most class definitions,
mypy doesn't need `Any`, and it provides no real useful info.
For clever monkeypatches, you should provide a more specific
type than `Any`.
These commands are super boring standard Docker commands,
so this probably isn't helpful for anyone who is familiar
with building Docker images... but I had to consult docs
to work out the right commands again today, so they'd help me.
Provision was failing at tools/setup/postgres-init-dev-db
with this in the log:
Sep 20 02:27:01 + sudo -i -u postgres psql ''
[sudo] password for circleci:
The issue is that the old version of this line (from Circle upstream)
only lets the `circleci` user sudo to root -- not to other users, or
not directly anyway -- because sudoers syntax is complicated. Fix it,
after studying `man sudoers`.
This is only an optimization -- if this list is missing anything,
we'll get to it in `provision` in the actual build. That's important,
because we want an existing image to work fine for testing new
versions of our codebase, including changes that may install more
packages in `provision`.
What this does accomplish is keeping provision's `apt-get install`
fast, by leaving it very little work to do.
The list comes from looking at the APT output during provision in an
actual run without this step, and leaving out two packages which
aren't available at this stage, because we get them from PPAs:
postgresql-9.3-pgroonga and postgresql-9.3-tsearch-extras.
Install `jq` with APT -- that's a lot simpler to read than this
explicit download.
And coalesce several commands, following Docker upstream's
recommendation and avoiding unnecessary overhead.
This is nearly the same as Circle's version, linked in the comment.
I've
* changed the FROM line to get Ubuntu,
* added a couple of distro packages to compensate, and
* revised the comments.
This seems to have been causing the travis production suite to fail.
It's a direct consequence of removing travis' giant library of apt
sources.list files; now that those are gone, there aren't copies of
all these extra packages available anyway.
This method was new in Tornado 4.0. It saves us from having to get
the time ourselves and do the arithmetic -- which not only makes the
code a bit shorter, but also easier to get right. Tornado docs (see
http://www.tornadoweb.org/en/stable/ioloop.html) say we should have
been getting the time from `ioloop.time()` rather than hardcoding
`time.time()`, because the loop could e.g. be running on the
`time.monotonic()` clock.
tools/setup/generate_zulip_bots_static_files now starts off by
deleting static/generated/bots/ (if it hasn't been removed already)
so that outdated static files from older versions of the zulip_bots
package don't supress errors in the main repo that would otherwise
break.
For more info, see #7542Fixes: #7542.
The `re.match` function in the Python stdlib is a trap for the unwary,
with surprising and asymmetrical semantics; we should probably add a
lint rule to ban it entirely. The docstring says:
> Try to apply the pattern at the start of the string, [...]
In other words, it effectively adds a `^` at the start (or `\A`, where
the distinction matters.) It's bad enough that this differs from what
grep, sed, perl, less, and every other tool I can think of do when
looking for matches to a regex; on top of that, it treats the
beginning of the string differently from the end, for no obvious
reason. The function that does what the rest of the world understands
by "match against this regex" is `re.search`.
In this case, it's unlikely that anyone intended for comments with
URLs, or `api_url` references, to miss out on their respective
exceptions to the long-line rule if they happen to start after the
first column. So fix those rules by just switching to `re.search`
with the same pattern.
I think Markdown URL references may have to start at the beginning of
the line, so I've left a `^` there to preserve -- but now make
explicit -- the `re.match` behavior.
Tweaked by tabbott to move changes from the next commit that are
required for this to pass tests into this commit.
Note that this exports a few items that were not previously exported.
This is part of our efforts to change our integrations/webhooks
docs to follow the same sort of numbered-list format as our /help
docs. In order to indicate that paragraphs separated by newlines
are part of the same numbered-list point, every paragraph must be
indented 4 spaces.
This both improves the comment to be more readable, and also uses the
new and improved exclude feature to limit the exclusion to just the
webhook fixtures (where it's needed).
Also fixes a mypy error.
The previous exclude rules only allowed excluding a directory (and
things in subdirectories would silently still be linted). Anyone
using this would expect it to exclude a directory tree, so we make it
do that.
The readthedocs theme overrides a few settings in their layout template.
We might want to change some settings back to their default values.
This commit copies the original readthedocs layout file from
https://github.com/rtfd/sphinx_rtd_theme/blob/master/sphinx_rtd_theme/layout.html
to _templates/layout.html, and excludes it from lint and template checks.
Addresses #7417.
This fixes a bug where, when a user is unsubscribed from a stream,
they might have unread messages on that stream leak. While it might
seem to be a minor problem, it can cause significant problems for
computing the `unread_msgs` data structures, since it means we need to
add an extra filter for whether the user is still subscribed, either
in the backend or in the UI.
Fixes#7095.
This commit modifies `test-locked-requirements` to use some caching
so that we don't need to use the `update-locked-requirements` tool
everytime for checking the validity of locked requirements as it is
slow.
