Code style and conventions

Be consistent!

Look at the surrounding code, or a similar part of the project, and try to do the same thing. If you think the other code has actively bad style, fix it (in a separate commit).

When in doubt, send an email to with your question.

Lint tools

You can run them all at once with


You can set this up as a local Git commit hook with


The Vagrant setup process runs this for you.

lint runs many lint checks in parallel, including

  • JavaScript (ESLint)
  • Python (Pyflakes)
  • templates
  • Puppet configuration
  • custom checks (e.g. trailing whitespace and spaces-not-tabs)


Please don’t put any passwords, secret access keys, etc. inline in the code. Instead, use the get_secret function in zproject/ to read secrets from /etc/zulip/secrets.conf.

Dangerous constructs

Misuse of database queries

Look out for Django code like this:

for bar in Bar.objects.filter(...)
if  bar.baz < 7]

This will make one database query for each Bar, which is slow in production (but not in local testing!). Instead of a list comprehension, write a single query using Django’s QuerySet API.

If you can’t rewrite it as a single query, that’s a sign that something is wrong with the database schema. So don’t defer this optimization when performing schema changes, or else you may later find that it’s impossible.

UserProfile.objects.get() / Client.objects.get / etc.

In our Django code, never do direct UserProfile.objects.get(email=foo) database queries. Instead always use get_user_profile_by_{email,id}. There are 3 reasons for this:

  1. It’s guaranteed to correctly do a case-inexact lookup
  2. It fetches the user object from remote cache, which is faster
  3. It always fetches a UserProfile object which has been queried using .select_related(), and thus will perform well when one later accesses related models like the Realm.

Similarly we have get_client and get_stream functions to fetch those commonly accessed objects via remote cache.

Using Django model objects as keys in sets/dicts

Don’t use Django model objects as keys in sets/dictionaries – you will get unexpected behavior when dealing with objects obtained from different database queries:

For example, UserProfile.objects.only("id").get(id=17) in set([UserProfile.objects.get(id=17)]) is False

You should work with the IDs instead.

You should always pass the update_fields keyword argument to .save() when modifying an existing Django model object. By default, .save() will overwrite every value in the column, which results in lots of race conditions where unrelated changes made by one thread can be accidentally overwritten by another thread that fetched its UserProfile object before the first thread wrote out its change.

Using raw saves to update important model objects

In most cases, we already have a function in zephyr/lib/ with a name like do_activate_user that will correctly handle lookups, caching, and notifying running browsers via the event system about your change. So please check whether such a function exists before writing new code to modify a model object, since your new code has a good chance of getting at least one of these things wrong.

Naive datetime objects

Python allows datetime objects to not have an associated timezone, which can cause time-related bugs that are hard to catch with a test suite, or bugs that only show up during daylight savings time.

Good ways to make timezone-aware datetimes are below. We import timezone function as from django.utils.timezone import now as timezone_now and from django.utils.timezone import utc as timezone_utc. When Django is not available, timezone_utc should be replaced with pytz.utc below.

  • timezone_now() when Django is available, such as in zerver/.
  • when Django is not available, such as for bots and scripts.
  • datetime.fromtimestamp(timestamp, tz=timezone_utc) if creating a datetime from a timestamp. This is also available as zerver.lib.timestamp.timestamp_to_datetime.
  • datetime.strptime(date_string, format).replace(tzinfo=timezone_utc) if creating a datetime from a formatted string that is in UTC.

Idioms that result in timezone-naive datetimes, and should be avoided, are and datetime.fromtimestamp(timestamp) without a tz parameter, datetime.utcnow() and datetime.utcfromtimestamp(), and datetime.strptime(date_string, format) without replacing the tzinfo at the end.

Additional notes:

  • Especially in scripts and puppet configuration where Django is not available, using time.time() to get timestamps can be cleaner than dealing with datetimes.
  • All datetimes on the backend should be in UTC, unless there is a good reason to do otherwise.

x.attr('zid') vs.

Our message row DOM elements have a custom attribute zid which contains the numerical message ID. Don’t access this directly as x.attr('zid') ! The result will be a string and comparisons (e.g. with <=) will give the wrong result, occasionally, just enough to make a bug that’s impossible to track down.

You should instead use the id function from the rows module, as in This returns a number. Even in cases where you do want a string, use the id function, as it will simplify future code changes. In most contexts in JavaScript where a string is needed, you can pass a number without any explicit conversion.

JavaScript var

Always declare JavaScript variables using var:

var x = ...;

In a function, var is necessary or else x will be a global variable. For variables declared at global scope, this has no effect, but we do it for consistency.

JavaScript has function scope only, not block scope. This means that a var declaration inside a for or if acts the same as a var declaration at the beginning of the surrounding function. To avoid confusion, declare all variables at the top of a function.

JavaScript for (i in myArray)

Don’t use it: [1], [2], [3]

jQuery global state

Don’t mess with jQuery global state once the app has loaded. Code like this is very dangerous:

$.ajaxSetup({ async: false });
$.ajaxSetup({ async: true });

jQuery and the browser are free to run other code while the request is pending, which could perform other Ajax requests with the altered settings.

Instead, switch to the more general $.ajax_ function, which can take options like async.

Translation tags

Remember to tag all user-facing strings for translation, whether they are in HTML templates or JavaScript editing the HTML (e.g. error messages).

State and logs files

Do not write state and logs files inside the current working directory in the production environment. This will not do what you expect, because the current working directory for the app changes every time we do a deploy. Instead, hardcode a path in – see SERVER_LOG_PATH in for an example.

