WTF Bites



  • @blakeyrat it is non-standard, but there's no warning box. And you did say that literally all of the non-standard ones had a warning box.



  • @anotherusername Oh good job, you win a trophy for proving me wrong. I'll just go seppuku now.


  • Considered Harmful

    @vault_dweller said in WTF Bites:

    Pop quiz. What is the result of the following query in T-SQL:

    select ISDATE('2013-12-006')
    

    FILE_NOT_FOUND?


  • 🚽 Regular

    Regarding Navigator.oscpu:

    There's Bug 1120892 in Mozilla's Bugzilla: "Consider removing or standardizing navigator.oscpu". Last comment in there says it was submitted for standardization and links to Bug 27820 in W3's Bugzilla, where it sits with status "RESOLVED MOVED", whatever that means.

    In any case, there's a note in https://html.spec.whatwg.org/#client-identification :

    If the navigator compatibility mode is Gecko, then the user agent must also support the following partial interface:
    partial interface NavigatorID {
      [Exposed=Window] boolean taintEnabled(); // constant false
      [Exposed=Window] readonly attribute DOMString oscpu;
    };
    

    (the link "navigator compatibility mode" links to that very same paragraph)

    So, it's standardized, it seems.


  • 🚽 Regular

    @pie_flavor said in WTF Bites:

    I love it when IntelliJ freezes.

    [screenshot of floating tooltip]

    Heh. I've seen that effect with Visual Studio too, even without it freezing.


  • Considered Harmful

    No, it's not Yet Another April 1st RFC:


  • 🚽 Regular

    Satellite Internet (SI) Specification says:

    SI works by using a space network of satellites that surrounds the whole earth, connected to each others using fiber optic cables

    I dunno, sounds legit.



  • @laoc said in WTF Bites:

    No, it's not Yet Another April 1st RFC:

    No it is not. It is much worse. You could actually implement IP over Avian Carriers, but you could not implement this as it is in direct violation of laws of physics, specifically orbital mechanics.

    @zecc said in WTF Bites:

    Satellite Internet (SI) Specification says:

    SI works by using a space network of satellites that surrounds the whole earth, connected to each others using fiber optic cables

    I dunno, sounds legit.

    Well, having satellites connected with fiber optic cables would be possible. It would be somewhat tricky to extend them, because the cable you send forward would tend to raise and fall back and the one you send back would tend to descend and race ahead, but it should be possible. However you could only do it for satellites on the same orbit. And probably couldn't be the geostationary one, because that one is already pretty crowded so you'd risk hitting another satellite with the cable.


  • Considered Harmful

    @bulb said in WTF Bites:

    Well, having satellites connected with fiber optic cables would be possible. It would be somewhat tricky to extend them, because the cable you send forward would tend to raise and fall back and the one you send back would tend to descend and race ahead, but it should be possible. However you could only do it for satellites on the same orbit. And probably couldn't be the geostationary one, because that one is already pretty crowded so you'd risk hitting another satellite with the cable.

    Not to forget the fact that once around the geostationary orbit is roughly a quarter million kilometers.

    I just found the dude is also the author of the IPv10 spec.



  • *le sigh*

    gfortran: error: unrecognized command line option ‘-fno-fpe-trap=’; did you mean
    ‘-fno-fpe-trap=’?

    Yes. Yes, I suppose I did.

    (The problem is that I have a global -ffpe-trap=invalid,zero,stuff, but want to reduce to just -ffpe-trap=stuff for a single file. I was hoping I could override the first -ffpe-trap with a later one, but that didn't pan out. Neither did the subsequent attempts. I gave up when I got the above message.)


  • 🚽 Regular

    @cvi Reminds me of when I once got an error like (paraphrasing as I don't remember and CBA to check) "Expected QLabel instance, got QLabel instance instead."

    Turns out I was importing PySide classes in one script and PyQt4 classes in another.



  • @laoc said in WTF Bites:

    IPv10 spec

    There is only one possible answer to that:

    xkcd:927


  • Discourse touched me in a no-no place

    @zecc said in WTF Bites:

    @cvi Reminds me of when I once got an error like (paraphrasing as I don't remember and CBA to check) "Expected QLabel instance, got QLabel instance instead."

    Turns out I was importing PySide classes in one script and PyQt4 classes in another.

    That in turn reminds me of the kinds of “fun” you can get with complicated class loader hacking in Java.


