We only want to change cases where we're talking about the hostname; HTTP
requests should still go to staging.humbughq.com for now.
Before this commit is deployed the hostname of staging.humbughq.com should
be changed to staging.zulip.net on the VM.
(the same for prod)
(imported from commit 7412530773f720ac227f40061c9ddb1a851e19bb)
lb0.zulip.net will proxy connections to the relevant backend servers.
Depressingly, SSL certificate verification of the backend servers is not
performed at this time, see:
<http://trac.nginx.org/nginx/ticket/13>
The above-mentioned bug has existed since 2011, but a CVE was not
allocated until January. The nginx developers don't seem to care. Sigh.
In any case, this is of somewhat limited impact at Humbug, since we can
have reasonable confidence that communications within AWS are not
subject to active MITMs. Passive MITM is not a concern, because the
traffic *is* in fact encrypted.
(imported from commit c96e1235fc17192c7452e0417a1309cfcda62de2)
The older certificate had an "Organization: Luke Faraone" which was sub-
optimal. This new certificate is also based on a key that has not been made
available to our app instances in the past.
(imported from commit 0a8070bb16e4c640e6b4c06afee07ccf88495327)
On EC2-VPC we have the ability to attach multiple addresses to one
interface, and multiple interfaces to one machine.
We should configure those interfaces whenever our system boots, and
ideally whenever networking is restarted.
This commit adds a script that is executed once eth0 is brought up that
proceeds to configure all subsequent interfaces, real and virtual.
The script is configured to be installed (along with the helper script
that calls it) on all systems via Puppet.
(imported from commit fdc153ef649edbb8fedd40ff4d77262aae593c39)
* Do the string processing to extract the current token once
when autocomplete is started in `source` rather than repeating
it once per potential completion in `match`.
* Store a property for whether we're completing emoji or mention
instead of performing different tests on each step.
This reduces the lag on emoji autocomplete.
(imported from commit 7aa3f762eaf4286608809a09a3cddbbe3dfcabf9)
Fixes issues with e.g. foo@hamlet.com (where hamlet is a user)
The \b was ineffective because @ made it always on a word boundary.
Instead, use the negative lookbehind trick from the URL regex.
(imported from commit fdca9bd686e4f8747e67b412cba1fa7c5c9391aa)
This allows us to e.g. match "#1329" in "Bug #1329", even though the
place between a space and a # is not a word boundary.
Also this commit factors out some repeated code used for both in-message
and subject filters.
(imported from commit 5f7d80a58e76e51ea07fed050c88c5251faaaacd)
Tagged [manual] because we need the Zulip-latest.dmg,
zulip-latest.exe, and Linux updates to actually exist on prod.
(imported from commit 67a514a3f8844b939163ca330144ae80cf028132)
This sort of browser-based OS detection isn't perfect, and can be
really fragile, but in this case it's not a big deal if we don't get
it exactly right; the user can always pick a different one.
(imported from commit 70fdaf8f82252b17a3ffc347e239db5963363fbd)
Since f8fbf70c8502 had troubles in FF and people did not like the
message-info-on-hover, this commit hopefully will work better. Tables
are likely the culprit with using dropdowns in FF, so when we stop having
the messagelist be at table, we can switch back to dropdowns which are relatively
positioned rather than absolutely on the page.
(imported from commit c49a5cd45f3439d089146771e2aa0cee6431125a)
This didn't /seem/ to be hurting anything before, but we really
shouldn't be doing it more than once.
(imported from commit 8d4ad7f8376227db723259ec9293262233b0a69a)
This way if you refer to "trac #253" in the subject, it's super
convenient to get from your recipient bar to the ticket.
A note on performance: this part of rendering for 1000 messages takes
about 3.5ms for messages with 1 match; this is small compared to the
overall time for to_dict_uncached for that many message objects, so I
think this is OK for now.
(imported from commit 5bdc2b8415d7599d59eb554739f545c485b78d5a)
Renamed show, hide and clear to show_box, hide_box, and clear_box,
to make it a little more obvious what each one does and make them
greppable.
Also, hide and clear weren't being used by any other modules, so
these functions are no longer exported.
This resolves Trac ticket #1503.
(imported from commit 904d596ef5b8dc2154dc01ed7c9c99a54dc9b31e)
From Trac ticket #1503:
All of the calls to compose.set_mode() should be calling compose.start()
instead. Nearly all are impossible to trigger while already composing,
and by calling compose.set_mode() they just do the stuff in set_mode()
twice. The only case where this would change behavior, that I can see,
is if you press C, Shift-Tab to unfocus the compose box without
dismissing it and then press c (or vice versa). I think it's okay if we
clear the input fields in that unlikely case.
(imported from commit ba7f181ec9d1df90a443b0a754462a3a201dcabb)
Previously we were checking for webkit before checking for
$('<audio>') support, but that cuts out non-webkit browsers that support
the basics of the HTML5 audio api.
Now we actually run the feature test in order to enable it, and only
check for webkit when enabling webkit-specific desktop notifications
(imported from commit 851eed86af167d0530f7e1793e2ca1f9b4cdd71d)
For completions, since we have no representation of categories, users
might expect lexical ordering.
(imported from commit d2e27dd4d8452df6ac48d44551802d7e65519115)
Instead of `compressed_js` in the `compress` library, we now use
`minified_js` from the `minified_js` library. (CSS is not affected by
the change.)
(imported from commit 07aea9efbc83d30b354a7e044070563e93f7525a)