JS array/object manipulation

For generic functions that operate on arrays or JavaScript objects, you should generally use Underscore. We used to use jQuery’s utility functions, but the Underscore equivalents are more consistent, better-behaved and offer more choices.

A quick conversion table:

   $.each → _.each (parameters to the callback reversed)
   $.inArray → _.indexOf (parameters reversed)
   $.grep → _.filter
   $.map →
   $.extend → _.extend

There’s a subtle difference in the case of _.extend; it will replace attributes with undefined, whereas jQuery won’t:

   $.extend({foo: 2}, {foo: undefined});  // yields {foo: 2}, BUT...
   _.extend({foo: 2}, {foo: undefined});  // yields {foo: undefined}!

Also, _.each does not let you break out of the iteration early by returning false, the way jQuery’s version does. If you’re doing this, you probably want _.find, _.every, or _.any, rather than ‘each’.

Some Underscore functions have multiple names. You should always use the canonical name (given in large print in the Underscore documentation), with the exception of _.any, which we prefer over the less clear ‘some’.

More arbitrary style things


Indentation is four space characters for Python, JS, CSS, and shell scripts. Indentation is two space characters for HTML templates.

We never use tabs anywhere in source code we write, but we have some third-party files which contain tabs.

Keep third-party static files under the directory zephyr/static/third/, with one subdirectory per third-party project.

We don’t have an absolute hard limit on line length, but we should avoid extremely long lines. A general guideline is: refactor stuff to get it under 85 characters, unless that makes the code a lot uglier, in which case it’s fine to go up to 120 or so.

Whitespace guidelines:

  • Put one space (or more for alignment) around binary arithmetic and equality operators.
  • Put one space around each part of the ternary operator.
  • Put one space between keywords like if and while and their associated open paren.
  • Put one space between the closing paren for if and while-like constructs and the opening curly brace. Put the curly brace on the same line unless doing otherwise improves readability.
  • Put no space before or after the open paren for function calls and no space before the close paren for function calls.
  • For the comma operator and colon operator in languages where it is used for inline dictionaries, put no space before it and at least one space after. Only use more than one space for alignment.


Don’t use == and != because these operators perform type coercions, which can mask bugs. Always use === and !==.

End every statement with a semicolon.

if statements with no braces are allowed, if the body is simple and its extent is abundantly clear from context and formatting.

Anonymous functions should have spaces before and after the argument list:

var x = function (foo, bar) { // ...

When calling a function with an anonymous function as an argument, use this style:

$.get('foo', function (data) {
    var x = ...;
    // ...

The inner function body is indented one level from the outer function call. The closing brace for the inner function and the closing parenthesis for the outer call are together on the same line. This style isn’t necessarily appropriate for calls with multiple anonymous functions or other arguments following them.


$(function () { ...

rather than

$(document).ready(function () { ...

and combine adjacent on-ready functions, if they are logically related.

The best way to build complicated DOM elements is a Mustache template like zephyr/static/templates/message.handlebars. For simpler things you can use jQuery DOM building APIs like so:

var new_tr = $('<tr />').attr('id',;

Passing a HTML string to jQuery is fine for simple hardcoded things:

foo.append('<p id="selected">foo</p>');

but avoid programmatically building complicated strings.

We used to favor attaching behaviors in templates like so:

<p onclick="select_zephyr({{id}})">

but there are some reasons to prefer attaching events using jQuery code:

  • Potential huge performance gains by using delegated events where possible
  • When calling a function from an onclick attribute, this is not bound to the element like you might think
  • jQuery does event normalization

Either way, avoid complicated JavaScript code inside HTML attributes; call a helper function instead.


Don’t use the style= attribute. Instead, define logical classes and put your styles in external files such as zulip.css.

Don’t use the tag name in a selector unless you have to. In other words, use .foo instead of We shouldn’t have to care if the tag type changes in the future.

Don’t use inline event handlers (onclick=, etc. attributes). Instead, attach a jQuery event handler ($('#foo').on('click', function () {...})) when the DOM is ready (inside a $(function () {...}) block).

Use this format when you have the same block applying to multiple CSS styles (separate lines for each selector):

selector2 {


  • Scripts should start with #!/usr/bin/env python and not #/usr/bin/python (the right Python may not be installed in /usr/bin) or #/usr/bin/env python2.7 (bad for Python 3 compatibility). Don’t put a shebang line on a Python file unless it’s meaningful to run it as a script. (Some libraries can also be run as scripts, e.g. to run a test suite.)

  • Scripts should be executed directly (./, so that the interpreter is implicitly found from the shebang line, rather than explicitly overridden (python

  • The first import in a file should be from __future__ import absolute_import, per PEP 328

  • Put all imports together at the top of the file, absent a compelling reason to do otherwise.

  • Unpacking sequences doesn’t require list brackets:

    [x, y] = xs    # unnecessary
    x, y = xs      # better
  • For string formatting, use x % (y,) rather than x % y, to avoid ambiguity if y happens to be a tuple.

  • When selecting by id, don’t use when you mean E.g.

    recipient = Recipient(, type=Recipient.HUDDLE)

    should be written as

    recipient = Recipient(, type=Recipient.HUDDLE)

    in case we ever change the primary keys.


All significant new features should come with tests. See testing.

Third party code

When adding new third-party packages to our codebase, please include “[third]” at the beginning of the commit message. You don’t necessarily need to do this when patching third-party code that’s already in tree.