That’s not good at all. 🙁

Fortunately it's getting less entrenched in a lot of places.

Turns out excluding more than half the population ends up being bad for business, among other problems. 😉

How is your game dev going?

@DMGregory Those are some great names and works to check out.

Celeste is awesome.

Going alright. :) We just branched for first submission.

And at home, I'm prototyping some rhythm game mechanics I might try using at a game jam next weekend.

I'm a bit sad that Flash is effectively dead now. I loved showing Anna Anthropy's Dys4ia to my students as a concise example of using gameplay mechanics to convey meaning.

I imagine I would have liked that.

I wonder if I can find some playable version of it.

I've got it as an .swf, if you have a standalone flash player to run it.

When showing it to my students, I do give content warnings for misgendering, transphobia, nudity and blood (just a little, very pixelated).

hmm I have Adobe Flash but not sure the last time I ran it.

Did the EOL make it unusable?

Nope, just doesn't run in browsers anymore.

sweet.

Wait... somehow this version appears to still be playable.

freegames.org/dys4ia

yup. I just found that

Looks like it might be a slightly older/glitchier version though... got stuck on one of the last vignettes...

the bricks?

Yeah

same

@DMGregory I just played “Happy Wheels” a lot of the time.

But they created a new JavaScript enabled version, so I still play.

It's a shame. The ending is quite good.

Morning!

Good morning. :)

Gonna get my 5G chip tomorrow!

First one?

Yep.

Let's hope they forget to turn on the side effects.

Good luck! I've still got a two week wait till my first.

Yes; well I've also ordered an suit made of asbestos to block the signal.

@DMGregory Yeah, we've been "lucky" here. A tip I heard was to look often if there were new spots opening up.

Yeah, Toronto's been pretty swamped, so nothing in my area has had capacity for under-40 until very recently.

Yeah, I see. For us, since we live in "the suburbs", we can chose spots either in the city, or on the south shore, giving us more options.

Welp. This code works on my machine but not on coliru...

Yep ;)

Hmm, hold on.

Apparently, I am.

Hmm, well I am in my test app. I'm not in the main app, and the language standard is also not set.. (???)

Looks like I have a couple of things to fix there.

Ah, the c++17 flag is set "manually"

I suppose I should make it working in VS and on coliru, not sure I'd like this to bite me in the ass when I upgrade to vs 2019 (or 21/22)..?

Ah, yes I could try that one to see if it would compile under a more recent version of VS!

Turns out, I used std::apply to iterate over the tuple elements

template<typename T>
T fromJson( const Json& data ) {
  T object;

  // Thanks to stackoverflow.com/a/54053084
  std::apply(
      [&object, &data]( auto&&... args ) {
        ( recursivelySetData( object, args, data ) , ... ); }
    , T::properties
    );

  return object;
}

Yes, much more simple that other options!

The issue I seem to be having with g++ is that it will generate a templated function even if it's not needed..

(It will "raise" the static_assert( false...) that there is in it. )

/// This is the "model" template that is used to extract a JSON value into a
/// value usable by our object. If it is invoked, it will raise an error at
/// compile time as all used specializations should be defined.
template<typename T>
T asValue( const Json& data )
{
  static_assert(
      false,
    "unknown type. make sure a specialization for asvalue for this type has "
      "been defined." );

  return t();
}

/// This is the template specialization for T == double.
template<>
double asValue( const Json& data )

arrr

hold on

fixed

If I am to call asValue<int>(...) in vc++, I'll get the static_assert, presumably because the function will be generated by the compiler. If I don't call asValue<int>(...), I don't get the assert. This works fine in vc++, but it complains in g++ whether I call the function or not.

Hmm, still the same. An option would be to declare the first template, but not define it. That would make a linker error.

Nice!

I'll see if I can hack this with template stuff, or use the linker approach (which is much less clear).

So here the idea is that the compiler will not know until it instantiate the templated function whether Assert is true or false?

Yes; so the conclusion would probably to not use this approach :P

Hmm, right, probably; I did not consider vectors yet.

The answer proposed about using another function "works" too.

template<typename T>
constexpr bool retFalse()
{
  return false;
}

/// This is the "model" template that is used to extract a JSON value into a
/// value usable by our object. If it is invoked, it will raise an error at
/// compile time as all used specializations should be defined.
template<typename T>
T asValue( const Json& data )
{
    static_assert(
      retFalse<T>(),
      "unknown type. make sure a specialization for asvalue for this type has "
      "been defined." );

  return T();
}

review queue full...

unbelievable...

This whole thing is annoying. I have made something similar to load data from XML; now I'm doing this for JSON too...

Yeah, I suppose using properties like this can open up those doors.

Here we're using the XML for our "environments"/Entities-and-Components (what the scenes are made up), while JSON (and another "proprietary" text format) are used for other program settings.

(we only read, we never write)

@Tyyppi_77 Thanks a bunch for your help!

Did they just changed the font on the main site?

Apparently.

Heh... /shrug.

Esc is the worst for chat.

I noticed the font change too. I don't think their "GitHub’s rationale for switching" for switching is an apples to apples comparison. When I looked at their before & after screenshots, the differences were too subtle for me to spot.

I'm sure a bunch of "regular users" will change the default font using a user script...

Not really a fan of Cascadia Mono either. I tried it for a day when the upcoming VS update was announced.

Yeah, I say they changed the font too... I'll see when it happens. Then change if need be..

I'm not a fan of them changing it either.

I mean, people can change it if they want, why push this to everyone?

/shrug

It sounds like if I uninstall Cascadia, it would fall back to Consolas, so I guess there's that.

I like Cascadia's license better than that Consolas, so on paper I would have agreed with the change. I just don't like using it as a typeface for work.

I wasn't familiar with the Segoe UI Mono option. Considering it doesn't have a dotted or slashed zero, I'm surprised it made the cut for mono width type consideration.

w.r.t. potential confusion with capital O?

Yes.

Yeah okay, I guess it is a weird choice. I suppose "well you'll use the context, right?".

It's not nearly as bad as some fonts, but for coding, I'm not interested in paying an extra mental tax that I could just as easily avoid.

Does it occur often though that you have to think whether it is a zero or a capital o?

Admittedly, I don't have data on that, even wrt myself. I will admit to being a very sub par typist though. I have mistakenly keyed lower case o instead of 0. As to the often enough part? I suppose that's debateable.

But in my day, I had t' walk uphill both ways t' the computer with m' stack a punch cards. My zee-rows had slahses 'n m' ohhs didn't. Now a days I can't tell em apart no more, 'n the kids won't git off m' lawn.

;)

Sometimes I type English things with how French is pronounced or written. Sometimes I type French stuff with how English things are pronounced or written. In my case, I have more to complain to my own brain than to how a character looks like!