Since we are doing a assert call checking a url without no page.waitForSelector
or similar calls, we add this to avoid flakes where the assert is called
when we are in process of navigation to the page. Also, we use Promise.all
here, and call page.waitForNavigation first, to avoid race conditions where
form is submitted quickly and navigation occurs and then we are stuck waiting
for a navigation. There were no flakes in the CI related to this issue, but
I expect someone with a slow hardware would probably stumble upto this in future.
We were missing the step for clicking the checkbox to show the subdomain
input field. There was a flake because of this issue in CI where the
"testsubdomain" input was getting typed into the name field. This fixes
that flake.
There were two reason there were flakes in login tests. One problem was
that the assert call for checking that the login page url was happening
too early in rare cases. To fix this issue we wait for email input field
element in the login page. The second issue was the sometimes the url was
/login/ instead of /accounts/login/. The time between the redirect were
super close here too (around ~10ms) so I am suprised this didn't occur flake
locally more often.
Server logs that show the redirect:
2020-05-26 18:55:27.065 INFO [zr] 127.0.0.1 GET 200 36ms (db: 5ms/2q) (+start: 14ms) /accounts/login/ (unauth@zulip via ?)
2020-05-26 18:55:27.074 INFO [zr] 127.0.0.1 GET 200 28ms (db: 3ms/2q) (+start: 23ms) /login/ (unauth@zulip via ?)
To take a screenshot on failure where we have our common error
handling code in common.run_test method we need to have access to
the Page instance. Currently, we create a new Page in the test method
we pass into run_test, so we pass in a new Page instance to that method
so the test and the run_test method have access to the same Page. And,
then we take a screenshot on failure. It will be saved as `failure-${num}`
in var/puppeteer directory.
This is mostly just moving code from the `activity.js`
tests, but I also now explicitly cover the "100%"
use case (i.e. all four folks in the huddle are present).
Computed indexes into these raw objects should be guarded with
Object.prototype.hasOwnProperty; make our accessors do this
automatically and use them consistently.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
This is helpful because if the user pastes multiple queries in the
searchbox and there are invalid search operators, then it is visible
through the typeahead.
The main reasoning for this change is as follows:
* When the search bar contains multiple search queries
but no search results, the last search operand does
not get displayed.
This happens due to the fact that filter object
contained 2 terms having the operator key value as
"search" instead of a single term where operator is
"search" and operand is a single string containing
the space seperated search queries. This condition
occurs for search_pills_enabled case only because
we used to Filter.parse the query twice
(once for the `base_operators` and once for the
`suggestion_operator instead of doing both at once).
Thus the `search_query` value inside the
`narrow.show_search_query` function which only
selected the operands of the first term displayed
an incomplete result.
* Another benefit of this commit is to display the narrow
operators in the URL fragment the same way as when
search_pills_enabled = False.
For example, On entering the queries in the mentioned
order -> 'is: starred', 'abc', 'def', 'is: private',
'ghi'. This is the URL:
Previously:
/#narrow/is/starred/is/private/search/abc.20def/search/ghi
Now (same as pills disabled case):
/#narrow/is/starred/is/private/search/abc.20def.20ghi
* We are also able to de-duplicate the non-typeahead search
query code path.
Commit 9b78a73e36 (#15005) made some of
our poorly written Casper tests fail. Now they’re just as poorly
written but passing again.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
The people.js tests were using _add_user function to add
cross realm bots. The problem is that _add_user function
doesn't properly simulates the adding process as it doesn't
add the user in cross_realm_dict as well.
To solve this and eliminate the need of calling
people.initialize(), which means the params obj needs to be
defined, we extracted the whole logic of adding a cross realm
user into a separete function, add_cross_realm_user.
This makes it so that search_suggestion.js
does not depend on activity.js.
That dependency hasn't really been "elegant"
for quite some time, but it will become particularly
unnecessary when we go to remove the "Group PMs"
section from the right sidebar.
This commit introduces a temporary wart
where we have these two functions with the
same name in a sort of unnecessarily
complicated code stack:
activity.process_loaded_messages
huddle_data.process_loaded_messages
But we will eliminate the former function
very soon, and our message-related codepaths
will just call the `huddle_data` version
directly.
TESTING NOTES:
Now that `huddle_data` is a tiny leaf
module, it's super easy to just use the
real implementation of what was formerly
called `activity.get_huddles()` (and is
now in `huddle_data`).
When I first wrote this commit, introducing
the real implementation of `get_huddles` exposed
some bugs that I fixed in the immediately
prior commits to this.
When the tests were originally written,
I believe `activity.js` had some annoying
`jQuery` dependencies that made it hard
to unit test against. We've slimmed it over
time to be mostly just a "controller" module.
But even in its current state it would have
been a bit of a bloated dependency.
The other friction for using the actual
version of `get_huddles` was setting up
the message data, but that's pretty minor.
If you have a group PM where some users have
three-digit user_ids and some with four-digit
user_ids (or similar), a huddle could effectively
be ignored when determining the order of
search search suggestions.
Basically, we need a way to canonically sort
user_ids in "huddle" strings, and it's somewhat
arbitrary whether you sort lexically or sort
numerically, but you do need to be consistent
about it.
And JS is not exactly helpful here:
> [99, 101].sort()
[ 101, 99 ]
This is a pretty obscure bug with pretty low
user-facing consequences, and it was never
reported to us as far as I know, but the fix
here is pretty straightforward.
We have had similar bugs of slightly more consequence
in the past. The reason this bug has shown
up multiple times in our codebase is that every
component that deals with huddles has slightly
different forces that determine how it wants
to serialize the huddle. It's just one of those
annoying things. Plus, bugs with group PMs
do tend to escape detection, since most people
spend most of their time either on streams
or in 1:1 PMs.
