WTF Bites
-
@polygeekery said in WTF Bites:
Here's what happens when sparkies try to do network cabling:
Fucking sparkies. That join of what appears to be a black stranded patch cable to the blue CAT5E is a meathook abortion.
There are a fair number of Ethernet connections in my building that are actually a 2-pair connection that's sharing the 4-pair cable with another 2-pair connection.
(Note: that works for 10 or 100 Mbit, but Gigabit Ethernet uses all 4 pairs...)
-
@polygeekery It passes a continuity test and that's all that matters, right?
The reason I got called in was because he couldn't find where it was in the patch panel. No time at the other end. As soon as I plugged my toner in I checked for a tone at the toner and nothing. That splice job shorted out the toner and killed its signal.
-
@timebandit said in WTF Bites:
@polygeekery If you're not using POE, Ethernet only uses 4 wires anyway, just patch the working ones
POE or NOE is
-
-
-
@pie_flavor said in WTF Bites:
And just like that, all Mac gaming died.
As opposed to a week ago when it looked like this:
-
@blakeyrat It'd be far more accurate to say that it looked like this:
https://i.imgur.com/tzES0Wg.png
Since, y'know, that's all people play on Mac.
-
My company recently changed to different VPN software. Now, every time you go to login you get a page like this:
So you
click here
and then it takes you to the login page. Why?!? Just take me there to begin with!
-
@boomzilla Because people apparently can't configure F5s even though the Web-based interface is pretty damn intuitive, at least to me. I was able to set up a relatively complex re-encryption scenario in a test lab in a couple hours having never seen an F5 before. My guess is people are fooling around in the shell more than they should be and breaking stuff that's hard to find or fix.
-
@boomzilla It's a conspiracy by Big IP
-
@boomzilla said in WTF Bites:
My company recently changed to different VPN software. Now, every time you go to login you get a page like this:
So you
click here
and then it takes you to the login page. Why?!? Just take me there to begin with!Is "click here" a unique link every time, or could you just update your "login" bookmark to point there?
I use a site that leaves a zombie session if you don't "log out" properly. Next time you visit the login page, it says you're already logged in, but following the link to continue says that your session's no longer valid and you need to log in again. So I changed my bookmark to the log out link, which terminates any zombie session I had before it redirects me to the login page.
-
@anotherusername said in WTF Bites:
Is "click here" a unique link every time, or could you just update your "login" bookmark to point there?
I think it sees an old cookie in my browser and takes me there regardless.
@anotherusername said in WTF Bites:
I use a site that leaves a zombie session if you don't "log out" properly. Next time you visit the login page, it says you're already logged in, but following the link to continue says that your session's no longer valid and you need to log in again. So I changed my bookmark to the log out link, which terminates any zombie session I had before it redirects me to the login page.
I'm not aware of a logout link like that for this.
-
@boomzilla well, there's always private mode... then you know there's no old session to worry about.
edit: although, I wonder if it's complaining that there isn't any session. The text certainly seems to indicate this. That would be pretty dumb.
-
@anotherusername said in WTF Bites:
edit: although, I wonder if it's complaining that there isn't any session. The text certainly seems to indicate this. That would be pretty dumb.
Yes, it's complaining that there's an invalid/missing session, when given the page you accessed, it expects there should have been a valid one by that point.
-
@pie_flavor What game is this?
-
@pie_flavor What game is this?
It looks like the health bar from Minecraft, but I'm not sure why the hearts are different sizes.
-
@anotherusername said in WTF Bites:
I'm not sure why the hearts are different sizes.
#AllLivesMatter
-
@anotherusername When you're near death, your health bar wiggles as a visual alert that you should get the hell out of dodge.
-
@anotherusername said in WTF Bites:
It looks like the health bar from Minecraft, but I'm not sure why the hearts are different sizes.
They're not different sizes but at different heights because they bounce.
-
@heterodox, @TwelveBaud and @anotherusername thanks.
-
@pie_flavor Why are the hearts mis-aligned?
-
@topspin Read the posts right above yours.
-
Changelog snippet:
- Fixed a crash on the Georgia Peach Logo when launching Paladins for players in India.
Wha... Ho... You know what, I don't want to know...
-
Changelog snippet:
- Fixed a crash on the Georgia Peach Logo when launching Paladins for players in India.
Wha... Ho... You know what, I don't want to know...
Unexpected success dereferencing a non-null pointer?
-
Today's WTF Bite is NodeBB. Not sure how long it's been like this, but any time the NodeBB tab is visible and I'm viewing the topics list, it's hogging 50% of a CPU core. See, you can tell when I switched to the NodeBB tab, vs. when I switched to the GMail tab:
That's not even doing anything -- just sitting there, on "recent". All I did was click the tab to bring it to the front.
The reason appears to be that it's doing something which triggers repaints. Constantly.
Note that it doesn't seem to happen inside a topic (and I haven't tested it elsewhere). It's something about the topics list that does it.
edit: this seems specific to "Unread" and "Recent". The categories list doesn't cause the CPU to spike, nor does viewing the topic list for a single category.
-
@anotherusername said in WTF Bites:
it's doing something
Does it show you what it's painting? Or calling? Or something? I don't know dev tools for Firefox (or Chrome even), but you'd think that would be something that said tool would help you find out...
-
@tsaukpaetra well, there's a
MozAfterPaint
event that's supposed to give the bounding client rectangles that were painted (oncedom.send_after_paint_to_content
is enabled inabout:config
; by default it's only available to extensions), but I couldn't access theclientRects
property -- despite the console showing that it had one:
Expanding the object doesn't even showclientRects
...
And now that I've restarted the browser, it's not doing the repaints at all. Despite the fact that it was doing it on two separate computers earlier today. Maybe there was an update in the last ~hour~ that fixed it.