Fixes: #6969.
This commit renames various source requirements files like `dev.txt`,
`mypy.txt` etc to `dev.in`, `mypy.in` etc and various locked requirements
files like `dev_lock.txt`, `mypy_lock.txt` etc to `dev.txt`, `mypy.txt`
etc. This will help in emphasizing to the user that *.in are actually
input to `update-locked-requirements` tool which should be run after
updating any of these.
This often can cause minor caching problems.
Obviously, it'd be better if we had access to the AST and thus could
do this rule for UserProfile objects in general.
This prevents the caches in /srv from growing to fill up the disk --
e.g., on my laptop after 6 months of regular development the venv cache
was 12G and the NPM cache 5G, making them by far the largest disk hogs
on the machine.
It costs about 0.4s, apart from any time spent actually removing
things. This is a little annoyingly slow to be adding to every
provision, and seems like it could be optimized, but I think already
worth it as is.
We already do this by default in tools/build-docs, but since we
migrated test-documentation to not run that directly (to disable
collapsing), we need to add the recent parallelism fix here too.
It saves about 5-10s when running this test suite for me, which is
good, but definitely leaves me feeling like there could be more
improvement.
In this commit we add new dependencies needed for running thumbor.
Also we add the script for creating the virtual environment ready
for thumbor.
Note: Thumbor will use python2 and thus have different virtualenv
dedicated to it.
Credits to @TigorC and @joshland as well for there work on this.
In this commit we add a new option which could be used to specify
python version. When 'py2' is specified, future/futures are not
removed from the requirements lock file generated.
Even where this is actually used for a temporary checkout, it obscures
the relationship between this and $TMPDIR -- and some of our logic
depends on that. In other places, it isn't actually even a checkout.
In all cases, the expanded version is clearer.
This script, and tools/update-prod-static which it relies on,
have kept getting more complex since this conditional was added
in 2013, and the places we rely on GNU features have probably
multiplied beyond `mktemp -d`. It's unlikely this works on
macOS with BSD tools now, and it'd be hard to maintain that way
if it did; drop the pretense.
There's no need to remove this file here -- the whole tree will be
removed a few commands later, and the `tar` command we do first, to
supplement our tarball with various generated files, is quite
selective and wouldn't look at this file anyway.
With the new portico work we've done, the help documentation does
sorta depend on the database if you're logged in. So it's best to
just require it for these tests.
This commit helps reduce clutter on the navigation sidebar.
Creates new directories and moves relevant files into them.
Modifies index.rst, symlinks, and image paths accordingly.
This commit also enables expandable/collapsible navigation items,
renames files in docs/development and docs/production,
modifies /tools/test-documentation so that it overrides a theme setting,
Also updates links to other docs, file paths in the codebase that point
to developer documents, and files that should be excluded from lint tests.
Note that this commit does not update direct links to
zulip.readthedocs.io in the codebase; those will be resolved in an
upcoming follow-up commit (it'll be easier to verify all the links
once this is merged and ReadTheDocs is updated).
Fixes#5265.
The CSS linter was pretty hard to reason about. It was
pretty flexible about certain things, but then it would
prevent seemingly innocuous code from getting checked in.
This commit overhauls the pretty-printer to be more composable,
where every object in the AST knows how to render itself. It
also cleans up a little bit of the pre_fluff/post_fluff logic
in the parser itself, so comments are more likely to be "attached"
to the AST node that make sense.
The linter is actually a bit more finicky about newlines, but
this is mostly a good thing, as most of the variations before
this commit were pretty arbitrary.
This adds the "--disallow-any=generics" option to run-mypy, which no
longer permits:
- inheriting from "list"; use "List[sometype]" (or a TypeVar)
- generic types with no following square brackets specifying the type
(even if initially 'Any')
Any (and '...' for Callable) is a lot easier to search for than an
absence of square brackets, and should improve overall typing quality.
Armaan Ahluwalia
Shubham Dhama
Tim Abbott
Anders Kaseorg
Yashashvi Dave
Vishnu Ks
Shubham Padia
Greg Price
Raymond Akornor
Aditya Bansal
Sampriti Panda
Robert Hönig
Yago González
Steve Howell
Sivakar Sithamparanathan
Eeshan Garg
Michael J. Sullivan
Max Nussenbaum
Aman Ghebreweldi
Joshua Pan
RobbieClarken
Darshan Markandaiah
Priyank Patel
Emilio Schadt
Preston Hansen
Alyssa Wagenmaker
Tarun Kumar
Ben Reeves
neiljp (Neil Pilgrim)
sinwar
Gooca
Harshit Bansal
Umair Khan
Balaji2198
rht
ihsavru
Rhea Parekh
Brock Whittaker
Rohitt Vashishtha
Cynthia Lin
cPhost
Tommy Ip
Symt
picapi_
David Rosa Tamsen