This is a pure code extraction. The current
code is buggy with respect to user_ids with
different lengths of digits, i.e. it does
a naive lexical sort instead of a numerical
sort. We'll fix that in the next commit.
We already have a loading indicator for fetching older
messages. Thus it makes sense to implement the same
for displaying newer messages.
We set the display of `bottom-messages-logo` to none,
to prevent displaying two loading indicators during
the initial message load.
Fixes#15060.
I consolidate most of our users toward the top
of the file, so that we don't have to clutter
up individual tests. This also avoids some
confusion where charles/maria got repeated
in different tests with different ids.
I also introduce a couple four-digit ids to
try to expose more bugs related to sorting.
Note that it's still easy to keep tests
isolated here, as we have always been able
to cheaply re-initialize `people.js` and then
add individual users back.
There are still some tests where it makes
sense to just declare users locally, especially
if we are mutating their data.
There are a few minor incidental cleanups here,
mostly involving replacing hard coded ids
with things like `maria.user_id`.
This creates a little bit of noise in some
tests where we don't care about users, but
it's worth avoiding confusion about which
users exist at which time. Also the noisy
aspects here may actually catch regressions.
Finally, if the noise gets annoying, we can
do things like rename "Ted" not to collide
with the "Test" stream.
Using "bob" as the current user was a bad
choice, as our convention is to use "me" or
"myself" or "alice" for the current user.
It also particularly complicated the tests
around Group PMs.
Now we have both "bob" and "myself", which
makes the intentions of the tests a little
more clear.
This commit adds code to live update the message edit history.
Message edit history is fetched and rendered again if the edit
history modal is open.
This also adds 'data-message-id' attribute to 'message-history'
when opening history modal element which is used for checking
whether the history modal opened is of the message which is
edited.
Fixes#15051.
This code generates the timestamp string to be shown to the user
from the given timestamp in unix format using moment.js.
We also render the timestamp in a pill.
Previously, we handled this code only in message_list_view.js.
Now we support rendering stream descriptions and some dynamic
elements can be rendered in them, so we extract this new module
and use it in both the places.
This adds support for syntax like: !time(Jun 7 2017, 6:30 PM) so that
everyone sees the time in their own local timezone. This can be used
when scheduling online meetings, etc.
This adds some hardcoded values for timezones, because of there
being no sureshot way of determining the timezone easily. However,
since the main way of using the feature should be a typeahead for
entering the time, this shouldn't be cause of much concern.
Fixes#5176.
We wrap the [reset] anchor tag in a button so that we can set 'disabled'
attribute on it. We change the styles to hide the [reset] button and the
pencil icon when the widget is disabled.
We also need to call `e.preventDefault()` in the event handler since now
the anchor tag behaves as a button.
Add methods to extract recent topics from received messages.
Process new messages as they are received.
Use new messages received from the server to extract recent_topics.
Node tests added.
In common.js, we now have a single browser instance for the whole
test. When we update the test-js-with-puppetter to spawn a single
node process, like we do for node tests, we will save time on having
to open an new browser for every test + puppetter start up cost.
We will also add some more helpers here like a method for
filling out a form easily etc.
Before fixing `test-js-with-puppeteer` to to use the right webpack
setup, redirects weren't happening properly. The line this commit
removes did a redirect to the register page. This line is removed
as the redirect will happen automatically as expected with fixing
`test-js-with-puppeteer`.
Previously, we tried to read the value from page_params, which was just
a hack to make the calling code look cleaner. We now remove that hack
and thus, our dependency on page_params existing. Now, if the caller
does not specify a default value, we'll use the null-value.
This also creates a new init() function to cleanly wrap the code that
makes changes to the opts passed to the widget.
JSON.parse behaves as we want for numbers but for strings, we would
throw an error like 'unexpected token at position 0'. This meant we
couldn't read back the value set by `$input.data('val', 'text')`.
We had removed this function from the codebase when we switched to
using dropdown_list_widget. This was accidentally left as it is when
making that change.
Previously, we handled these updates in server_events_dispatch
and could accidentally call widget.render() before initializing
the widget.
Original report: https://chat.zulip.org/#narrow/near/875608.
The sync_realm_settings function ensures that if the settings are
not open, any updates are a noop.
When a user changes its avatar image, the user's avatar in popovers
wasn't being correctly updated, because of browser caching of the
avatar image. We added a version on the request to get the image in
the same format we use elsewhere, so the browser knows when to use the
cached image or to make a new request to the server.
Edited by Tim to preserve/fix sort orders in some tests, and update
zulip_feature_level.
Fixes: #14290
* Remove old topic and reprocess both old and new topic to ensure
that we are correctly storing the last_msg_id of users in the
topic. Also, Handle topic's stream (& topic) edit updates.
* Add function to get all messages in a topic in message_utils.js.
* Send topic edit event to recent_senders.
* Add func get sorted list of recent_senders to topic.
The function will be useful to handle topic edits in Recent Topic UI.
We no longer use `/json/users` in the codepath
for bot settings (admin side).
We also specifically don't load human users when
we load bots, so you no longer have to pay for
the server round trip as a side effect of loading
bots. Instead, there is a dedicated `set_up_bots`
entry point.
We also get the bot ids directly from `bot_data` now.
This commit, to some degree, builds on the prior commit
that had us hydrate data from `people.js` instead
of the payload from `/json/users`.
We want to move toward having list consumers
pass us in a list of ids that we hydrate later
in the process. This should help live-update
scenarios. The next commit will describe the
benefits in a bit more detail, using the
concrete example of our bot settings table
in the org settings.
A slightly longer-term goal here is to be
able to ask `list_render` to re-render a particular
id, and this moves us closer to that. But even
before that, this change should eliminate a class
of bugs dealing with stale data, such as when
you manually patch a list (with direct jQuery
hacks) but then later go to sort/filter the rows.
