Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

INT should round towards negative infinity instead of towards zero #96

Closed
dnprod opened this issue Mar 7, 2019 · 5 comments
Closed

INT should round towards negative infinity instead of towards zero #96

dnprod opened this issue Mar 7, 2019 · 5 comments
Assignees
@robhagemans
Labels
bug fixed

Comments

@dnprod
Copy link

dnprod commented Mar 7, 2019

Bug report

i thought the int() function should behave like the C runtime floor() function, but it seems to misbehave
on values greater than -1.0 and smaller than 0.0
int(-0.6) returns zero, i expected -1

am i misunderstanding some subtely of the int() function, or is it really a bug?

Program
10 POS1=1.6
20 POS0=.6
30 POSM0=-.6
40 POSM1=-1.6
50 PRINT " val: ",POS1,POS0,POSM0,POSM1
60 PRINT "int expect: ",1,0,-1,-2
70 IPOS1=INT(POS1)
80 IPOS0=INT(POS0)
90 IPOSM0=INT(POSM0)
100 IPOSM1=INT(POSM1)
110 PRINT " got: ",IPOS1,IPOS0,IPOSM0,IPOSM1
120 IIPOS1=INT(1.6)
130 IIPOS0=INT(.6)
140 IIPOSM0=INT(-.6)
150 IIPOSM1=INT(-1.6)
160 PRINT" still got: ",IIPOS1,IIPOS0,IIPOSM0,IIPOSM1

Ok
run
val: 1.6 .6 -.6 -1.6
int expect: 1 0 -1 -2
got: 1 0 0 -2
still got: 1 0 0 -2
Ok

PC-BASIC version: 1.2
Operating system version: win7/64

thank-you for a cool retro toy :)
@dnprod
Copy link
Author

dnprod commented Mar 7, 2019

in case hardware instructions are involved in the math, my machine is
AMD Phenom II X4 830 @ 2.80 GHz with 16.0 GB (15.7 GB usable) memory

@robhagemans
Copy link
Owner

Thanks for reporting! This looks like a bug as INT(-0.5) should give -1 as per documentation and GW-BASIC behaviour. I'm a bit surprised I missed this before...

Rob

@robhagemans robhagemans added the bug label Mar 8, 2019
@robhagemans robhagemans self-assigned this Mar 8, 2019
@robhagemans robhagemans changed the title int() function misbehaves with teeny tiny negative numbers INT should round towards negative infinity instead of towards zero Mar 8, 2019
@udhos
Copy link

udhos commented Mar 8, 2019 •
edited
Loading

I think that INT(-0.1) should give -1 as well.

See:

PRINT INT(98.89)
 98
PRINT INT(-12.11)
 -13

From: https://hwiegman.home.xs4all.nl/gw-man/INT.html

It is interesting to compare with FIX():

http://www.antonis.de/qbebooks/gwbasman/fix.html

  • GW-BASIC INT() is similar to C floor().
  • GW-BASIC FIX() is similar to C trunc().

@dnprod
Copy link
Author

dnprod commented Mar 8, 2019

the code i was looking at (calculates moon rise/set time) seem to employ the sequence
value = num - int(num)
several times to map numbers into a positive value between 0 and 1 before tossing them into trig functions.
the bug allowed a negative value to slip into the later calculations.
if this trick was commonly used, i could imagine this bug creating other weird bugs in trig-using
tools (audio ?) since the downstream math would not be expecting negative values.
in my case the moon was up in the daytime every day -- only fatal to goths :)

robhagemans added a commit that referenced this issue Mar 9, 2019
@robhagemans
add test negative_INT_FIX for issue #96
1fa33d1
robhagemans added a commit that referenced this issue Mar 9, 2019
@robhagemans
fix issue #96: INT should round towards negative infinity - this was …
843da93 
…broken for INT(x) in range -1 < x < 0
robhagemans added a commit that referenced this issue Mar 9, 2019
@robhagemans
fix issue #96: INT should round towards negative infinity - this was …
1484002 
…broken for INT(x) in range -1 < x < 0
robhagemans added a commit that referenced this issue Mar 9, 2019
@robhagemans
add test for issue #96
560464d
@robhagemans
Copy link
Owner

fixed by commits 1484002 (1.2 branch) and 843da93 (2.0 branch)

robhagemans added a commit that referenced this issue Sep 17, 2020
@robhagemans
add test negative_INT_FIX for issue #96
de83e93
robhagemans added a commit that referenced this issue Sep 17, 2020
@robhagemans
fix issue #96: INT should round towards negative infinity - this was …
5caea8c 
…broken for INT(x) in range -1 < x < 0
robhagemans added a commit that referenced this issue Sep 26, 2020
@robhagemans
add test negative_INT_FIX for issue #96
885aeed
robhagemans added a commit that referenced this issue Sep 26, 2020
@robhagemans
fix issue #96: INT should round towards negative infinity - this was …
e8a36e1 
…broken for INT(x) in range -1 < x < 0
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@robhagemans robhagemans

Labels
bug fixed
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@udhos @robhagemans @dnprod