This is designed to help PostgreSQL have better specificity and
locality in its indexes. Subsequent commits will adjust the code to
make sure that we use these indexes rather than the `realm_id`-less
versions.
We do not add a `realm_id` variation to the full-text index, since
it is a GIN index; multi-column GIN indexes are not terribly
performant, require the `btree_gin` extension for `int` types (which
requires superuser privileges on PostgreSQL 12 and earlier), and
cannot be consistently added concurrently on running instances.
After all indexes have been made, we also run `CREATE STATISTICS` in
order to give PostgreSQL the opportunity to realize that recipient and
sender are highly correlated with message realm, allowing it to
estimate that `(realm_id, recipient_id)` is likely as specific as
matching a given `recipient_id`, instead of as likely as matching
`realm_id` times matching a `recipient_id`. Finally, those statistics
must be filled by `ANALYZE zerver_message`, which is run last.
Using `COUNT(*) FILTER (WHERE ...)` allows getting counts of different
subsets with only one giant join. This makes the query significantly
more performant.
Matching the topic exactly, as opposed to case-insensitively, is not a
common operation, and one that we want to make difficult to do
accidentally. Inline the single use case of it.
We now have a `realm_id` on Message; use it, rather than having to
check the sender's realm. This is theoretically different for
cross-realm bots, but these changes are all in tests where that does
not apply.
This algorithm existed in multiple places, with different queries.
Since we only access properties in the UserMessage table, we
standardize on the much simpler and faster Index Only Scan, rather
than a merge join.
We don't have a "Manage user" tab in your own profile, so it would
throw an exception to do the default thing here, and the mechanism for
editing your own settings should be the main "your profile" panel,
which is a bit bigger and more user-friendly.
I'm not totally convinced that the extra logic for not having a
"Manage user" tab is worth it; the alternative of figuring out a
non-confusing label would not be terrible here. But this version is
non-broken, and it'll be easy to change later if we change our mind.
This commit adds Tippy tooltips to the edit pencil button
in the user profile. Now, when a user opens another user's
profile, they will see the tooltip 'Manage user', and when
they open their own user profile, they will see the tooltip
'Edit profile'.
This commit unsets the padding of the bot icon and, as the anchor
tag has been replaced with the i tag, we have also set the color
and pointer. This is done so that when the user hovers over the
icon, it will appear as a clickable button.
This commit moves the bot icon to the left of the name in the user
profile and adds a margin-left and margin-right to the name to ensure
there is enough space between the icon, name, and edit pencil.
This commit redirects the edit pencil in the user profile modal to
the 'manage user' tab whenever a user with permission to manage
other users via the user profile opens another user's profile.
However, we still want to redirect the edit pencil to
'settings/profile' if the user opens their own profile. The user
management permission is granted if the user is an admin or the
owner of the bot. However, we do not want system bots to have
access to the edit pencil or the 'manage user' tab. Therefore,
a new variable called 'can_manage_profile' has been introduced
to manage all these permissions, and the CSS has been updated
accordingly.
To redirect to the manage user tab without opening another modal,
I have extracted the toggler. This toggler will store the
component, and if the edit pencil button is clicked, we can use
the goto function to redirect to a different tab.
Changed the id names of both the edit pencil icons to explain
better of what they do.
This commit increases the gap between the labels and input/select
of the user profile manage user/bot tab for clearer visibility.
It also reduces the gap between the 'deactivate user' button and
the last input to make it look better.
This commit disables the "Deactivate User" button inside the
"manage user" tab in the user profile modal. Additionally, a
tooltip is applied to the button by wrapping it inside a div
with the tooltip applied.
This commit fixes the inconsistent height of the user profile when
switching tabs. We now have four tabs in the user profile:
"Profile," "Streams," "User Groups," and "Manage User." However,
the "Manage User" tab has footer buttons that need to remain
sticky at the bottom without changing the overall height of
the modal.
To achieve this, we wrapped the footer inside a div element and
assigned it a class called "manage-profile-tab-footer." The main
body of the user profile is given a height of 60vh. However, for
the "Manage User" tab, we decreased this height to 52vh and
allocated the remaining 8vh height to the div element to
accommodate the footer buttons. This ensures that the user
profile maintains consistent height when switching between tabs.
However, we have a 1px border in the footer, so to ensure that
the height is consistent, we reduce the height of the modal
content by 1px, making it calc(52vh - 1px).
To implement this, CSS code was added to the "popover.css" file
and accessed through the class specified in the "user_profile.js"
file.