We will now re-hydrate the items in those scenarios.
For the below payloads we want `owner_id` instead
of `owner`, which we should deprecate. (The
`owner` field is actually an email, which is
not a stable key.)
page_params.realm_bots
realm_bot/add
realm_bot/update
IMPORTANT NOTE: Some of the data served in
these payloads is cached with the key
`bot_dicts_in_realm_cache_key`.
For page_params, we get the new field
via `get_owned_bot_dicts`.
For realm_bot/add, we modified
`created_bot_event`.
For realm_bot/update, we modified
`do_change_bot_owner`.
On the JS side, we no longer
look up the bot's owner directly in
`server_events_dispatch` when we get
a realm_bot/update event. Instead, we
delegate that job to `bot_data.js`.
I modified the tests accordingly.
The prior version of "me" confusingly had the same
user_id as one of our bots, so I fixed that.
I also avoid using a test email of 'owner@zulip.com',
which is confusing for earlier tests where I haven't
established "me" as the actual owner of any bots.
Since production testing of `message_retention_days` is finished, we can
enable this feature in the organization settings page. We already had this
setting in frontend but it was bit rotten and not rendered in templates.
Here we replaced our past text-input based setting with a
dropdown-with-text-input setting approach which is more consistent with our
existing UI.
Along with frontend changes, we also incorporated a backend change to
handle making retention period forever. This change introduces a new
convertor `to_positive_or_allowed_int` which only allows positive integers
and an allowed value for settings like `message_retention_days` which can
be a positive integer or has the value `Realm.RETAIN_MESSAGE_FOREVER` when
we change the setting to retain message forever.
This change made `to_not_negative_int_or_none` redundant so removed it as
well.
Fixes: #14854
We shouldn't be checking for #zfilt here, since we haven't done
anything that should cause #zfilt to load. Instead, we verify whether
messages have loaded into the DOM (the condition we actually want) by
checking whether at least one message row is in the DOM.
This adds a way to keep track of max_message_id of a
stream and fetch it using the method get_max_message_id().
This will be useful for sorting streams by most recent
activity which will be implemented in the upcoming commit.
Essentially rewritten by tabbott to have a coherent tracking system,
and provide documentation.
Part of #10794.
Test fails at default timeout value indeterministically
because we have to wait for the server to start and the
above tests to pass, which takes more than the Default Timeout
of this test.
Hence, 4 x Default timeout value is kept.
When switching from Private Messages narrow to
All messages narrow, stream list max-height was not
correctly updated. Stream list max-height was calculated
before new height were updated by browser for
All message narrow.
Inshort:
Stream list max-height was being updated before the browser could
render height for `#global_filters`. Calling resize after narrow
completes removes this issue.
This exists in all versions of the desktop app that we still support,
and will eventually let us delete a bit of annoying compatibility code
from the desktop app’s injected JavaScript.
Signed-off-by: Anders Kaseorg <anders@zulipchat.com>
Previously, the message and event APIs represented the user differently
for the same reaction data. To make this more consistent, I added a
user_id field to the reaction dict for both messages and events. I
updated the front end to use the user_id field rather than the user
dict. Lastly, I updated front end and back end tests that used user
info.
I primarily tested this by running my local Zulip build and
adding/removing reactions from messages.
Fixes#12049.
In passing, we also get coverage for
list_render.get.
This test also implicitly demonstrates that
you can call `widget.sort` directly from some
external event handler; in other words, you
are not locked into the widget's conventions
for setting up <th> tags.
The check here was too late, and it should
have given a blueslip error. We obviously
don't expect these errors at runtime; this
is a convenience for developers creating
new widgets.
This covers how we wire up the filter, and it
covers using `filterer` instead of `predicate`.
I also go away from the strange length-based
predicate that I had in the original test.
The intention behind the original test was
to show that filters could be more than simple
string-matching, but that was just a strange
way to demonstrate it.
1. Replaced the deactivate and reactivate buttons with icons.
2. Added (you) near the current user name to denote his/her account in
the entire user list.
Tweaked by tabbott to reuse the (you) formatting from the right
sidebar here for readability and consistency.
Fixes#6313.
This commit:
- Switches margin for padding on the search closed icon, to ensure we
cover the region to the right of icon as clickable area.
- Applies the click handler that initiates the search to the second
last element of the navbar:
- This will most commonly be the narrow_description element, but may
also be the entire navbar eg in the case of "ALL" or "starred".
Applying this change to user names in "group-pm-with: ..." based
narrows is a little questionable, but there are no other triggers
on these names so this change makes sense for now.
- The narrow_description may also contain links, which need to be
handled correctly so that the behave like links should. We work
around the onClick on the narrow_description, by applying a
handler to <a> tags and invoking stopPropagation.
- We also add CSS to change the cursor to a pointer to make the
search icon change color on hover over the clickable area to
indicate that the search box can be opened with a single click.
- However, since <a> tags are handled differently, we add a hover
listener which makes sure it behaves appropriately. We also increase
the vertical padding of the <a> tags so they cover the entire
vertical navbar region.
We shouldn't add redundant data to page_params. Since we already have
page_params.realm_notifications_stream_id, we can use that value instead
of creating page_params.notifications_stream.
We, however, still need the name of the notifications stream to render
it in templates. Thus we create stream_data.get_notifications_stream().
This commit removes most of the duplicate logic for the stream selection
dropdowns for the settings: `realm_signup_notifications_stream_id` and
`realm_notifications_stream_id`.
We also make minot changes to DropdownListWidget to accomodate the stream
rendering of the format: `#stream_name`.
We finally switch to using stream_ids instead of stream_name everywhere
which makes reading data from page_params simpler.
