This commit adds call_function_periodically helper function
which will be used to call functions periodically using
setTimeout. Currently, this new function is used to send
presence requests and trying reload.
We retry reloading repeatedly at an interval of 30 seconds,
to handle the case where window.location.reload has no
immediate affect.
Previously, setInterval was used for this, but this commit
replaces it with nested setTimeout calls.
This change will help us in avoiding a large number of
requests to `/` in case when browser tries to "catch up"
pending calls after unsuspend.
This finishes up the work from 057ee6633a.
We had this "get_reload_topic" helper for user that has "topic"
encoded as "subject" from a older release. This is unlikely to cause
problem now because we no longer use "subject" for reloading since 2018.
Signed-off-by: Zixuan James Li <p359101898@gmail.com>
Previously, we deleted all reload tokens on each reload, which
created a race condition if there were multiple tabs open.
Now, we continue to delete tokens after using them, but if a
token is not used it is preserved for a week before being deleted.
Fixes#22832.
We have observed infrequent storms of accesses (tens of thousands of
requests to minute) to `/` after an event queue expires. The current
best theory is that the act of reloading the page itself triggers a
focus event, which itself triggers a reload before the prior one had
had time to do anything but send the network request.
Since the `focus` event here is merely as a backstop in case the
synchronous reloading and deferred reloading fail, we need only run it
once.
Prevent a non-immediate reload from being scheduled while an immediate
reload is already in progress. This is highly unlikely in practice,
but is a reasonable safeguard.
Prefer a regexp match over using String#replace to strip expected
prefixes and suffixes because (a) it implicitly verifies that the
input has the expected format and (b) it won’t unexpectedly strip from
the middle of the string.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
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.
ES and TypeScript modules are strict by default and don’t need this
directive. ESLint will remind us to add it to new CommonJS files and
remove it from ES and TypeScript modules.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
Prettier would do this anyway, but it’s separated out for a more
reviewable diff. Generated by ESLint.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
Prettier would do this anyway, but it’s separated out for a more
reviewable diff. Generated by ESLint.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
The orig_initial_pointer variable was part of the implementation for
ensuring server-initiated reloads preserve the user's selected message
and scroll position (so that they are not disruptive). Previously,
the logic did some unnecessary contortions to ensure the two goals:
* The `pointer.js` logic knows what the server thinks the pointer is.
* The `message_fetch.js` logic knows what anchor to use to center it's
home view fetch.
It's a lot cleaner to do this by not mutating page_params.pointer.
This is not always a behavior-preserving translation: _.defaults
mutates its first argument. However, the code does not always appear
to have been written to expect that.
Signed-off-by: Anders Kaseorg <anders@zulipchat.com>
We now treat util like a leaf module and
use "require" to import it everywhere it's used.
An earlier version of this commit moved
util into our "shared" library, but we
decided to wait on that. Once we're ready
to do that, we should only need to do a
simple search/replace on various
require/zrequire statements plus a small
tweak to one of the custom linter checks.
It turns out we don't really need util.js
for our most immediate code-sharing goal,
which is to reuse our markdown code on
mobile. There's a little bit of cleanup
still remaining to break the dependency,
but it's minor.
The util module still calls the global
blueslip module in one place, but that
code is about to be removed in the next
few commits.
I am pretty confident that once we start
sharing things like the typeahead code
more aggressively, we'll start having
dependencies on util. The module is barely
more than 300 lines long, so we'll probably
just move the whole thing into shared
rather than break it apart. Also, we
can continue to nibble away at the
cruftier parts of the module.
The compose_state.recipient field was only actually the recipient for
the message if it was a private_message_recipient (in the sense of
other code); we store the stream in compose_state.stream instead.
As a result, the name was quite confusing, resulting in the
possibility of problematic correctness bugs where code assumes this
field has a valid value for stream messages. Fix this by changing it
to compose_state.private_message_recipient for clarity.
Aman Agrawal
Sahil Batra
Zixuan James Li
evykassirer
Alex Vandiver
Anders Kaseorg
YashRE42
Anders Kaseorg
Tim Abbott
Tim Abbott
Anders Kaseorg
Steve Howell