[asterisk-bugs] [JIRA] (ASTERISK-26493) Is REMAINDER behaving in intended way?

[cloos (JIRA)]

    [ https://issues.asterisk.org/jira/browse/ASTERISK-26493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=232820#comment-232820 ] 

cloos commented on ASTERISK-26493:
----------------------------------

This got me curious.

REMAINDER() just calls remainderl(3) from libm.

That says that the quotient is rounded to the nearest int, so negative values are not unexpected.

For an always non-negative result you want to use modulus (infix %).


> Is REMAINDER behaving in intended way?
> --------------------------------------
>
>                 Key: ASTERISK-26493
>                 URL: https://issues.asterisk.org/jira/browse/ASTERISK-26493
>             Project: Asterisk
>          Issue Type: Bug
>      Security Level: None
>          Components: Core/General
>    Affects Versions: 14.0.1
>         Environment: Linux 4.4.0-38-generic #57-Ubuntu SMP Tue Sep 6 15:42:33 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
>            Reporter: Jonathan Harris
>            Severity: Minor
>
> I was having issues with getting minutes and seconds from seconds, using some code provided by another user, using remainder I was getting negative seconds.
> James Thomas helped me find a workaround, but from what he wrote, I wonder if REMAINDER is behaving as expected?
>  the following test dialplan shows what I mean:
> {code}
>  exten => 7,1,Verbose(Context: ${CONTEXT} Exten:${EXTEN}) 
>     same => n,Set(seconds=57)
>     same => n,While($[${seconds} <= 400]);
>     same => n,Set(minutes=$[FLOOR(${seconds} / 60)])
>     same => n,Set(myRemainderSec=$[REMAINDER(${seconds},60)])
>     same => n,SET(myModSec=${MATH(${seconds}%60,int)})  
>     same => n,Verbose(1,Seconds:${seconds} = Minutes:${minutes} Remainder Seconds:${myRemainderSec} modulo seconds:${myModSec})
>     same => n,Set(seconds=$[${seconds}+3])
>     same => n,EndWhile()
> {code}
> This is the output:
> {code}
>  Seconds:57 = Minutes:0 Remainder Seconds:-3 modulo seconds:57
>  Seconds:60 = Minutes:1 Remainder Seconds:0 modulo seconds:0
>  Seconds:63 = Minutes:1 Remainder Seconds:3 modulo seconds:3
>  Seconds:66 = Minutes:1 Remainder Seconds:6 modulo seconds:6
>  Seconds:69 = Minutes:1 Remainder Seconds:9 modulo seconds:9
>  Seconds:72 = Minutes:1 Remainder Seconds:12 modulo seconds:12
>  Seconds:75 = Minutes:1 Remainder Seconds:15 modulo seconds:15
>  Seconds:78 = Minutes:1 Remainder Seconds:18 modulo seconds:18
>  Seconds:81 = Minutes:1 Remainder Seconds:21 modulo seconds:21
>  Seconds:84 = Minutes:1 Remainder Seconds:24 modulo seconds:24
>  Seconds:87 = Minutes:1 Remainder Seconds:27 modulo seconds:27
>  Seconds:90 = Minutes:1 Remainder Seconds:-30 modulo seconds:30
>  Seconds:93 = Minutes:1 Remainder Seconds:-27 modulo seconds:33
>  Seconds:96 = Minutes:1 Remainder Seconds:-24 modulo seconds:36
>  Seconds:99 = Minutes:1 Remainder Seconds:-21 modulo seconds:39
>  Seconds:102 = Minutes:1 Remainder Seconds:-18 modulo seconds:42
>  Seconds:105 = Minutes:1 Remainder Seconds:-15 modulo seconds:45
>  Seconds:108 = Minutes:1 Remainder Seconds:-12 modulo seconds:48
>  Seconds:111 = Minutes:1 Remainder Seconds:-9 modulo seconds:51
>  Seconds:114 = Minutes:1 Remainder Seconds:-6 modulo seconds:54
>  Seconds:117 = Minutes:1 Remainder Seconds:-3 modulo seconds:57
>  Seconds:120 = Minutes:2 Remainder Seconds:0 modulo seconds:0
> {code}
> James Thomas wrote:
> {quote}
> All I can tell you is where -3 comes from. 
> From http://www.voip-info.org/wiki/view/Asterisk+Expressions :
> REMAINDER(x,y) computes the remainder of dividing x by y. The return value is x - n*y, where n is the value x/y, rounded to the nearest integer. If this quotient is 1/2, it is rounded to the nearest even number.
> -3 comes from:
> n = x/y = 957/60 = 15.95 which rounds to 16
> n*y = 16*60 = 960
> x - 960 = 957-960 = -3
> I'm not mathematically gifted either but I think the n is the problem. it shouldn't be the rounded result it should be the integer part of x/y (n=15)
> Can you just use modulo instead: ${MATH(${myNum}%60,int)}
> {quote}
> The reason I'm filing this as a potential bug is that all other tools I've tried round it correctly, but REMAINDER acts differently. So just checking...



--
This message was sent by Atlassian JIRA
(v6.2#6252)