'get_active_message_people` function is added which returns active
users who have sent the messages that are currently showing up in
the feed.
typeahead fetches the users from 'get_active_message_people` instead
of `get_message_people` and thus shows only active users in the
mention typeahead and excludes deactivated users.
Fixes#14310
This commit makes it so that inline (recipient bar) topic edits follow
a different path from full message row edits in `message_edit.js`.
This commit:
- deletes `.save()` endpoint and replaces all calls to it with
`.save_message_row_edit()` and `.save_inline_topic_edit()`
- deletes `.end()` endpoint and replaces all calls to it with calls to
either ".end_message_row_edit()" and ".end_inline_topic_edit()".
Some extraneous zrequires were added in
3bc818b9f7
This is not a huge deal, but it makes it
appear as if data modules are dependent
on things that they don't really care
about. The tests should provide a bit
of signal on how "deep" an object's
dependencies go.
The set_up_muted_topics_ui and templates have been
refactored to use list_render.
This is done to support filtering and sorting of
the muted stream topics.
This also includes the addition of a new Date muted header.
The div containing options for filtering streams was placed in the
centre. Aligned it towards the right. Had to pass a special check
variable in subs.js:540 to add the specific class for this purpose.
This was a specific scenario where this sort of CSS was to be added,
hence had to make a specific case.
Also, fixed the bottom border color of the search streams bar for night
mode.
Previously, we would always pick up the stream and topic name from
compose_state. This would work for message edits as well when the
composebox was open.
Now, if we are in a message edit, we get the stream and topic of the
message being edited before falling back on trying to populate using
the composebox state.
Fixes#14545.
This commit changes the code to show user according to emails based
on email_address_visibilty_values and the type of user.
1. email_address_visibility = admins,members and guests
Typeaheads are shown according to original emails.
2. email_address_visibility = admins only
Typeaheads are shown according to original email to admins which
were previously shown according to system-generated email of
form "user10@zulipdev.com".
For non-admins, typeaheads are not shown according to emails as
they are not visible in the typeahead itself to non-admins.
3. email_address_visibility = nobody
Typeaheads are not shown according to emails for all type of users.
This commit moves the get_visible_email function to people.js
as this function will be used in other places and people.js seems
relevant file for this.
Tests are added to get full coverage.
We change the user facing interface to allow specifying expected
number of error messages (default=1). Now an average test can look
like:
```
// We expect 3 error messages;
blueslip.expect('error', 'an error message', 3);
throwError();
throwError();
throwError();
blueslip.reset();
```
This change adds a toggle widget to the "add streams" page that
lets the user change the sort order of the streams list. So far,
this supports sorting by stream name, by number of subscribers,
or by estimated weekly traffic.
Previously, in narrow viewports, the "filter"
option would disappear, which was very confusing.
This commit moves the filter streams input to the
next line, making it visible at all viewport widths.
@showell modified the commit message and got Casper
tests passing.
Fixes#12898.
Some uploads can be rejected in the frontend, like when the file
size is too big, without sending the file to the server. Remove the
'Uploading file...' message from the compose box in such cases.
Now, the system uses word='' and an editing=True for rendering an
form for addition of alert words. This is a very vulnerable
way to implement said feature and this commit fixes that.
The addition form has been moved to alert_word_settings.hbs
thereby rendering it only once but always. Now, we do not have
to manually add an empty word and editing for the form to be
rendered.
As part of refactoring, the editing parameter has also been
removed as there is no purpose left.
This updates the logged-in top navbar to display the stream/message
name, number of users, and description. It also replaces the search
bar with a search icon that expands into a full-width search bar.
Co-authored-by: Max Nussenbaum <max@maxnuss.com>
Fixes: #164.
Fixes: #5198.
In ee0d4541b4, we renamed the topic_date
-> stream_topic_history, and in the process renamed some local object
properties from .name => .topic_name, and accidentally change the
type for the data from the server as well.
The test fixtures were incorrectly migrated in the same way, so we fix
that as well.
`stream_topic_history` is a more appropriate name as this
module will contain information about last message of a
stream in upcoming commits. Function and variable names
are changed accordingly like:
* topic_history() -> per_stream_history()
* get_recent_names() -> get_recent_topic_names()
* name -> topic_name
If a file cannot be added for upload because of restrictions in frontend
we call cancelAll immediately in 'info-visible' callback. This would
prevent files that are already added to be cancelled but does not cancel
files that are yet to be added. So we use break to prevent any more files
from being added.
Calling uppy.cancelAll() when a batch of uploads is completed
result in the cancelation of any other batch of uploads that is
in progress. This case happens when a user uploads some files
and then tries to upload another bunch of files before the existing
upload is completed.
The form for entering alert words has been moved above the list
of words.
The list of words will be presented alphabetically rather than
time of addition.
We already know which list widget a `<th>`
tag is associated with when we set up the
event handler, so it's silly to read data
from the DOM to find that widget again
when the handler runs.
This commit eliminates a whole class of possible
errors and busy work.
For some widgets we now avoid duplicate redraw
events from this old pattern:
widget = list_render.create(..., {
}).init();
widget.sort(...);
The above code was wasteful and possibly
flicker-y due to the fact that `init` and
`sort` both render.
Now we do this:
widget = list_render.create(..., {
init_sort: [...],
});
For other widgets we just clean up the need
to call `init()` right after `create()`.
We also allow widgets to pass in `sort_fields`
during initialization (since you may want to
have `init_sort` use a custom sort before the
first render.)
Finally, we make the second and third calls
eliminate the prior updates from the previous
widget. This can prevent strange bugs with
double-reversing columns (although that's
been prevented in a better way with a recent
commit), as well as avoiding double work
with sorting.
