We received a complaint about the generation of multiple duplicate
drafts for a single message. It was discovered that the likely cause
of this was how we were handling clients that were frequently
suspending/unsuspending, we would initiate a reload when we discovered
this, and expect the `beforeunload` handler to save the draft. This
behaved correctly, however, we would also save the compose state and
fill it in via `preserve_state` in reload.js. The important detail
here is that `preserve_state` would not encode and preserve the
`draft_id` for the current message, partly because it had no way of
knowing the `draft_id` of the draft... since we have not saved it yet,
the `beforeunload` event happens after `preserve_state`. As such,
performing any action that would trigger a draft to be saved, eg
pressing Esc to close the compose box, would save a duplicate draft of
the same message.
To resolve the above bug, we (1) ensure that we call
`drafts.update_draft()` in `preserve_state`, this returns a draft_id
to us, which we (2) ensure that we encode as part of the url and (3)
set on the `#composebox-textarea` as a `draft-id` data attribute,
which we check the next time we try to save the draft, post reload.
Note that this causes us to save the draft twice, once from
preserve_state and then again from the `beforeunload` handler, but we
do not add two drafts since the second update_draft call just edits
the timestamp because it finds the `draft-id` data attribute on the
`#composebox-textarea` set by the first call.
Previously, opening a draft and closing it without changing the
content would cause us to show the "saved as draft" tooltip. This was
annoying and would cause annoying UX after fixing a bug related to
saving drafts when reloading, as such, this commit removes the above
behaviour by introducing a simple check on whether the draft contents
are edited.
Fixes the link to the Neil Green presentation on TypeScript
vs Coffee Script vs ES6.
This is a change from slides to a video becasue the slides are
no longer available.
A page can have either `white` (from `landing_page.css`) or `gray`
(from `portico.css`) background color depending on
webpack chunking order. So, this fixes that bug.
In message header search bar, users didn't use to get any typeahead
suggestions if a normal filter follows search filter.
E.g.: query => foo bar stream:D
In the above case, users didn't use to get any typeahead suggestions.
This was because we had set that the callers of 'parse' function can
assume that the 'search' operator is present in the last in the query.
Because of which `get_search_result` function (in search_suggestion.js)
didn't use to show any typeahead suggestions as it used to assume that
the latest typed query is for search filters.
Fixes part of #19435.
The warning was fixed in python-jose 3.3.0, which we pulled in with
commit 61e1e38a00 (#18705).
This reverts commit 1df725e6f1 (#18567).
Signed-off-by: Anders Kaseorg <anders@zulip.com>
OIDC config features a get_secret call (so it requires adding an import)
as well as having a bunch of its instructions in the form of comments on
the various keys of the config dict - thus users should really update
settings.py to fetch all of that.
In https://github.com/jorisroovers/gitlint/pull/246 I split the
gitlint package into gitlint and gitlint-core, where the latter avoids
pinning exact versions of its requirements so we can use it again.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
We directly pass operators to remove dependency on narrow_state
module. This avoids a circular dependency of `filter` module
which is evident on the `/devtools/integrations/` page.
This commit extends dialog_widget class by adding a new
optional paramter validate_input which will be a function
to validate the inputs in the dialog and will be called
before showing the spinner and calling the on_click function.
Currently, the password change modal uses this paramter to
validate that the old and new password inputs must not be
empty. Since the spinner will not be initiated in the case
where form is invalid, we need not hide the spinner after
showing the error and thus we can simplify the code to use
ui_report.error to show the error messages of empty fields.
This fixes unexpected cursor repositioning behavior when the cursor was
positioned before or inside the "Quoting..." element.
The comments document the new logic, but roughly we aim to just
preserve the logical position of your cursor after replacing the
placeholder.
We also factor out a shared variable for the "Quoting..." string which
will allow us to tag it for translation in a future commit.
If you used "Quote and reply" to start composing a message, and
started typing before receiving the original message body from the
server, we ended up resetting your cursor to the start of the line
after the quote for two reasons:
* We were incorrectly fetching the pre_cursor for our replacement
operation before doing the server fetch, which meant we ignored any
editing done while waiting for the server to respond.
* Worse, we actually fetched the original cursor position before
inserting the "[Quoting...]" placeholder text. So we were guaranteed
to have at least some amount of error in the cursor position.
Fixes#20379.
Continuing the efforts to reduce dom trashing from the previous
commits, here we remove the third forced reflow by reordering the call
to $(".top-messages-logo").show() via narrow.reset_ui_state(), such
that it happens before the other DOM writes in
recent_topics_ui.hide().
Tweaked by tabbott to avoid adding an unnecessary if/else statement
around recent_topics_ui.hide.
This is a prep commit towards pushing the reset_ui_state calls upwards
into recent_topics_ui, in order to prevent a forced reflow. One side
effect of this change is that we add a call to
`narrow_banner.hide_empty_narrow_message()` from `narrow.activate()`.
This likely has no visible effect, in that the message list rendering
process would end up setting the narrow_banner state correctly, but
logically it could in the future avoid a banner from a stale banner
incorrectly appearing before we've rendered the current view.
This is a prep commit towards extracting a reset_ui_state function
which we can call from here, narrow.activate(), recent_topics.show()
and recent_topics.hide().
We want that function because it will enables us to prevent a forced
reflow when navigating from recent_topics to stream: xyz.
Going through the description of commit
a150b9b0ae is highly recommended since
this is a related issue.
We had received a complaint on chat.zulip.org about navigation with
the keyboard `n` key being significantly slow (~5 seconds), the first
time `n` was pressed when starting from the recent topics view.
It was difficult to reproduce the amount of lag that was reported, but
running chrome with the profile tab set to 4x slowdown helped get
close to it.
Based on profiling from the original report, as well as locally with
chrome set to 4x slowdown, led to the realisation that recent topics
to stream navigation involved a lot of dom thrashing, and so this
series of commits aims to prevent this path from causing forced
reflows.
In this commit, we reorder the calls to $(...).show() in
recent_topics_ui.hide(), this prevents the first reflow in this path,
most likely because displaying message_feed_container before
message_view_header_underpadding was guaranteed to cause style
recalculations since the underpadding is visually above the message
container.
This causes a net 60 ms decrease in the first renarrow, an ~ 70 ms
increase in the second renarrow and an ~ 5 ms increase in the third
renarrow but, more importantly, it eliminates one reflow and sets on a
path where we can achieve strong gains in subsequent commits.
Migrates the `/update-subscription-settings` api endpoint to the
`ignored_parameters_unsupported` model, which is also currently used
by `/update-settings` and `update-realm-user-settings-defaults`.
This change is a step towards preparing for an eventual migration to
have all endpoints return an `ignored_parameters_unsupported` block.
Previously the `/update-subscription-settings` endpoint returned a
copy of the data object sent in the request.
Fixes#15307.
We show "Please enter your password" error inside the modal
if the "Old password" input is empty and "Please choose a new
password" error if the "New password" input is empty and do
not send a request to server.
Fixes#19901.
Sahil Batra
Anders Kaseorg
YashRE42
Eeshan Garg
Ganesh Pawar
AEsping
Aman Agrawal
akshatdalton
Mateusz Mandera
Tim Abbott
Manan Rathi
Lauryn Menard