-
@polygeekery Whoever did that in the first place needs to be fired out of a cannon aimed toward the ocean.
-
@brisingraerowing said in WTF Bites:
@polygeekery Whoever did that in the first place needs to be fired out of a cannon aimed toward the ocean.
Sparkies should just stay away from data cabling. Six or months ago we had to go in and recable a brand new office space because the sparkies fucked it up so badly. Cables pinched, kinked and torn. One section had the insulation ripped off because they tried to pull it like Romex and it scuffed on ductwork. Probably 8-10' with no insulation and the twisted pairs laying in disarray. Flying leads left hanging out of the keystone jacks and not trimmed.
Years ago we had to recable another new office space because sparkies had wired network drops like outlets. They daisy chained them. That one took longer than I would like to admit to figure out the problem on. Most of that time was because we never even considered anyone could be so stupid.
This was the first time I have seen telephone splices used on CAT5 though.
Fucking sparkies.
Although, they have made me a lot of money over the years. Keep up the bad work. Keep undercutting us by 50% and I will keep charging more to come fix your inevitable fuckups. Why do it right when you can pay to have it done twice?
-
@anotherusername Now it's doing it again.
-
@pie_flavor said in WTF Bites:
@blakeyrat It'd be far more accurate to say that it looked like this:
https://i.imgur.com/tzES0Wg.png
Since, y'know, that's all people play on Mac.@Joel continuously told me that he can't play GW2 because he uses a Mac and would have to reboot to Windows.
Now that there's an official Mac client, he had to come up with a different excuse.
-
@anotherusername I think maybe it has to do with the topic/post timestamps.
I set up a
MutationObserver
to see what was changing. It logged a whole bunch of mutation events. Here's one:
Of note are the
removedNodes
andaddedNodes
. That event was caused by removing a textabout 4 hours ago
node and then adding a textabout 4 hours ago
node. In other words, it changed nothing.
-
@anotherusername If only there was some way to represent that timestamp that wouldn't have to be updated relative to the current time.
-
@hungrier If only there was some way to compare two strings to see if they're identical and if so, not update the DOM with an identical text node.
-
@anotherusername String comparison? Sounds expensive, better re-render the whole page just to be safe.
-
Blame where blame is due, that should be timeago.js's fault, not NodeBB's.
Edit: assuming timeago isn't being misconfigured. I have no idea.
-
@zecc They made their bed, they have to lie in it.
When you decide to use a library in your product, you also decide to include all its misfeatures and bugs.
-
Blame where blame is due, that should be timeago.js's fault, not NodeBB's.
Edit: assuming timeago isn't being misconfigured. I have no idea.
Well, it only seems to be doing it on
/recent
and/unread
. And only sometimes?
-
@anotherusername ok, it's definitely not that. I'm having constant repaint events even when the timestamps aren't being updated (which appears to be set on an interval).
-
@anotherusername I figured it out. It's 's damn avatar! It's triggering constant repaints even when it's not visible on the screen.
Now the question is why Firefox is chugging 50% of a CPU to render a fairly small animated GIF, and why it's doing so even when the GIF is scrolled off the screen.
-
@anotherusername said in WTF Bites:
Now the question is why Firefox is chugging 50% of a CPU to render a fairly small animated GIF, and why it's doing so even when the GIF is scrolled off the screen.
Mine's sitting at around 5% (actually less, that an accumulation for all the firefox processes) according to taskmgr.
-
@dcon well since I have 4 CPU cores, 50% of one core is ~12%. But still... I tried firing up Chrome and it was running at around 5% for the same page.
edit: then again, I'm not sure that image is the whole explanation...
-
@anotherusername said in WTF Bites:
why Firefox is chugging 50% of a CPU
Oh yeah, now I see it too in Chrome...
Wow....
-
@anotherusername String comparison? Sounds expensive, better re-render the whole page just to be safe.
I hear there's an ember.js update coming next month that solves all of this.
-
@anotherusername said in WTF Bites:
That event was caused by removing a text about 4 hours ago node and then adding a text about 4 hours ago node. In other words, it changed nothing.
That's not fair. The first time it meant 'slightly more than 4 hours ago' and it updated to mean 'almost 5 hours ago'.
-
public abstract class Component<T extends Component> public static abstract class Builder<T extends Component, B extends Builder<T, B>> public class Type<T extends Component, B extends Component.Builder<T, B>> extends Component<Type<T, B>> public static class Builder<T extends Component, B extends Component.Builder<T, B>> extends Component.Builder<Type<T, B>, Builder<T, B>>
elsewhere:
public interface CollectionValue<E, C extends Collection<E>, V extends CollectionValue<E, C, V, I>, I extends ImmutableCollectionValue<E, C, I, V>> extends Value<C>, Iterable<E> {
-
Quote from a lecture in my Software Engineering Basics class:
Thorough testing is impossible
Testing everything (all combinations of inputs and initial conditions) is possible only in trivial cases. You must use risk analysis and prioritization.
Pinging @HardwareGeek.
-
@gąska I don't see what's wrong. I might just replace thorough with exhaustive.
-
@gąska I don't see what's wrong. I might just replace thorough with exhaustive.
What's wrong is that it's entirely possible to test everything for every case - it's just that programmers are and CBA to write proper tests and employ techniques that reduce number of tests necessary via static analysis. It's only impossible because we don't even try. Hardware guys do try, and they have appropriate methodologies all figured out so they can actually guarantee that no set of inputs will ever break their chip.
-
@gąska to exhaustively black-box test even a super simple function like
bool is_odd(uint64_t)
you need to test 2^64 = 18,446,744,073,709,551,616 inputs. That is simply not feasible and doesn't scale to anything more complex.