This code has always been kind of convoluted
and buggy, starting with the first
sorting-related commit, which put filtering
before sorting for some reason:
3706e2c6ba
This should fix bugs like the fact that
changing filter text would not respect
reversed sorts.
Now the scheme is simple:
- external UI actions set `meta` values like
filter_value, reverse_mode, and
sorting_function, as needed, through
simple setters
- use `hard_redraw` to do a redraw and
trigger external actions
- all filtering/sorting/reverse logic on
the *data* happens in a single, simple
function called `filter_and_sort`
We put this in `scroll_util` to make it more likely
we will eventually unify this with other scrolling
logic. (A big piece to move is ui.get_scroll_element,
but that's for another PR.)
And then the other tactical advantage is that we get
100% line coverage on it.
I changed the warning to an error, since I don't
think we ever expect scrolling at the `body` level,
and I don't bother with the preview node.
I pushed this risk commit to the end of
a PR that had a bunch of harmless prep
commits at the front, and I didn't make
it clear enough that the last commit (this
one) hadn't been tested thoroughly.
For the list_render widget, we can simplify
the intialization pretty easily (avoid
extra sorts, for example), but the cache aspects
are still tricky on subsequent calls.
For some widgets we now avoid duplicate redraw
events from this old pattern:
widget = list_render.create(..., {
}).init();
widget.sort(...);
The above code was wasteful and possibly
flicker-y due to the fact that `init` and
`sort` both render.
Now we do this:
widget = list_render.create(..., {
init_sort: [...],
});
For other widgets we just clean up the need
to call `init()` right after `create()`.
We also allow widgets to pass in `sort_fields`
during initialization (since you may want to
have `init_sort` use a custom sort before the
first render.)
We had a bug where if your peer mentioned you in
message, but then edited the message not to mention
you, the latter wouldn't reset your unread counts
for "Mentions". And the same problem would happen
vice versa.
The fix basically extracts `update_message_for_mention`
and makes sure it handles all combinations of
unread/mentioned flags, instead of assuming
any invariants about which directions of change
are possible.
And then we call that new function from
`message_events.js` whenever we get message
edit events.
Fixes#14544
We use a somewhat more realistic message, mostly
to prep for testing some mention/unread stuff in
a subsequent commit.
We also set message booleans.
Unfortunately, `recent_senders` is kind of awkward
for checking a single message, since its only
public API is for sorting. I don't bother with it.
But I do check the `topic_data` interaction.
Breaking the Casper tests into smaller tests
will make it a lot easier in the future to
hone in on test flakes.
Having small tests adds little overhead--most
of the slowness comes from starting the server.
The only extra steps here are logging in and
entering "Manage Organization", which is two
lines of code.
We split out the custom profile test first,
since the code for custom profiles has the
annoying property that it can only run once
before failing, as it has the side effect
of creating a field name that can't be reused.
We only need to run loops to test flakes, so
this isn't an immediate blocker.
The function message_send_error was messing up
on calls to message.get when we were passing in
string versions of `local_id`. Now we pass in
float ids.
This fixes a traceback where we tried to set
`.failed_request` on to an `undefined` value
that we had instead expected to be a locally
echoed message from our message store.
We stop using `local_id_counter`, which was just noise,
and instead we just make the test more realistic:
- Use 123.04 for our local id on the message that
we're simulating sending.
- Use 127 as the message id that the server gives
us back in the success payload.
We still stub echo functions, but for
one of our stubs (`try_deliver_locally`)
we now exercise one its actual callees
in the stub (`echo.insert_local_message`).
And we're still stubbing some callees
of `echo.insert_local_message`, since
that has all kinds of unwanted side
effects, too.
The main piece we want from
`insert_local_message`, for now,
is somewhat realistic handling of
our local message ids.
We also add a little sanity check
that our timestamp does get plumbed
through to `local_message.insert_message`.
Option is added to video_chat_provider settings for disabling
video calls.
Video call icon is hidden in two cases-
1. video_chat_provider is set to disabled.
2. video_chat_provider is set to Jitsi and settings.JITSI_SERVER_URL
is none.
Relevant tests are added and modified.
Fixes#14483
This adds a new realm setting: default_code_block_language.
This PR also adds a new widget to specify a language, which
behaves somewhat differently from other widgets of the same
kind; instead of exposing methods to the whole module, we
just create a single IIFE that handles all the interactions
with the DOM for the widget.
We also move the code for remapping languages to format_code
function since we want to preserve the original language to
decide if we override it using default_code_clock_language.
Fixes#14404.
This is a prep commit for changes to the top navbar, it adds helpers
to filter.js which will help control the behavior of some aspects of
the redesigned navbar.
Modified by tabbott to add comments, internationalization tags on the
strings, support streams:public, and change various title strings.
We fix this by adding a more expressive data function, with tests, for
whether a filter is on UserMessage data, which would mean that
streams:public could never add additional matches.
We now use `assert.throws()` to test that we're
properly calling `blueslip.fatal`.
In order to not break line coverage here, we have
to remove an unreachable `return` in `stream_data.js`.
Usually we test `fatal` for line coverage reasons.
Most places where we use `blueslip.fatal` fall in
these categories:
* the code is theoretically unreachable, but
we have `blueslip.fatal` for defensive reasons
* we have some upstream bug that we should just
fix
* the code should recover gracefully and just
use blueslip.errors()
It's possible that we should eliminate `blueslip.fatal`
from our API and just throw errors when really important
invariants get broken. This will make it more obvious
to somebody reading the code that we're not going to
continue after the call, and `blueslip` already knows
how to catch exceptions and report them.
When we redraw the left sidebar, we need to tell the
topic list to clear its data structures (and do other
stuff like hiding its popover), since we are clearing
its parent container.
The commit f0e18b3b3e
introduced this regression in late January 2020.
