[3.3.2] Useless error message is wrong

[Matt]

Code: Select all

class ThatStupidFuckingErrorMessage:Actor{
	override void postbeginplay(){
		super.postbeginplay();
		setorigin((12,198273,42),false);
//		if(max(abs(pos.x),abs(pos.y),abs(pos.z))>=32768){destroy();return;}
		A_PlaySound("baron/sight");
	}
	states{
	spawn:
		TNT1 A 100;
		stop;
	}
}

Expected: "Invalid sound position (12.000000, 198273.000000, 42.000000) for actor of class ThatStupidFuckingErrorMessage"

Actual: "Invalid sound position (12.000000, 42.000000, 198273.000000) for actor of class ThatStupidFuckingErrorMessage"

[Rachael]

All this talk of X and Y's are giving me the Z's.

(... Sorry, couldn't resist with the zany nature of this bug report)

I've taken a look at the source and the values being passed to the string parser are "technically" correct, at least in terms of how it is written, so I am going to let Graf or Chris review this one. And it is using the actor's position and velocity directly, not translating it, so it should be correct right from the outset.

[_mental_]

Swapped axes in 8a6ae50.

Could you please report issues in a more polite way?
This warning is not stupid nor useless. If you think so, create a separate topic with a suggestion to make it optional or to remove it completely.
I believe you want an engine that accepts everything you passed to it and ignores all bogus data silently.
And then, you will spent half a day in desperate attempts to figure out why something doesn't work, blaming developers for their engine being too friendly.

[Matt]

And then, you will spent half a day in desperate attempts to figure out why something doesn't work, blaming developers for their engine being too friendly.

And instead, because of your mistake, I spent weeks trying to figure out what the hell was wrong when it would have been patently obvious had the numbers been right to begin with.

And yes, this is completely useless. There is nothing that is served by this error message.

If the problem is all actors placed out of bounds, then there should be a warning immediately when the actor itself is out of bounds, not just when a sound plays.

If the problem is actually a problem with the sound being able to handle out-of-bounds sources, then I am welcome to hearing an explanation why. However, not once was that ever forthcoming, from you or anyone else, and on its face it simply makes no sense - after all, before you added this warning no one had ever noticed it being a problem.

This entire "validation" process has been a total waste of everyone's time.

[Rachael]

Matt, please calm down, and try to see the other side of this argument, not just your own.

Do not make this escalate into anything. You know what will happen if it does.

You don't have to agree with the developers' decisions, but that does NOT give you grounds whatsoever to use the kind of toxic language you've used here in this thread.

[Matt]

Okay, fine. What is the issue that this warning is designed to address?

The major source of my frustration is the very fact that I haven't been given the other side of this. There was no mention at all on the 3.3.0 announcement and the only explanation was mental's flippant "There is no point to use bogus position for sound source" which explains exactly nothing. Added this paragraph while mental was posting, disregard

[_mental_]

viewtopic.php?t=57632

[Rachael]

If you absolutely must have code consisting of one-liners to simplify your own code, then write wrapper functions containing the checks and call those, instead.

[_mental_]

Errare humanum est.

Initially I wanted to report only infinities and NaNs.
However, placing actors outside of theoretical map bounds isn't a very good idea.
It may work today and will break tomorrow because of unrelated changes.
That's why I decided to handle this situation too.

Please note that I asked about other opinions before adding the warning.
I'm OK to remove it or make it optional along with OpenAL message.

[Graf Zahl]

@Matt:

If you need to check for such messages to silence them, better ask yourself what YOU did wrong that these bogus values get in there in the first place. These are a clear indicator of something going wrong in your mod!

[Matt]

Graf:

I remember those OpenAL errors and I only ever got them in much more restricted contexts than where we're getting these error messages now. Each time I got that old message, it took very little time to realize that an actor was spawning or warping to some totally inappropriate place (like a couple miles underground or something, or disappearing altogether). Despite being cryptic gibberish it was still better at helping me find actual errors than what we're getting now.

32768 is a very small number. It translates to under 800m. It is not at all unreasonable to assume that such a distance would be handled without problem in a modern limit-removing sourceport. In my case the offending actors were acting completely within the parameters they were given.

If the lower limit is needed to keep everything else running smoothly (limit memory use for frequent checks, etc.) that's fine, but don't go around blaming me, or any other modder, for your decision to restrict the bounds to something much smaller than the engine could actually handle.

[Graf Zahl]

32768 is the outer bounds of map geometry. In other words: You are placing a sound outside the map!
Why? To have this limit lifted you have to present a reasonable use case.

Also, check your tone, please! If you go on ranting and insulting like you did in this thread I can promise you that someone will act on it.

[Matt]

I apologize for responding in kind, Graf.

[wildweasel]

I apologize for responding in kind, Graf.

As much as I can understand the frustration, I'm disappointed in how that frustration is being expressed. This could have gone a lot worse, I admit, but it feels like there are a lot of ways it could have gone better. I think Rachael's put it a lot better than I can at this point.

It may be worth distancing yourselves from the topic for a while and try to see things from an outside perspective.