WTF Bites
-
The YouTube app on my “Smart” TV (a Panasonic) just crashed for the third time this day. It’s never done that before. I had turned the TV off and on again after the second crash in the hope that’d fix it.
It does ask you to submit a crash report but we all know these go to www.google.com/dev/null.
I have no idea how the TV updates its “apps” or if it even still gets updates at all, but I’m fully assuming Google changed something server sidefor the sake of changeto move fast and break things and from now on it’ll stay broken forever.
I hate computersmodern tech.
-
@topspin my LG smart TV gets app updates occasionally. I know this because it will tell me when there's an update and then it generally reboots. If yours hasn't ever done that then it probably isn't getting updates.
-
The most annoying clickbait.
https://www.youtube.com/watch?v=aBppb2quqkE
First half of this 16 minute video explains what timezones even are. Second half of the video explains what the international date line even is and why it changes. It's only at 12:00, with 2.5 min remaining (because of minute and a half of post-video ads), he even mentions the discrepancy for the first time. And the reason? Dunno, someone tell me in the comments I'll be glad.
14 minutes just to see that the video doesn't actually explain what it claims to be about in the thumbnail. Fucking fuckhead.
-
14 minutes just to see that the video doesn't actually explain what it claims to be about in the thumbnail. Fucking fuckhead.
Don't forget to like and subscribe.
I save time by pronouncing the "fucking fuckheads" judgement ahead of time.
-
@Applied-Mediocrity I'm more of a JIT condescension guy. Saves resources on videos I don't even bother to watch.
-
status: the virtual machine's clock is drifting away from the host's. How the bloody fuck....
-
In the past few days I've seen new videos come up in my subscriptions from channels I don't recognize. First one was a channel whose content actually looked somewhat interesting, and although I didn't recall ever seeing any of the videos on it, it was fine. Today, a new video popped up from a channel simply called "Sounds", which only has that one video. The channel has been around since 2014, so I can only guess that it was at some point a real channel with real content, then (scene missing) and now it's this nearly empty spam farm looking thing with "Scary Sounds for Halloween - 3 Hours"
-
@Tsaukpaetra said in WTF Bites:
How the bloody fuck....
@Tsaukpaetra-certified software running on @Tsaukpaetra-certified hardware.
-
@Zerosquare said in WTF Bites:
@Tsaukpaetra-certified software
Yeah, apparently this is a thing.
I'm sure there's a Very Good Raisin why clocks drift in VMs, but for now I'll have to figure out why the domain controller that's supposed to be in charge of all the clocks isn't asking for the time.
-
If that makes you feel any better, my PC's clock randomly decides to run really slowly, and/or to jump one or more hours back in time. It's a known issue with the mainboard, with no fix (but rebooting usually cures it for a while). It causes annoying side-effects, such as outgoing e-mails replies that appear to have been sent way before what they're replying to. People have asked me whether I was doing that on purpose...
-
@Zerosquare said in WTF Bites:
If that makes you feel any better, my PC's clock randomly decides to run really slowly, and/or to jump one or more hours back in time. It's a known issue with the mainboard, with no fix (but rebooting usually cures it for a while). It causes annoying side-effects, such as outgoing e-mails replies that appear to have been sent way before what they're replying to. People have asked me whether I was doing that on purpose...
My car's AV head unit has this weird (semi-related) thing--when it gets set, it drifts about 4 minutes slow and then stop drifting. Basically, it's always 4 minutes slow except for about one day after setting it. Even when set to manual mode (ie "don't ask anyone for time updates").
to actually figure it out, because I normally have Android Auto running with my phone plugged in, which completely replaces that clock-display UI with the phone time (which is correct).
-
I'm the Committee Chair for my son's Scout Troop. One of my annual responsibilities is renewing the charter. There's usually training, because stuff changes every year. In particular, more and more goes online, which is mostly good. Anywho, today I got an email about this year's training, which is next Tuesday. Which is a bit of a problem because we always have our Committee meetings on the last Tuesday of the month.
1st obvious
...something that looks like a button but is neither a link nor a button:
Second more subtle and less obvious
...a suspicious time:
CST? But Daylight Savings Time doesn't end for more than a month after the meeting! Since registration is currently impossible, I guess it doesn't really matter.
-
@Zerosquare said in WTF Bites:
It causes annoying side-effects, such as outgoing e-mails replies that appear to have been sent way before what they're replying to.
I was under the (apparently mistaken) impression that the server is what determines the timestamps of messages.
Hold on, I'm going to be doing a bit of evil...