That commit made topic_list use a vdom to avoid
unnecessary updates. Before that, topic_list did
a lot of brute-force redraws, which covered up the
fact that we weren't having stream_list telling it
when the rug was being pulled out from under it.
The boundary between stream_list and topic_list
has always been kind of complicated code, since
topic lists get embedded into the stream list.
The main interactions, though, are basically:
* topic_zoom.clear_topics() - you're leaving
a narrow that may or may not be zoomed
* topic_list.clear() - you're about to redraw
stream items in the unzoomed stream list
* topic_list.rebuild(stream_li, stream_id) -
you're building or updating a topic list
for the newly active stream
Fixes#14465
In case of video embeds, the previous logic used
`data-src-fullsize` or `src` as a key to look
for the metadata of video in `lightbox.open()`,
but while parsing, the key used while storing
the metadata was the video ID.
This doesn't make any sense because video's data
could never be accessed from `asset_map` and we
always needed to lookup the DOM for this.
This commit fixes this by using $img.attr('src')
as a key for `asset_map` for both, images and
videos. Since `src` is the link of preview image
in case of video embeds, it will always uniquely
determine the video ID and we won't loose
anything with the change in how videos handle
things.
Part of #14152.
This changes the payload that is used
to populate `page_params` for the webapp,
as well as responses to the once-every-50-seconds
presence pings.
Now our dictionary of users only has these
two fields in the value:
- activity_timestamp
- idle_timestamp
Example data:
{
6: Object { idle_timestamp: 1585746028 },
7: Object { active_timestamp: 1585745774 },
8: Object { active_timestamp: 1585745578,
idle_timestamp: 1585745400}
}
We only send the slimmer type of payload
to clients that have set `slim_presence`
to True.
Note that this commit does not change the format
of the event data, which still looks like this:
{
website: {
client: 'website',
pushable: false,
status: 'active',
timestamp: 1585745225
}
}
This adds a new test_realm_integer, replacing test_realm_boolean for
testing integer fields like realm_create_stream_policy,
realm_invite_to_stream_policy, and realm_invite_required in dispatch.js.
Fixes#12284
Fields like realm_email_address_visibility and realm_bot_creation_policy
were strings instead of integers in page_params obeject in
settings_org.js tests.
Also use values struct defined in settings_config.js and setting_bots.js
instead of direct values for improving readability.
If folks use an overly broad selector for message rows,
they will accidentally include drafts from the drafts
dialog, which won't have zids. More specific selectors
will be more efficient and possibly prevent strange
behaviors.
For testing convenience, we extract the message.
The UI in the `#settings/notifications` page is updated similarly
to what is done in the `update_global_notifications` path present
in the `server_events_dispatch` file.
This setting is being overridden by the frontend since the last
commit, and the security model is clearer and more robust if we don't
make it appear as though the markdown processor is handling this
issue.
Co-authored-by: Tim Abbott <tabbott@zulipchat.com>
Signed-off-by: Anders Kaseorg <anders@zulipchat.com>
While we could fix this issue by changing the markdown processor,
doing so is not a robust solution, because even a momentary bug in the
markdown processor could allow cached messages that do not follow our
security policy.
This change ensures that even if our markdown processor has bugs that
result in rendered content that does not properly follow our policy of
using rel="noopener noreferrer" on links, we'll still do something
reasonable.
Co-authored-by: Tim Abbott <tabbott@zulipchat.com>
Signed-off-by: Anders Kaseorg <anders@zulipchat.com>
This function returns a list of objects to create a
list_render object, and each item contains the streams
whose atleast one notification setting differs from the
default set by the user.
This is done by comparing the global settings in the
`#settings/notifications` page with those settings
present in the subscribed streams.
Work towards #9228.
This flag was used to delay unread count updates while the bankruptcy
modal was visible. Now that bankrupcty is no longer a modal, we don't
need this flag at all.
Before this we were monkey-patching in the
function `waitForSelectorText` into the
`casper` namespace, but only if you called
`common.initialize_casper`.
This would cause confusion if you expected
that function to be documented by Casper.
Now we just add the helper to `common` in
the `common` namespace.
We also avoid having to reason about what
`this` means by just using `casper` inside
the implementation of `wait_for_text` now.
And we don't bother with a return code that
none of our callers were using, anyway.
We removed the phantom_page_loaded logic in
b13265d135
(July 2017).
Now we just say that the page is loaded
to the console, which can possibly help
us debug glitches where the tests are
executing too early.
We added a really nice feature recently,
called `--interactive`, which lets you loop
through Casper tests without having to restart
it every time.
I am renaming it to `--loop` and adding a few
features:
- The first loop will just run without you having
to tell it to start. (This means you don't have
to sit there while waiting for webpack to finish
and for the server to start, just to launch
the tests again.)
- You specify how many loops you want to run,
which means in the success case, it won't
just keep going forever--it will eventually
stop, giving you an opportunity to refine
the test further without re-launching.
We now trim the headers inside of
`get_rendered_messages`, since any
sane caller of that function just
wants nicely trimmed headers.
(Note that we're now doing the
string manipulation inside of
Zulip code, not Casper code, which
is why I didn't reuse normalize_spaces.)
Starred messages from muted topics were not shown in the starred
messages view. Condition for muting_enabled is modified accordingly
such that the starred messages from muted topics is shown in the
starred messages narrowed view.
Node tests are updated accordingly.
Fixes#13548
We've noticed that many production organizations don't set either an
organization description or profile picture, even large open source
organizations that could definitely take advantage of this feature.
This adds a top-of-page banner that bugs organization administrators
to add an organization description and profile picture, generally
starting on the second login (as we only do it on page load after
notifications are configured).
Significantly tweaked by tabbott to get the right user experience.
