[SOLVED] UnigineScript incorrectly parses numeric literals

[simon.wade_]

There appear to be a few issues with UnigineScript parsing float and double literal values. Here's an example you can try:

 

void testLiterals()

{

float f;

 

// Let's try and use this value

f = 4.3E-7f;

log.message("4.3E-7f is %g\n", f);

 

// So maybe scientific notation is the issue...

f = 0.00000043f;

log.message("0.00000043f is %g\n", f);

 

// How about doubles?

f = 4.3E-7;

log.message("4.3E-7 is %g\n", f);

 

// And again with the notation

f = 0.00000043;

log.message("0.00000043 is %g\n", f);

 

// Maybe it's a printing issue?

log.message("4.3 is %g\n", f * 10000000);

 

// *sobs*

}

 

Results in the following:

 

4.3E-7f is 1e-006

0.00000043f is 1e-006

4.3E-7 is 0

0.00000043 is 0

4.3 is 0

 

Obviously this is not so great. Sadly, I'm not sure which SDK version I'm running (will hopefully find out tomorrow), but the log file tells me I'm running version 2.47 of the Interpreter (if that helps). Let me know if you'd like any additional information and I'll hunt it down for you :D

[frustum]

The result from interpreter 2.48 is:

 

:!./test.usc

4.3E-7f is 4.3e-07

0.00000043f is 4.3e-07

4.3E-7 is 4.3e-07

0.00000043 is 4.3e-07

4.3 is 4.3

 

There were no changes regarding literals parsing recently.

[simon.wade_]

Interesting... I just tried it from the command line with interpreter 2.48 and it worked fine. However, if I tried running the same thing from within the editor - by jamming the code in to the init() function in the scratch project - I get:

 

4.3E-7f is 0

0.00000043f is 0

4.3E-7 is 4.3e-007

0.00000043 is 4.3e-007

4.3 is 4.3

 

So at least the doubles are working, but the floats are still wacky :(

[sergey.shandar]

I've experienced the same problem. Can anyone else confirm it?

[manguste]

Please, provide us with a test scene for reproduction, it would be most helpful for solving this issue.

[frustum]

It was a really strange interpreter issue... Interpreter used a default variable compare operator to remove constant duplicates. And from this operator's point of view 1.0e-6f is equal to 4.7e-7f... The problem is fixed.