Allow addressing multiple people with @

[ashnur]

singular, current case:
!somefactoid @ nick1

I think it would be nice to have:

!somefactoid @ nick1, nick2, nick3 too.

[ljharb]

Personally I'd prefer multiple @ signs, although that'd be slightly more characters. Both syntaxes would be nice.

[ashnur]

@ljharb you mean
!somefactoid @nick1@nick2@nick3 ?

I omitted the whitespace because I am unsure if it's needed at all.

[gkatsev]

whitespace delimited would be best.
!factoid @foo @bar @baz

[ljharb]

yes, with the whitespace.

[imbcmdth]

The commit supports comma and space (and no delimiter since @ isn't valid in nicknames). It does it's best not to capture @ in the "command" portion of the string.

For instance, !g @media will catch @media as an intent (it does that currently, anyway) but it won't catch !g @media queries since the possible intent isn't followed by an intent or the end of a line.

[eligrey]

Really good point, maybe we should change the @ syntax to something less ambiguous like @user1 user2 etc !command here

[gkatsev]

We shouldn't change the command for several reasons.

• familiarity in-channel
• expectation based on similar usage
• intuitiveness

[ljharb]

Also, the command is required, the user names are optional. Optional things go last.

[robotlolita]

Bracketing the names in an @user would make it unambiguous in all situations, I think:

!g @media @(foo)
foo @(blah)

Edit: can also support multiple users

foo @(user1 user2)

[gkatsev]

Perhaps

!g @media -- @bar
!foo @bar
!something -- @bar

So, the -- would be optional.

[imbcmdth]

@robotlolita - Except then you can't mindlessly tab-complete. Right now I tend to @so<tab> but that proposal would require @so<tab>) - I can see a lot of people missing the ending ) and screwing up intents.

EDIT: Heck, I just screwed up the intent!

@gkatsev - Would anyone remember the -- since it would be required so very infrequently? It seems to me that complicating the parsing for @media is a bit silly. I can't even think of another "@"-bearing search that would present an issue.. we aren't #css!

[robotlolita]

That seems to work pretty well too, it's also more English-like, rather than Lispy

[gkatsev]

It's how unix lets you pass extra arguments to a subscript that you're running.

[eligrey]

@robotlolita parenthesis have query ramifications on google and I sometimes use it for more complex searches

[imbcmdth]

The only potential issue I see with -- is that it is ambiguous for something like:

>>> var a = 10; setTimeout(print, 1000, a); a-- @nickname

[robotlolita]

Hm, if you make sure there's whitespace before it, it's likely to be less of a problem a -- is still valid, tho.

[gkatsev]

Make it three?

[ashnur]

4! :)

[robotlolita]

Yeah, b ---a is still valid :P

[ashnur]

--- @ somenick is not

[ashnur]

would this maybe work? http://jsbin.com/sivod/1/edit?js,console,output

[ashnur]

with delimiter --- and 'g @media' as testcase: http://jsbin.com/sivod/3/

[ashnur]

with delimiter @ http://jsbin.com/sivod/5/edit

it would look like !g @media @@somenick @someothernick