-
@boomzilla said in WTF Bites:
I'm the Committee Chair for my son's Scout Troop. One of my annual responsibilities is renewing the charter. There's usually training, because stuff changes every year. In particular, more and more goes online, which is mostly good. Anywho, today I got an email about this year's training, which is next Tuesday. Which is a bit of a problem because we always have our Committee meetings on the last Tuesday of the month.
1st obvious
...something that looks like a button but is neither a link nor a button:Second more subtle and less obvious
...a suspicious time:CST? But Daylight Savings Time doesn't end for more than a month after the meeting! Since registration is currently impossible, I guess it doesn't really matter.
They've sent out a new email with a fixed link. On my confirmation email there was an address to email if you have any questions. I have asked for clarifications on the timing of the webinar.
-
Status: Switched branches in VS 2022, that was fine. Opened a code file with Ctrl+T, that was fine as well. Switched to an already open tab with appsettings.json, VS stopped responding, then wouldn't close until it crashed after a couple minutes
-
And here I thought the Exclusive Website was a joke...
devRant - Need my vaccine receipt to go to a wedding. And I’m in line to get the pdf document. On the one hand this is some cool implementation for a load balancer, on the other hand this so strange…...
Connect with fellow developers over fun rants about tech
I fully expect the next step to be asking for a callback email. Or you could just wait in line for the next available server...
-
@Tsaukpaetra said in WTF Bites:
@Zerosquare said in WTF Bites:
@Tsaukpaetra-certified software
Yeah, apparently this is a thing.
I'm sure there's a Very Good Raisin why clocks drift in VMs, but for now I'll have to figure out why the domain controller that's supposed to be in charge of all the clocks isn't asking for the time.
Is your DC in a VM? I had that issue, where the host would sync to the DC, which would drift, which would cause the host to sync farther off, which caused the DC to sync farther off, which caused the host to sync a little farther off, which...
-
@sloosecannon said in WTF Bites:
@Tsaukpaetra said in WTF Bites:
@Zerosquare said in WTF Bites:
@Tsaukpaetra-certified software
Yeah, apparently this is a thing.
I'm sure there's a Very Good Raisin why clocks drift in VMs, but for now I'll have to figure out why the domain controller that's supposed to be in charge of all the clocks isn't asking for the time.
Is your DC in a VM? I had that issue, where the host would sync to the DC, which would drift, which would cause the host to sync farther off, which caused the DC to sync farther off, which caused the host to sync a little farther off, which...
Yeah, but they don't talk to each other.
Apparently, the system clock is used only at boot to determine the initial clock time, and from there the clock is free to drift wherever it wants, regardless of the host's clock.
Or, at least, that's the observed behaviour.
-
@Tsaukpaetra said in WTF Bites:
Or, at least, that's the observed behaviour.
It's probably the right one, too. Allowing for the possibility that host's clock may not be UTC and VM can be whatever, it's a horrible can of worms. Better have external time source.
-
For the first time someone managed to really fuck up Git. The javascript guy (yeah, predictable) was told to move some code (let's say 'bar' directory) into a submodule, so that it can be shared by another app. What happens now when you try to clone:
$ git clone --recurse-submodules git@initech:foo.git Cloning into 'foo'... $ cd foo $ git submodule status fatal: no submodule mapping found in .gitmodules for path 'src/bar' $ ls -a src/bar/ [nothing] $ touch src/bar/dupa; git status nothing to commit, working tree clean
-
@sebastian-galczynski I wouldn't call it really fucking up Git. He simply failed to commit the
.gitmodulesfile that tells git where to clone the submodule from, which is just a case of forgetting to commit files that happened to people in all version control systems since SCCS.… you should be able to just re-add the module if you know which repo the code went into, but I don't know the details as submodules are the one feature of Git I've been stubbornly avoiding.
-
@Bulb
Yep, I know what happened here, but it's still weird. It looks like a design error in git (but it probably isn't, since Linus is a smart guy).
-
-
@sebastian-galczynski said in WTF Bites:
It looks like a design error in git (but it probably isn't, since Linus is a smart guy).
The problem with initialization of submodules is hard enough that even Linus didn't come with a really good solution. Git records the revision ID it is referencing, which is the right thing to do, but then to initialize the checkout you need a pointer to find that revision. An that pointer can change over time independent of the history, so a file inside the revision isn't actually a reasonable place to store that information. But in the common case the usability ends up better then creating yet another kind of metadata so they added it there anyway.
