anyway, I guess the point was, is your screen height really small? is your window perhaps not maximized?

Window is maximized. Screen resolution is 1366X768

ok

@JasonC Thanks :)

Yeah so the scroller is off by the topbar margin then

Yeah. Right now I compensate by adding a margin-top to the container, that was just driven by getting the "load older messages" button in the right place.

The easiest thing to do is probably to just offset all the anchor elements upwards.

That's weird. There's no anchor links. It must be scripted.

I suppose I could insert an actual div spacer up there, but...

Well I already put a white backdrop behind the top bar to cover its opacity, but its position is fixed. I guess I could make it a block and put it in the right place.

@JasonC - Just tie into the .reply-info element's click event, delay by 20 milliseconds, and scroll the page an extra [insert margin-top px amount here]

Ehhh that makes me feel funny inside.

It's either that, or override the actual code used by SE for the scrolling (which from experience is a pain, although somewhat rewarding)

I think I can do it properly just by rearranging the layout a bit.

And as a bonus it could solve any other unforeseen vertical offset issues.

Once you go down the unintended side effects path, forever will it dominate your userscript. Consume you, it will.

Oh nope, hack it is.

The scrollbars on the body. So anything I put up there gets scrolled with the chat.

Can't do it unless I move the scrollbar to one of the inner containers, which sucks.

"[the third attempt to create] the `Stack Overflow for startup marketing'"

@Guero presumably there'll be a round of funding so that they can buy the answers & learn how to market those answers.

Can somebody ping me

@JasonC Sure thing.

Or do you want a response to an earlier message?

No that's fine I'm testing something else too.

Okay.

Darn it. What triggers the little blue notification icon to appear on the chat.meta.stackexchange.com/ topbar? The one on the right.

Sometimes it appears with chat pings, sometimes it doesn't.

@Shokhet \o

You can ping yourself if you manually reply to one of your own messages, like explicitly type :id at the front where id is your own message id.

I think invites also appear there

@JasonC I'm pretty sure it's any uncleared ping. I haven't cleared my self-ping yet, and I got a blue 1

Huh.

Oh wait it has something to do with... not uncleared, but whether you've been active in the room or not I think.

Wait can somebody ping me one more time

I did turn off your script first, though.

@JasonC Sure.

@Shokhet Now trying it with the script on.

Dammit, nothing.

I got one a few minutes ago then it disappeared.

Maybe... reply to a much older message?

@Shokhet I got a blue number both on chat.meta.SE.com, and in the top bar in this room.

And clicking on that number goes where I expected it to.

@Shokhet One minute

Yeah. The blue number only appears for an uncleared ping after a page refresh.

And only goes away after unclearing, and then refreshing.

Except this one hasn't gone away yet :0

@JasonC @JasonC

@Shokhet New refresh and now it's gone.

#define JasonC buffalo

@JasonC Come on

JasonC JasonC JasonC JasonC @JasonC

@Shog9 Cute.

Success!

I went from no blue thing to a "2", which I think might be my new personal blue thing record.

Although mine's not going away now either. Maybe it's refreshed slowly or something.

Wait a minute or two, and then refresh again. Worked for me.

This is why debugging minimized code is such a pain sometimes

function f() {
 $n = En() <= (i ? 15 : 5)
}

Oh it's great.

(this is part of the scroll workflow for the reply link)

the rabbit hole is deep

I've spent the last 3 days staring at minified SE javascript. So many letters.

do you use jsbeautifier.org ?

No Chrome has a prettify button built in.

Also a trick to find unknown call stacks is to run the performance profiler while you perform an action, then you can spot all the calls.

I just read teh source

where is the prettify button?

ah, the {}

cool, TIL

And it's great because going to function defs from the console links up to the correct location in the prettified source.

I'm super impressed with chrome's dev tools.

@TravisJ Also performance tab to make tracing easier; e.g. when clicking a reply link:

hey @Jason, any particular reason you're using GM_[value FNs] for prefs vs. localstorage?

@JasonC - To be honest, I just prefer reading the source code.

@Shog9 Cross domain storage. I dunno how to do that with local storage without adding other dependencies or kludges.

that's fair

also avoids issues with https

@JasonC - I don't think it is feasible to override the defined scroll function, as it seems to have other implications as well. In addition, it would require including hundreds of lines of code just to have the opportunity to overload the existing functionality.

I guess beta's don't get shuttered anymore either.

They just limp on into oblivion

Can we call it Gamma?

(gamma)

:)

@Guero That's now how beta works..

There is no way BioInformatics is getting shut down. First, it is the fastest growing science exchange. Second, it has the highest rate of questions answered out of all the science exchanges. Third, an exchange will only be shut down if it fails to moderate itself, which judging from the meta activity there is not an issue. Fourth, it is already at 6.6 questions per day, which places it on pace to graduate (10 questions per day) in one of the quicker timelines. Sixth, off by one errors.

It seemingly already has more traction than Economics, just not the coverage yet.

Apparently programmers make bad economists.

Bioinformatics will be fine

It's more like a SE 1.0 -> 2.0 upgrade from an existing community (biostars) than a new one.

Someone should PSA there: "this site will not be shutdown. kthxbai"

@Guero - For ref (if you were thinking of meta'ing that): meta.stackexchange.com/questions/257652/… and meta.stackexchange.com/questions/257614/…

Those two cover the whole "don't worry, you're site is fine" message, with just a hint of "mod it or lose it"

ok, either way

yvette?

I read. I'm a reader. I hear things.

bioinformatics is a cool site, looking at some of its posts

@TravisJ @Shokhet Reply scrolling will be fixed in next update. It actually turned out to be rather straightforward to just reimplement the reply scrolling from scratch.

I watch the dom for new reply-infos, blow away the old click handler, and use a new one that just does the scroll and highlight. Not too bad. Works really well.

But update will have to wait until later tonight I gotta run.

pastebin.com/4Fi4gJBW, in a nutshell.

@JasonC when you do, I got a pull request for you - lets the script work in Firefox.

WHAT

OMG

yeah, that's what you'll say when you read it too

OMG OMG OMG

for different reasons

Haha

Also lets you debug it in Firefox, which honestly is probably the better motivation

Oh man. Thanks. And, I really gotta run I'll check it all out later. Oh boy oh boy oh boy.

source map hints are teh shit

@JasonC Yaay! Thank you :D