  • Considered Harmful

    @bulb said in WTF Bites:

    @laoc said in WTF Bites:

    IPv10 spec

    There is only one possible answer to that:

    You mean that's the standard answer? I propose adding another standard answer: :facepalm:



  • @blakeyrat said in WTF Bites:

    @anotherusername Oh good job, you win a trophy for proving me wrong. I'll just go seppuku now.

    If you didn't mean literally, you shouldn't have said "literally". 🤷♂



  • @zecc said in WTF Bites:

    Regarding Navigator.oscpu:

    There's Bug 1120892 in Mozilla's Bugzilla: "Consider removing or standardizing navigator.oscpu". Last comment in there says it was submitted for standardization and links to Bug 27820 in W3's Bugzilla, where it sits with status "RESOLVED MOVED", whatever that means.

    In any case, there's a note in https://html.spec.whatwg.org/#client-identification :

    If the navigator compatibility mode is Gecko, then the user agent must also support the following partial interface:
    partial interface NavigatorID {
      [Exposed=Window] boolean taintEnabled(); // constant false
      [Exposed=Window] readonly attribute DOMString oscpu;
    };
    

    (the link "navigator compatibility mode" links to that very same paragraph)

    So, it's standardized, it seems.

    I suspect it falls into a separate category, "not standardized but on track to become standardized"... the warning box that I saw on other properties said, "This feature is non-standard and is not on a standards track."


  • ♿ (Parody)

    @anotherusername said in WTF Bites:

    @blakeyrat said in WTF Bites:

    @anotherusername Oh good job, you win a trophy for proving me wrong. I'll just go seppuku now.

    If you didn't mean literally, you shouldn't have said "literally". 🤷♂

    Just call him a moron idiot and move on.



  • I come home from work to play guitar, not look at more yellow screens of death...

    0_1510344065848_d42e761d-3037-412e-8981-dafc605cac87-image.png



  • @maciejasjmj If you've played for centuries, you have too much time on your hands 🤷🏼♂


  • Considered Harmful

    Update on the Flayer's Bell:

    The numbers in the post counter no longer unhighlight when you scroll. I'm not sure whether this is indicative of the highlight handling being better, or worse.

    Edit: Actually, a lot of the weirdness has changed. Highlighting the bell tooltip now highlights all the tooltips, and highlighting the chat tooltip will also highlight the search tooltip (but not the bell tooltip) (and not vice versa). And any menu with a highlighted tooltip now has its contents highlighted - no more mouseover weirdness. Someone seems to have drained the bell of its powers.



  • Go home machine learning, you're drunk.
    0_1510458218953_3dbd968a-79c7-4844-ac6f-68f960f65f20-image.png



  • @lb_ said in WTF Bites:

    Go home machine learning, you're drunk.
    0_1510458218953_3dbd968a-79c7-4844-ac6f-68f960f65f20-image.png

    8 minutes remaining? How can it possibly know what the video is about from the first five seconds?


  • Considered Harmful

    Speaking of 8 minutes remaining:

    This, in real-time, is how long it takes for my Downloads folder to arrange itself chronologically instead of alphabetically. Once per startup. There's only 365 items in the folder, including folders themselves which it ignores anyway.



  • 0_1510482884842_14627e8b-a066-4db8-8396-9ad04b54d74d-image.png

    The worst Windows API ever made: Event Tracing

    Pretty epic bad design.



  • This deprecation/non-standard thing reminds me of Microsoft's move to rename all functions of its C Run-Time Library that aren't part of the C standard, e.g. from open() to _open(). Can't say the intent was bad (though I at least hope it's disabled when compiling for WSUS), but the actual job was a bit heavy-handed:

    All such functions get the same warning, that says The POSIX name for this function is deprecated. Use (underscored name) instead, even for functions that are not, and have never been, part of the POSIX standard (such as getch()). And yes, they're all deprecated with the same macro, _CRT_NONSTDC_DEPRECATE, which means for Microsoft, everything in their CRT that's not part of the C standard is POSIX.



  • @ben_lubar said in WTF Bites:

    @lb_ said in WTF Bites:

    Go home machine learning, you're drunk.
    0_1510458218953_3dbd968a-79c7-4844-ac6f-68f960f65f20-image.png

    8 minutes remaining? How can it possibly know what the video is about from the first five seconds?

    Apparently the title and tags also count, not just the video content and auto-translated subtitles. But in my experience it usually declares me innocent after a day or two without any intervention from me, so maybe their algorithms assume guilty until proven innocent or they have been trained that new videos are bad and old videos are good. Who knows.


  • Notification Spam Recipient

    @pie_flavor said in WTF Bites:

    Once per startup. There's only 365 items in the folder, including folders themselves which it ignores anyway.