And similar problem pops up in many places in the domain of configuration management. For example in continuous integration/delivery how you build a software component depends on its version, so putting the build definition in the repository alongside the code makes a lot of sense, but it does not similarly make sense for deployment and testing that depends on the infrastructure nor for things like selecting which branches should be built. So you need some mix, and which parameters should go where is often quite difficult decision.
-
Status: Switched branches in VS 2022, that was fine. Opened a code file with Ctrl+T, that was fine as well. Switched to an already open tab with appsettings.json, VS stopped responding, then wouldn't close until it crashed after a couple minutes
Update: Merged from origin/develop into some branch. That worked fine, commit created and everything. Until a few seconds later when VS crashed
-
Today I was developing a feature to upload/download files to a Dropbox account in a desktop app written in C#. You would think that the Dropbox C# SDK would allow cancellation tokens in case the network was down to allow the user to cancel the operation, since the Dropbox client is on top of HttpClient anyway. But it doesn't, because in the planet the Dropbox C# SDK developers live in there is no such thing as slow internet. So I have to start a thread and use Thread.Abort and hope that it works.
-
Status: Yes, this is a copy of a disk into a smaller disk.
Somehow, it decided that 4th partition is not worth anymore .
-
@Tsaukpaetra said in WTF Bites:
that 4th partition is not worth anymore
It really wasn't, but I kept it anyways. It's just an archive Just In Case something fucks up and I happen to actually need the data.
-
Let's move on to a really shitty topic: pooping in space. Even Elon Musk admits that something has to be done...
Elon Musk Confirms "Challenges" With Toilet on Tourist Spacecraft
SpaceX CEO Elon Musk promised "upgraded toilets" after the Inspiration4 crew experienced "challenges with it this flight."
-
My phone is complaining that it is almost out of memory. Android itself is using ~4GB, and >11 GB is used by other stuff, leaving only a few hundred MB available. But the sum of all the things using that 11 GB, including "other", total to <7 GB. What happened to the other >4 GB?
-
@magnusmaster said in WTF Bites:
in the planet the Dropbox C# SDK developers live in there is no such thing as slow internet
Ah, the good old "works fine on LAN" methodology. Jira must be developed in a similar way.
-
@sebastian-galczynski It's always difficult to find time to test things with proper fault injection. Some time ago I even looked and found an appropriate(ish) tool for the job, the toxiproxy*. But I never got around to configuring it and actually testing something through it.
I am also thinking about creating a podman network definition that would route things through one as a canned bad network to make such tests easier.
*And now searching for the link to include I also found chaos toolkit that I'll like to look at to see where it could be useful.
-
@Bulb It often isn't even about faults or slow bandwidth, but latency. Making a lot of sequential requests works fine, until you move the host 1000km away. This is exactly why the cloud version of Jira is so slow - they make a lot of requests which must be completed in sequence for the page to load. Other frequent example are the N+1 queries generated by ORM misuse, although in that case the db is usually near the app server.
-
@sebastian-galczynski Toxiproxy can do latency, jitter, limited bandwidth, packet loss a a couple of ways of breaking connections, and has interface for the test suite to change them during the test run as desired. All should generally be tried.
-
It's always difficult to find time to test things with proper fault injection.
90% of those problems could be tracked down by a proper code review by someone who knows what they're doing. Unconditionally doing multiple requests to the same backend in quick succession is always a big red flag.
-
someone who knows what they're doing
Unfortunately such ones are in very short supply almost everywhere.
That's why I am thinking about creating a script (let's call it addis-abbeba) that would pull some containers and configure a podman pod with toxiproxy, chromium and/or firefox and xpra for attaching to the browser UI and set the toxiproxy to a quality corresponding to a typical mobile connection is some city in, say, Africa. And then just persuade the product owner to start it and demand that the site be presented to him in that browser.
-
@sebastian-galczynski said in WTF Bites:
Making a lot of sequential requests works fine, until you move the host 1000km away.
So much distance is not needed. Just have the database engine not on your local development computer, but on a different physical machine in the same LAN.
Source: Experience.
-
I never got around to configuring it and actually testing something through it.
Just get a subscription to Milwaukee PC instead. No configuration needed.
-
@BernieTheBernie said in WTF Bites:
@sebastian-galczynski said in WTF Bites:
Making a lot of sequential requests works fine, until you move the host 1000km away.
So much distance is not needed. Just have the database engine not on your local development computer, but on a different physical machine in the same LAN.
Source: Experience.Yeah, we have all of our development DBs in our datacenter dev/test enclave. Doing stuff as a developer when the app server is on your local machine, talking over the VPN, vs the test environment where the app server is on a VM on the same machine (or one sitting right next to it) is a massive difference. So doing stuff in development is almost always a worse experience than test, let alone production. Which is the way it should be, of course.