The default padding of the buttons inside footer is 20px, but
with the above solution on large screens the buttons are not
aligned properly, so removed the padding top and bottom
and instead applied the flex box and property to align the buttons.
This commit migrates the "Manage Bot" modal to the user profile modal,
with the same explanation as in the "Manage User" modal commit.
However, in this commit, we changed the permission of the "Manage User"
tab so that non-admin users can also see the "Manage Bot" tab if they
have created the bot. Additionally, since we can't make changes to
system bots, we check if the bot is a system bot and hide the
"Manage Bot" tab accordingly.
Fixes: #21806
This commit migrates the "Manage User" footer buttons to the user
profile modal. We made this change because we don't want the buttons
to scroll with the content; instead, we want them to be sticky at
the bottom at all times. Therefore, we moved the buttons to the user
profile modal.
This commit migrates the "manage user" modal to the "Full Profile"
modal. Refactored the show_edit_user_info_modal function since
earlier, we used to have a separate "Manage User" modal.
Consequently, we checked if we were coming from the
user_info_popover and then built the dialog widget accordingly.
However, with this new change, we no longer need to build the
dialog widget. Therefore, removed that part and now just pass
the container of the 'user profile modal' to get the content.
Previously, for deactivation, we used to have a separate dialog
widget. But now, since we have a dedicated function to handle
this case, refactored the deactivation code to use the
'confirm_deactivation' function.
Additionally, created two new functions to handle the loading
spinner. Since we will need these functions in the future for
the "Manage Bot" modal, we marked them as exported.
"Since we do not want to show the 'Manage User' tab to the user
looking at their own profile, in the 'can_manage_user' function
that we use to render the 'Manage User' tab, we check if the user
profile popover belongs to the same user. If it does, we set it to
'false' instead of 'true,' ensuring that the 'Manage User' tab is
not visible.
Added a new tab in the user profile modal 'manage user'.
Fixes: #21806
Rearranged some code to align with our current patterns and improve
readability. Renamed `show_user_group_info_popover`. No functional
changes in this commit.
This is a preparatory commit before we migrate `user_group_popover`
from Bootstrap to Tippy library.
The previous implementation was weirdly sharing the logic around
`current_message_info_popover_elem` with the user info popovers based
on a message; very likely an unfortunate latent bug caused by
copy/paste.
To address that, we need to add dedicated functions like
get_user_group_popover_items to avoid breaking keyboard navigation
with this extraction.
With the previous commits, we have now implemented the subscribe
widget with a separate heading. However, the font size and the gap
of both the heading of the widget and the streams list look bad,
so this commit fixes that.
Since the heading of both the widget and the list have the same CSS,
instead of duplicating it, we used the same class for both and
reduced the font size.
This commit adds a reset function for the subscribe widget so that
whenever someone subscribes to a stream from the dropdown, after
the subscription, the dropdown will return to its initial state
with no stream selected. Additionally, it will display a label with
the subscribe button disabled and a tooltip.
This commit adds a tippy tooltip to the subscribe button in
the user profile. We show this tooltip only when there is no
stream selected and the button is disabled. However, on any
change, we enable the button and hide the tooltip.
This commit adds a subscription widget to the user profile,
including the logic to prevent non-admin users from seeing the
subscription widget of other users. Additionally, as it is not
possible to subscribe generic bots to streams, and the user should
not be a deactivated user, we check for these conditions before
displaying the subscription widget.
To ensure that the alert for both subscribing and unsubscribing
appears on top of the subscribe widget, changed the location of
the alert to be displayed at the top.
Additionally, considering that no stream will be initially selected,
we have made the decision to disable the subscribe button. Once the
user selects a stream, we will enable the subscribe button
accordingly.
Changed the add_user_ids_to_stream function inside subscriber_api.js
to support self subscribe also, so that we don't have to duplicate the
logic in user_profile.js
Created a separate file for the subscribe widget called
user_profile_subscribe_widget.hbs.
Fixes: #18883
The closest($stream_row) logic is unnecessary, and worth cleaning up
because we'll be adding other calls from places that don't have a
stream row object.
narrow is a term that is intended to only apply to a message feed
view; it comes from "narrowing the set of messages you're looking
at from the All messages" so switching it to something different
makes sense.
Alex Vandiver
Lalit Kumar Singh
Daniil Fadeev
Tim Abbott
palashb01
Aman Agrawal
Anders Kaseorg