Quicktest flagging errors that don't (I think) exist

[Bob_Ross2]

Hey all! I’m currently coding my first game using choicescript and just came across this error. I created new code sharing the same issue as mine to make it more readable. There are no other lines being passed, no variables being changed other than what is shown.

*create xmana 20
*create xhp 20
*create xmaxhp 20
*create xdamage 0

*if xmana < 0
	*set xhp +xmana
	*set xmana 0

*rand xdamage 0 (xmaxhp - xhp)

This gives the following error:

QUICKTEST FAILED
Error: startup line 10: Invalid rand statement, min must be less than max: 0 > -20

From what I understand, the Quicktest is skipping the *if and immediately adding 20 to xhp, creating an error at the *rand statement. This shouldn’t be possible during a real playthrough, as the only time they addition takes effect is when xmana is negative, resulting in a substraction.

Is there a way around this, or did I make some obvious mistake?

[Cecilia_Rosewood]

You just created xmana as 20, so it’s not < 0. (smaller than) That’s why quicktest (and manually playing through it) skips that if-statement. I think you meant xmana > 0 (bigger than)

[Bob_Ross2]

No, it is supposed to be “<”. I’m checking if the current amout of mana you have is negative (which can be triggered by some other conditions). If so, health decreases by that amount and mana gets reset to 0.

Edit: In practise, the code wouldn’t break because the smallest it can get is *rand 0 0. But because Quicktest bypasses the *if statement, it causes *rand to go 0 -20.

Thanks for answering though.

[cup_half_empty]

Random test plays like a user, selecting choices at random. This means some lines may not be tested at all. That’s why it’s good to setup a high number of repetitions, to guarantee most if not all your code will be checked under different conditions.

Quick test on the other hand ignores conditionals, choices, or any branching mechanism and simply executes every line of the source file in one go. This can cause the sort of issue you’re facing.

To work around that you need to add conditionals at problematic points. There’s a special meta variable that is set during quick test, but you can use an always-falsy conditional really. Quick test will execute it anyway.

*if (false)
   "Fix values for quick test"

*if (choice_quicktest)
   "You can use the special variable for it too" 

As you can see, the code inside the if block only runs in quick test.

By the way, there’s also a similar meta variable for random test, to hide options that would cause a loop for example. It’s choice_randomtest.

[Stewart_Baker]

It seems like there are a few things going on here.

Cecilia is totally right that your “if” statement will never run where it is currently placed.

But the reason you’re getting this specific error is that you’re attempting to randomize a number between 0 and 0:

*create xmana 20
*create xhp 20
*create xmaxhp 20
*create xdamage 0

*if xmana < 0
	*set xhp +xmana
	*set xmana 0

*rand xdamage 0 (xmaxhp - xhp)

By the time the game hits that code, xhp and xmaxhp are both 20. 20-20 is zero, and as a result your command says to randomize a number between 0 and 0.

As the error notes, your first value (the hard-coded 0) must be less than the second value (which currently is 0 due to the value of your xmaxhp and xhp stats).

ETA: If you want that rand code to run after the stuff in the *if test, you need to indent it. Currently it is not indented, and choicescript will treat it as an independent line of code.

That said, your startup file is not really the right place to put either the code to check current mana or the damage generation code if you want to refer to it elsewhere in your game.

If these are lines of code you will want to check regularly during the game, I would recommend placing them in their own scene file and then placing that file after the *ending of the game.

You can then load these snippets every time you need them with the gosub_scene command.

[Bob_Ross2]

Alright, so if I understand, there isin’t a problem with my code but rather with the way Quicktest itself works? And that being the case, adding the condition choice_quicktest = false to my *if statement will supress the error and allow Quicktest to move on to actual bugs?

[Bob_Ross2]

Hey there, thanks for answering! I tested it out, and *rand value 0 0 isin’t the cause of the problem. In fact, running that line just always returns 0.

From my testing, the root issue is that the Quicktest compiler is executing the lines under the *if statement without checking the condition itself. This raises an issue because that condition is what prevents xhp from being positive and becoming larger than xmaxhp, thus resulting in the error you highlighted.

In actual gameplay, this shouldn’t be possible, since the line adding xmana to xhp can only trigger when negative, thus substracting from xhp rather than adding to it and avoiding the issue altogether.

As to adress your second point, this isin’t my actual code, but rather a smaller program I made that duplicates the same issue my actual code is facing. Resolving this program will allow me to solve my actual code. The *rand line isin’t indented, and that’s on purpose. But unrelated, you just reminded me *gosub_scene existed, which I’m going to start using on other things.

[Stewart_Baker]

Thanks, that’s good to know.

What happens in your smaller program if you change the values so that xmana is created with a negative value?

*create xmana 20
*create xhp 20
*create xmaxhp 20
*create xdamage 0

*set xmana (xmana - 22)

*if xmana < 0
	*set xhp +xmana
	*set xmana 0

*rand xdamage 0 (xmaxhp - xhp)

(ETA I forgot you can’t create negative variables. But you can subtract from them to make them negative after they’ve been created.)

