These calls were expensive and unnecessary. We already update
fade/unfade classes deeper in the call stack, when we render
the messages inside message_list.js.
(imported from commit 08fe028462b6d4569d9798a290dd7b26eb21fb01)
Use information from the server to figure out if we should prompt for
bankruptcy, rather than trying repeatedly inside load_more.
(imported from commit ccb8cb1ce482b8bf3d343e7324fef7981880282d)
1) The class Filter now lives in its own module.
2) The function canonicalized_operators() is now a class method on Filter.
3) The function message_in_home moved to filter.js and became private.
4) Various calling code had to change, of course.
5) Splitting out Filter helped simplify a few tests.
(imported from commit e41d792b46d3d6a30d3bd03db0419f129d0a2a7b)
Previously we would just discard the results of get_old_messages,
which meant that any messages sent either while you were doing the
tutorial or that you started out with (as in the case of the CUSTOMER3
experiment) would be lost until you reloaded.
(imported from commit f5280c091ab6ed7c2af6eb8fe49c0fa6b997ac97)
The compose_fade has three public exports:
set_focused_recipient
unfade_messages
update_faded_messages
All code was pulled directly from compose.js, except for the
one-line setter of set_focused_recipient. The focused_recipients
variable that used to be in compose.js was moved to compose_fade.js,
hence the need for the setter.
(imported from commit 462ca5d0d0bd58612d0197f3734a8c78de8c6d30)
There are also one or two places we don't need to use it for security
purposes, but we do so for consistencey.
(imported from commit aa111f5a22a0e8597ec3cf8504adae66d5fb6768)
On a page load or reload, the browser will, at an idle time, scroll to
the top of the page. We can't intercept this browser-induced scroll,
so to make sure it doesn't in interfere with our scrolling to the
correct place in your message feed, we let the browser scroll happen
before the work we do on page load.
(imported from commit f5f441ab90bcdb8404e05caea3c6da81a3a6fc1e)
This commit makes keep_pointer_in_view() less aggressive when
the pointer is toward the top of the screen. If the pointer is toward
the top of the screen, then as long as it's fully on the screen, we
don't pull the rug out from under the user and change the pointer.
An important benefit of this change is that we fix trac #1608,
which was a bug where autoscrolling interacted with
keep_pointer_in_view() to push messages past the top of the screen
before they were read.
(imported from commit e39926df99bfaedd5c0757f1241887ccd9b93fab)
The helper functions now return true when a message is on
screen, not when it's off screen, and the names have been
changed accordingly. I also eliminated the at_end parameter,
which was kind of abstract, by having message_is_far_enough_up
and message_is_far_enough_down handle those details.
(imported from commit cdb1543e430f49f23eb1b3a88d9aaff95ce7ca74)
* Move the state into a closure to duplicate it for multiple queues
* Use _.debounce instead of manual setTimeout fiddling
* Have it handle manipulating the message flags
(imported from commit 938f51fd666131a3cec5901d3f3fdd39e203b462)
Functions were supposed to check that messages were unread before
passing them to process_read_messages, but some didn't.
The `mark_messages_as_read` function was essentially that, so
take that name.
(imported from commit 2917fe30d2defb8a047ec32e1bc70d379779276b)
When you read messages in a narrow and then un-narrow, collapse
adjacent messages read in the narrow into a summary row that can
be clicked to expand those messages.
Scoped to staging with feature flags.
The implementation of this within our current MessageList is rather ugly.
(imported from commit bcb3a39d8c0c334136fe86318f18ead03f0f50bf)
Tim Abbott
Steve Howell
Jessica McKellar
Zev Benjamin
Leo Franchi
Scott Feeney
Steve Howell
Kevin Mehall
Zev Benjamin
Tim Abbott