Fixes#14019.
We now use `wait_for_message_fully_processed`
to check that messages are fully rendered.
Before this, we had loopholes where messages
sent outside the view were effectively ignored.
Now we explicitly ignore the check for the
one place we do that.
The more important behavior is for messages
that get sent to the current view.
Before this change, the older version of this
function declared victory as soon as we put the
server version of a locally echoed message into
the current message list's data.
This fixes flaky behavior with 07-stars in
particular, since we need the star icon
on our last message to be there before
we click on it.
Because this function is more robust now, we
can remove some redundant checks in 08-edit.js.
I think we could write this test better, but it's not a big deal for
this to break in the rare even that we change/remove one of the 2
strings it interacts with.
When you select a typeahead, it shouldn't
immediately do the action for you; you should
have to hit enter first. Even though 99% of
the time you're gonna confirm the typeahead,
it's jarring when you don't expect it.
You can still add a bunch of default streams
quickly, using only the keyboard, because
we have always had support for the enter
key saving. (and tab and enter also works)
This is a full-stack change:
- server
- JS code
- templates
It's all pretty simple--just use stream_id instead
of stream_name.
I am 99% sure we don't document this API nor use it
in mobile, so it should be a safe change.
We now only use `page_params.realm_default_streams` during
initialization, and then after that we use `stream_data`
APIs to get default stream ids and related info. (And
for the event that replace the data, we just update our
internal data structures as well.)
Long term we should have the server just send us ids here,
since we are now hydrating info from stream data in all places.
We only used get_default_stream_names() in a
test, so now it's being replaced with a function
that just gets ids.
We'll have use for get_default_streams_ids()
in an upcoming commit.
Now if a default stream gets deleted, we just
redraw the table. We always have a small number
of default streams, and the way that we were removing
rows without the actual consent of `list_render` was
really janky (and just a vestige of pre-list-render
code that never got fully ported).
This also makes us consistent with how we handle
added streams (i.e. just call
`update_default_streams_table`).
ASIDE:
Ideally we will update `list_render` at some point to
have an API for adding and removing elements. It does
allow you now to call `data()` to reset its data, but
for now we just build a new `list_render` object every
time.
We stopped needing this with
0329b67048
(Dec 2016).
The function sets `bot.can_admin`,
which was only used in `bot_data.get_editable`.
We removed two tests (and then put back
some test setup that needed to leak down
to the last test).
This is code simplification motivated
by a recent bug that we fixed with some
server changes, but which was really
caused in some sense by our client code
using an overly finicky
condition to check falsiness.
For cross-realm bots, the value of
`user.bot_owner_id` may be `null`, or it
may simply be `undefined`, depending
on whether the server passes `None`
or simply omits the field.
We don't want out client code to be
coupled to that rather arbitrary
decision.
We were doing a `!== null` check instead
of checking for falsiness, which led to
blueslip errors in the past. Because a
bot owner id could be plausibly 0, a falsiness
check would be brittle in a different way.
Now we avoid that ugliness by calling
`get_bot_owner_user`, which either returns
an object or `undefined`.
And then the caller can just do a concise
check for whether `bot_owner` exists.
And we also fix up the crufty code that
was putting `bot_owner_full_name` on to
the object instead of using a local.
We have a bug report for this again, although
it might be on an old branch.
Fixes#13621.
Instead of having logical expressions in templates, it's always preferred
to calculating them in javascript and pass the results as a context. It
also enhances the readability of templates and testing of such logic is
easier in js over templates.
The use case for this are small or fixed tables, which do not need
filtering support. Thus we are able to not include the unnecessary
search input inside the html parent container.
It is not used at present, but will be required when we refactor
the settings pages.
We also split out exports.validate_filter function for
unit testing the above condition.
Before this commit, the reactions code would
take the `message.reactions` structure from
the server and try to "collapse" all the reactions
for the same users into the same reactions,
but with each reaction having a list of user_ids.
It was a strangely denormalized structure that
was awkward to work with, and it made it really
hard to reason about whether the data was in
the original structure that the server sent or
the modified structure.
Now we use a cleaner, normalized Map to keep
each reaction (i.e. one per emoji), and we
write that to `message.clean_reactions`.
The `clean_reactions` structure is now the
authoritatize source for all reaction-related
operations. As soon as you try to do anything
with reactions, we build the `clean_reactions`
data on the fly from the server data.
In particular, when we process events, we just
directly manipulate the `clean_reactions` data,
which is much easier to work with, since it's
a Map and doesn't duplicate any data.
This rewrite should avoid some obscure bugs.
I use `r` as shorthand for the clean reaction
structures, so as not to confuse it with
data from the server's message.reactions.
It also avoids some confusion where we use
`reaction` as a var name for the reaction
elements.
Fixes#14254
You can test this on dev:
* do "-stream:Verona" in the search bar (the minus
sign negates the search here)
* reload the browser
You should see the same search (all streams besides Verona).
Apparently, this test was not allowing the browser to run between the
keypress to start edit and checking to see if message_edit_content appeared.
I'm not sure if this is what has been causing recent flakes, but it
was definitely wrong Casper code.
We had this API:
people.add_in_realm = full-fledged user
people.add = not necessarily in realm
Now the API is this:
people.add = full-fledged user
people._add_user = internal API for cross-realm bots
and deactivated users
I think in most of our tests the distinction between
people.add() and people.add_in_realm() was just an
accident of history and didn't reflect any real intention.
And if I had to guess the intention in 99% of the cases,
folks probably thought they were just creating ordinary,
active users in the current realm.
In places where the distinction was obviously important
(because a test failed), I deactivated the user via
`people.deactivate`.
For the 'basics' test in the people test suite, I clean
up the test setup for Isaac. Before this commit I was
adding him first as a non-realm user then as a full-fledged
user, but this was contrived and confusing, and we
didn't really need it for test coverage purposes.