You could also try putting your *if condition in parentheses. This shouldn’t be necessary but might potentially cause difficulties.

*if (xmana < 0)
	*set xhp +xmana
	*set xmana 0

This would be easier to troubleshoot if you can share the error and relevant code from your full game. I have found that sometimes it is actually the thing right before a reported error (or something around it) that is causing the problem instead.

[Stewart_Baker]

Creating a new project in CSIDE that has only the following code results in no errors for me on either quicktest or randomtest:

*title xmana test

*create xmana 20
*create xhp 20
*create xmaxhp 20
*create xdamage 0

Xmana is ${xmana}.  Xhp is ${xhp}. Xmaxhp is ${xmaxhp}. Xdamage is ${xdamage}.

*set xmana (xmana - 22)

You used 22 mana! Your mana is now ${xmana}.

*if (xmana < 0)
    *set xhp +xmana
    *temp mdamage (0 - xmana)
    *set xmana 0
    Using too much mana causes you mana burn. Oof! Ouch! You take ${mdamage} HP worth of damage, reducing you to ${xhp} HP.
    
    Xmana is ${xmana}.  Xhp is ${xhp}. Xmaxhp is ${xmaxhp}. Xdamage is ${xdamage}.

*rand xdamage 0 (xmaxhp - xhp)
*set xhp (xhp - xdamage)
Suddenly, a goblin hits you! You take ${xdamage} HP worth of damage, reducing you to ${xhp} HP.

Xmana is ${xmana}.  Xhp is ${xhp}. Xmaxhp is ${xmaxhp}. Xdamage is ${xdamage}.

*ending

Here is the output from a randomtest that shows fulltext:

*****Seed 0
Xmana is 20. Xhp is 20. Xmaxhp is 20. Xdamage is 0. 

You used 22 mana! Your mana is now -2. 

Using too much mana causes you mana burn. Oof! Ouch! You take 2 HP worth of damage, reducing you to 18 HP. 

Xmana is 0. Xhp is 18. Xmaxhp is 20. Xdamage is 0. 

startup *rand xdamage 2
Suddenly, a goblin hits you! You take 2 HP worth of damage, reducing you to 16 HP. 

Xmana is 0. Xhp is 16. Xmaxhp is 20. Xdamage is 2. 


Word count: 90
RANDOMTEST PASSED
Time: 0.017s

Now here is what I get when I try to run the same code when I comment out the *set xmana (xmana - 22) code that reduces xmana below zero (and the message about using 22 mana)":

startup
executing
QUICKTEST FAILED
Error: startup line 22: Invalid rand statement, min must be less than max: 0 > -20

So (at least according to a quicktest on this very small subset of your code) the problem is that your min and max values in the quicktest *rand are both zero. (As @liliarch notes, choicescript thinks (xmaxhp - xhp) is -20!)

Note that the code still runs, and when it does it doesn’t call that *if statement (which, again, it shouldn’t in this case, because xmana is still 20, which is not below 0):

*****Seed 0
Xmana is 20. Xhp is 20. Xmaxhp is 20. Xdamage is 0. 

Suddenly, a goblin hits you! You take 0 HP worth of damage, reducing you to 20 HP. 

Xmana is 20. Xhp is 20. Xmaxhp is 20. Xdamage is 0. 


Word count: 52
RANDOMTEST PASSED
Time: 0.014s

This suggests that there does not appear to be a problem with choicescript ignoring your *if condition, either.

[LiliArch]

Well, going by that error message choicescript (quicktest, whatever) does think that xmaxhp - xhp is -20.

[Stewart_Baker]

Whoops. Good catch!

[Bob_Ross2]

Thanks for taking the time to help me with this. Your answer does solve my problem, indirectly at least.

As @cup_half_empty notes, Quicktest works by executing all the lines of code at once in order, no matter the conditions. This causes an error with the way my code was organized, because ignoring the *if statement, xmana is getting added to xhp and becoming 40, resulting in the later operation (xmaxhp - xhp) giving -20.

I don’t know if this was intentional on your part or not, but reducing the value of xmana before the *rand operation effectively avoids this problem entirely. On your code, when Quicktest executes, it reduces xmana to -2 and thus xhp to 18 before the *rand operation, avoiding the error.

In regards to my actual code (which functions essentially identically to the snapshot I provided, but with much more… embellishment), the lines where the player can reduce the value of xmana is located after the problem section, but accessible chronologically to the player beforehand through a use of *goto.

So essentially, my problem was situated into two parts.

1. As @cup_half_empty indicated: “Quick test on the other hand ignores conditionals, choices, or any branching mechanism and simply executes every line of the source file in one go. This can cause the sort of issue you’re facing.”
2. The way the code was organized, which allowed this problem to even take place.

Realizing that the reducing xmana before reaching the following section prompted me to migrate the block of code to the bottom (using *goto), and thus forcing Quicktest to go through all the lines, reducing xmana, and preventing xhp from being too high.

*if xmana < 0
	*set xhp +xmana
	*set xmana 0