    0_1510542969065_6535aa1c-e4cc-4c64-8433-0792dcaa278b-image.png

    Vid

    On an SSD, after waiting two minutes for the first time.

    I guess checking all those thumbs.db is a tough thing to do.


  • Notification Spam Recipient

    @lb_ said in WTF Bites:

    so maybe their algorithms assume guilty until proven innocent or they have been trained that new videos are bad and old videos are good. Who knows.

    The first. Definitely the first.

    Did you know you can get your channel deleted instantly if three different people merely flag three videos for copyright?



  • @tsaukpaetra the Downloads folder always takes a very long time to load on old installs even if there's just a single file in it. Something about the constant addition and removal of files over time seems to wreak havoc on some internal processes.


  • Notification Spam Recipient

    @lb_ said in WTF Bites:

    @tsaukpaetra the Downloads folder always takes a very long time to load on old installs even if there's just a single file in it. Something about the constant addition and removal of files over time seems to wreak havoc on some internal processes.

    Internal pro---- what?? WTF does that even mean? What "internal process" could possibly be associated with the equivalent of dir *.*?


  • Notification Spam Recipient

    Status: Chrome, go home. You're drunk.

    0_1510548960318_3f0054b3-eea7-4b6a-b6cf-0bcb65dd097b-image.png



  • @tsaukpaetra said in WTF Bites:

    @lb_ said in WTF Bites:

    @tsaukpaetra the Downloads folder always takes a very long time to load on old installs even if there's just a single file in it. Something about the constant addition and removal of files over time seems to wreak havoc on some internal processes.

    Internal pro---- what?? WTF does that even mean? What "internal process" could possibly be associated with the equivalent of dir *.*?

    On FAT (and VFAT), the deleted entries are still there, just with the first character replaced by some marker (don't remember what it was). So if you list a directory where many files were deleted in the past, the system still may have to load and process a large chunk of data. Composed with the notorious slowness of file listing operation in Windows this might get noticeable.

    I don't know how much better NTFS is in this regard, but it may still hit some pessimal case in the directory structure in this case.


  • Notification Spam Recipient

    @bulb said in WTF Bites:

    @tsaukpaetra said in WTF Bites:

    @lb_ said in WTF Bites:

    @tsaukpaetra the Downloads folder always takes a very long time to load on old installs even if there's just a single file in it. Something about the constant addition and removal of files over time seems to wreak havoc on some internal processes.

    Internal pro---- what?? WTF does that even mean? What "internal process" could possibly be associated with the equivalent of dir *.*?

    On FAT (and VFAT), the deleted entries are still there, just with the first character replaced by some marker (don't remember what it was). So if you list a directory where many files were deleted in the past, the system still may have to load and process a large chunk of data. Composed with the notorious slowness of file listing operation in Windows this might get noticeable.

    I don't know how much better NTFS is in this regard, but it may still hit some pessimal case in the directory structure in this case.

    Nah. Turns out Windows decided that the downloads folder was Pictures, and required fetching thumbnails for each and every file. For executables this takes a while (because raisins) so the more you have the longer it takes for those, and then whatever else shenanigans to try producing icons for everything else.

    Switching it back to "general" fixed it. But that's another WTF to think about for later.



  • Our build server just had another of its hiccups. This time it somehow disconnected from the network share where it is uploading the build products and logs. Other times it gets another hiccups, but often also network-related. Usually a couple a day. And that's supposed to be a highly reliable server farm…


  • ♿ (Parody)

    @cartman82 said in WTF Bites:

    0_1510482884842_14627e8b-a066-4db8-8396-9ad04b54d74d-image.png

    The worst Windows API ever made: Event Tracing

    Pretty epic bad design.

    Bonus :wtf: - the extra braces around the for loop:

    // At the beginning of the program
    etw_event_trace Trace = ETWBeginTrace();
    ETWAddEventType(Trace, ETWType_ContextSwitch);
    
    // On each frame
    event EventBuffer[4096];
    int EventCount;
    while(EventCount = ETWGetEvents(Trace, sizeof(EventBuffer), EventBuffer))
    {
        {for(int EventIndex = 0;
            EventIndex < EventCount;
            ++EventIndex)
        {
            assert(EventBuffer[EventIndex].Type == ETWType_ContextSwitch);
            // process EventBuffer[EventIndex].ContextSwitch here
        }}
    }
    
    // At the end of the program
    ETWEndTrace(Trace);
    

    Also, that gave me some serious Win32 flashbacks.


  • 🚽 Regular

    @boomzilla Wasn't there a time when a variable declared in C++ inside a for loop was still valid after it?