-
-
someone who knows what they're doing
Unfortunately such ones are in very short supply almost everywhere.
That's why I am thinking about creating a script (let's call it addis-abbeba) that would pull some containers and configure a podman pod with toxiproxy, chromium and/or firefox and xpra for attaching to the browser UI and set the toxiproxy to a quality corresponding to a typical mobile connection is some city in, say, Africa. And then just persuade the product owner to start it and demand that the site be presented to him in that browser.
Thanks for the idea! Convention Internet is pretty equivalent and using Chrome's tools has proven insufficient...
-
So I looked at TVs the other day.
WTF 1: almost all of them had an F or G rating for energy efficiency, except for one that had an A. I assume that one used an older scale and they upgraded it in the meantime. Which would make sense, since for other things it was getting ridiculous that the bad ones had B or A ratings because the good ones were at A+++. However, if I’m correct with that assumption, I didn’t see anything on the label hinting to that. Nothing that would mark the “version” of the scale used on it, which makes it useless.
(And if I’m incorrect then all TVs except that one are a WTF)WTF 2: The power consumption is rated at 129 kWh / 1000h. Um, so you mean 129 W. Both the k and the h cancel out. What kind of retarded unit is that?!
WTF 3: What’s that right below? There it says 217 kWh / 1000 h. That’s 70% more! Which one is it?!
-
Based on the logo, I'd guess it uses more power if you enable HDR mode?
-
@Zerosquare said in WTF Bites:
Based on the logo, I'd guess it uses more power if you enable HDR mode?
The big question is ... does it get an A if you turn it off / put it into sleep mode?
-
I assume that one used an older scale
Yes, the scale for TVs changed some point earlier this year.
@Zerosquare said in WTF Bites:
I'd guess it uses more power if you enable HDR mode?
Yeah
-
@Zerosquare said in WTF Bites:
Based on the logo, I'd guess it uses more power if you enable HDR mode?
What the heck does it do to use 70% more power in HDR mode?!
I’ll have to remember never to use that.
-
@topspin brighter colours use more power - although a 70% increase is signifnicantly worse than typical, so maybe just don't buy a TV where the difference is that much. It's probably not the only compromise.
-
WTF of my day: So I tried my hand at Mechwarrior 5: Mercenaries again with the new DLC. Started a new campaign, got through the tutorial missions and selected a nice, easy "Go there and wreck shit" mission.
Was a bit irritated by the fact that I heard the distinctive sound of artillery shells flying over and impacting the dropship while my 'mech was still starting up. Bay doors opened, I walked outside ... and promptly received an artillery shell right to the face.
Turns out that artillery in this game after the patch seems to have gone to the "How not to shoot like Stormtroopers" workshop and are now dropping shells like any rational human would do it, given the capability: One in front of your mech (so if you don't stop you'll eat that one), one behind your mech (can't have you walking backwards now, can we?), one to the left (yeah, don't think about changing directions!), one to the right (yeah, we did not forget that right is to the opposite of left!) and the pièce de résistance - a final shell on top of your current position.
Oh, and if you do manage to avoid one salvo of artillery, one of your lance mates will surely be so friendly as to catch the one you missed.
The artillery is also located on the other side of the map and does not need spotters.
That was 30% of my armor gone before I was able to walk further than half a kilometer. Fuck that shit.
And don't talk to me about "realism" - because realism would mean that pretty much 99% of all conflicts in that universe would be solved by Tungsten rods dropped from 100 km above.
-
The power consumption is rated at 129 kWh / 1000h. Um, so you mean 129 W. Both the k and the h cancel out. What kind of retarded unit is that?!
Kilowatts-hour is the popular unit for electricity consumption.
I hope you're not expecting your regular consumer to do unit simplification in their head?
-
What kind of retarded unit is that?!
Only marginally more retarded than kWh. Using the pound-farce as the unit of retardation, I’d say it’s about 850 mlbf.
edit – some consideration later: it’s claimed that kWh are generally easier to work with (despite a suspicion that approximately nobody is regularly multiplying power consumption by hours of use).
But given the label under discussion, and the number of times I’ve seen people (yanks) write it as “kilowatts per hour” … clearly there’s an appealing comprehensibility to rates expressed with reciprocal time. (meanwhile, amperes are an SI base unit
)So, presumably we should really be metering energy with megajoules, and then power consumption would be listed in MJ/h, and everyone could be happy?
edit2: at least 1 kWh = 3.6 MJ is a saner equivalence than 1 MJ/h = 277.777777778 W…