We want to move more logic to stream_data to facilitate
testing.
Both before and after this commit, we essentially build a
new list of users for typeahead, but now the new list
excludes subscribed users. We can do even better than
this in a follow-up commit.
Before this commit, presence used get_realm_count()
to determine whether a realm was "small" (and thus
should show all human users in the buddy list, even
humans that had not been active in a while).
The `get_realm_count` function--despite a very wrong,
misleading comment--was including bots in its count.
The new function truly counts only active humans
(and no bots).
Because we were overcounting users before this change,
we should technically adjust `BIG_REALM_COUNT` down
by some amount to reflect our original intention there
on the parameter. I'm leaving it alone for now, though,
since we've improved the performance of the buddy list
over time, and it's probably fine if a few "big" realms
get re-classified as small realms (and show more users)
by virtue of this change.
(Also note that this cutoff value only affects the
"normal" view of the buddy list; both small realms
and large realms will show long-inactive users if you
do searches.)
Fixes#14215
We now restrict emails on the zulip realm, and now
`email` and `delivery_email` will be different for
users.
This change should make it more likely to catch
errors where we leak delivery emails or use the
wrong field for lookups.
Given that can_mark_messages_read is called whenever the blue box
cursor stops on a message and that it is calculated purely on the
basis of sorted_term_types, it makes sense to cache the result.
If you were in the "Starred messages" narrow and
your pointer was on a message with the stream/topic
of "social/lunch", we wouldn't move you to the unread
messages for that topic.
I fixed this by removing the code that looked at
the current message's topic. Instead, we only look
at the active narrow to figure out the "next" topic
to go to.
Fixes#14120.
Original email address is shown to admin users in subscriber list when
email_address_visibilty is set to "Admins only" by passing delivery_email
at required places. Email address are not shown to non-admin users when
visibility is set to "Admins only".
Tweaked by tabbott to fix a few bugs and dead code.
Fixes a part of #13541.
This intent is that we'll be able to reuse this when editing streams
as well.
* Rename method: filter_with_new_topic to filter_with_new_param.
* Fix tests and method calls.
This extends our email address visibility settings to deny access to
user email addresses even to organization administrators.
At the moment, they can of course change the setting (which leaves an
audit trail), but in the future only organization owners will be able
to change that setting.
While we're at this, we rewrite the settings_data.js test to cover all
the cases in a more consistent way.
Fixes#14111.
The file populates `windows.i18n`, so now
the file name matches our convention.
Note that the module really just initializes
`i18next` and then does this:
window.i18n = i18next;
It doesn't really add any functionality to
third party library.
Before this test, we were validating the behavior
of `i18next`, but we weren't validating our light
layer that sits on top of `i18next`, which currently
resides in the slightly misnamed `translations.js`
file.
The translations module is now so small that I'll
just quote it verbatim here:
import i18next from 'i18next';
i18next.init({
lng: 'lang',
resources: {
lang: {
translation: page_params.translation_data,
},
},
nsSeparator: false,
keySeparator: false,
interpolation: {
prefix: "__",
suffix: "__",
},
returnEmptyString: false, // Empty string is not a valid translation.
});
window.i18n = i18next;
We now just do `zrequire('translations')` to initialize
the `i18next` library, which allows us to have simpler
test setup and to actually exercise the above call to
`i18next.init`.
This change now gives us 100% line coverage of `translations.js`,
which of course isn't that hard to acheive (see above).
This extracts a new module with three
functions, which we will test with 100%
line coverage:
- show_email
- email_for_user_settings
- get_time_preferences
The first two break several dependencies
in the codebase on `settings_org.js`. The
`get_time_preferences` breaks an annoying
dependency on `page_params` within people.
The module is pretty cohesive, in terms that
all three functions are just light wrappers
around `page_params` and/or `settings_config`.
Now all the modules that want to call show_email()
only have to require `settings_data`, instead of
having a dependency on the much heavier
`settings_org.js` module.
I also make some of the unit tests here be more
full-stack, where instead of stubbing show_email,
I basically just toggle `page_params.is_admin`.
Explicitly stubbing i18n in 48 different files
is mostly busy work at this point, and it doesn't
provide much signal, since often it's invoked
only to satisfy transitive dependencies.
This follows the convention of other code calling into
add_sub_to_table of checking whether the stream settings overlay is
open (and thus in the DOM) before trying to rerender it.
We add these two functions to the API,
so that we no longer have `alert_words_ui`
using private data from `alert_word`:
alert_words.has_alert_word()
alert_words.get_word_list()
And to initialize the data, we have a proper
`initialize` method that is passed in only
the parameters that it needs from `ui_init`.
(We also move the step of deleting `alert_words`
from `page_params` to the `ui_init` module.)
Because it's a bit less cumbersome to initialize
`alert_words`, we now just it directly in the
node tests for `alert_words_ui`.
This is follow up to da79fd206a
I accidentally skipped over pm_conversations. Same
ideas as the bigger previous commit--we pass in params
to the initialize function and do the delete cleanup
within ui_init.
Priyank Patel
Steve Howell
Anders Kaseorg
Ryan Rehman
Tim Abbott
clarammdantas
Vishnu KS
sahil839
Rohitt Vashishtha
Aman Agrawal
Aman Agrawal
Dinesh
Pragati Agrawal
Tim Abbott
Hashir Sarwar
Anders Kaseorg
rebtung
Steve Howell
Jagan
YashRE42
vaibhavrajsingh2001
Roland Crosby
jiviteshjain
Divyanshu Agrawal
Abhishek-Balaji
Puneeth Chaganti
YashRE42
Stefan Weil
Wbert Adrian Castro Vera