    Maybe that's a habit gained from back then?


  • ♿ (Parody)

    @zecc I don't know, but that's a plausible explanation.



  • @zecc That's a thing that happened in C, never in C++ as far as I'm aware.



  • @zecc said in WTF Bites:

    @boomzilla Wasn't there a time when a variable declared in C++ inside a for loop was still valid after it?

    Maybe that's a habit gained from back then?

    Are you talking about declared in the for statement itself? Or the for-block? For the first, yes. Hence /Zc:forScope (default used to be no. It's now yes. And in VS2015: The /Zc:forScope- option is deprecated and will be removed in a future release.)



  • @lb_ said in WTF Bites:

    @tsaukpaetra the Downloads folder always takes a very long time to load on old installs even if there's just a single file in it.

    ... no? The Downloads folder isn't even special in any way whatsoever, except having an API call to retrieve its path.

    @tsaukpaetra said in WTF Bites:

    Internal pro---- what?? WTF does that even mean? What "internal process" could possibly be associated with the equivalent of dir .?

    I think the "internal process" of LB_'s brain is screwed up.



  • @lb_ said in WTF Bites:

    @zecc That's a thing that happened in C, never in C++ as far as I'm aware.

    What @dcon said. VisualStudio used to misbehave w.r.t. this. You can occasionally find the following workaround for the bad scoping in old C++ code bases:

    #define for if(0) {} else for
    


  • @blakeyrat said in WTF Bites:

    @lb_ said in WTF Bites:

    @tsaukpaetra the Downloads folder always takes a very long time to load on old installs even if there's just a single file in it.

    ... no? The Downloads folder isn't even special in any way whatsoever, except having an API call to retrieve its path.

    The Downloads folder is social in that it's where new downloads go for most modern browsers and that you usually delete downloaded files you don't need anymore, but you never delete the folder itself. I have consistently noticed slower performance after cycling many files through it, I assume it would happen to any directory but it just happens to be that the Downloads folder is the most noticeable example.

    If you actually Google the issue, you'll see that it's a widely known problem that many people experience, and it seems there's lots of different causes and solutions.

    @tsaukpaetra said in WTF Bites:

    Internal pro---- what?? WTF does that even mean? What "internal process" could possibly be associated with the equivalent of dir .?

    I think the "internal process" of LB_'s brain is screwed up.

    So you and @Tsaukpaetra are saying that there are no internal processes in an operating system? What is the Windows API then? Pixie dust? The word "process" can be used in multiple ways, you know. Thumbnail caching is an internal process from my perspective, for example.



  • You know how the Juicero was too expensive because it was overengineered?

    This is what happens when you under-engineer your devices instead:

    Review of the ION Bottleless Water Cooler – 02:17
    — Frank May


  • Notification Spam Recipient

    @lb_ said in WTF Bites:

    So you and @Tsaukpaetra are saying that there are no internal processes in an operating system?

    I don't recall saying anything of the sort.

    @lb_ said in WTF Bites:

    I assume it would happen to any directory

    I create and destroy billions of files in %temp% every day, yet going to it doesn't exhibit this problem.

    @lb_ said in WTF Bites:

    Thumbnail caching is an internal process from my perspective, for example.

    Is related to the actual problem, IMO, see above.


  • Garbage Person

    Had a nice dinner with the missus. They gave us a copy of the menu to take home:

    0_1510642871556_Untitled.png


  • 🚽 Regular

    @greybeard Lorem Tiramisum?


  • Java Dev

    Blizzard has a lot of support articles! Although I can't read any of them...

    0_1510651261094_blizz-support.PNG



  • @atazhaia 100 results, divided by zero results per page, equals MAX_INT pages.

    Computer math checks out!



  • @Tsaukpaetra said in WTF Bites:

    @PJH said in WTF Bites:

    getting corrupted.

    Speaking of corruption, I'm surfing the data on the company NAS and finding that a not insignificant portion of the files are essentialy files filled with 1 bytes.

    Unary encoding?

    @Tsaukpaetra said in WTF Bites:

    The AntiVirus on ReadyNAS doesn't give you that option, and instead, when it detects a "threat" it (apparently) just silently replaces all the bytes in the file with 0x01, popping an alert into the notifications menu that very few probably bother to read until the next email update (if that's even set up).

    Or that.



  • @anonymous234 said in WTF Bites:

    @atazhaia 100 results, divided by zero results per page, equals MAX_INT pages.

    Computer math checks out!

    Is it 100 results? Or is it more than 100 results?


Log in to reply