Regression since 1.4.1 - result of calculation involving java Long is converted to Double

[bouncysteve]

All of the following passed in 1.4.1. Only the first two pass in 1.5.0:

@math
Scenario: convert result of adding two ints to string
    * string result = 3 + 1
    # passes
    * match result == '4'
@math
Scenario: convert java Long to string
    * string result = new java.lang.Long(4)
    # passes
    * match result == '4'
@math
Scenario: convert result of adding int to java long to string
    * string result = new java.lang.Long(3) + 1
    # fails - converts to '4.0'
    * match result == '4'
@math
Scenario: convert result of adding java long to int to string
    * string result = 3 + new java.lang.Long(1)
    # fails - converts to '4.0'
    * match result == '4'
@math
Scenario: convert result of adding two java longs to string
    * string result = new java.lang.Long(3) + new java.lang.Long(1)
    # fails - converts to '4.0'
    * match result == '4'

[ptrthomas]

I'm not sure how to deal with this to be honest, and I don't think using java.lang.Long etc. is common and certainly has never been encouraged. I'll leave this open and see if any others report this

[bouncysteve]

Thanks for the prompt response. If this is unlikely to be fixed then I'll work around it in our tests, but if you have any pointers of what change may have introduced this, then I'm happy to work on a fix instead.

…

On Fri, 16 Aug 2024, at 1:56 PM, Peter Thomas wrote: I'm not sure how to deal with this to be honest, and I don't think using `java.lang.Long` etc. is common and certainly has never been encouraged. I'll leave this open and see if any others report this — Reply to this email directly, view it on GitHub <#2600 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA4QSWFER7VBXENL3TEZRYLZRXZHFAVCNFSM6AAAAABMUAMYS2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEOJTGQ3DCNBTGI>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[ptrthomas]

@bouncysteve the change is most likely the JS engine upgrade here: #2536 - which is why this is a hard problem to fix

JS engine changes have been so troublesome that we decided to write our own as a plan B, FYI: https://github.com/karatelabs/karate-js

just curious - why do you need to do this manipulation in the first place, can't you just use parseInt() instead of the java Long ?

[bouncysteve]

We do a lot of work with dates, for which we use Java DateTime classes, etc. Fortunately we have a utility class which marshalls a lot of this work and it has been straightforward for us to change the few methods which return longs to return ints, so I'm not pushing for a fix if it's hard to do, thanks.

…

On Fri, 16 Aug 2024, at 6:38 PM, Peter Thomas wrote: @bouncysteve <https://github.com/bouncysteve> the change is most likely the JS engine upgrade here: #2536 <#2536> - which is why this is a hard problem to fix JS engine changes have been so troublesome that we decided to write our own as a plan B, FYI: https://github.com/karatelabs/karate-js just curious - why do you need to do this manipulation in the first place, can't you just use `parseInt()` instead of the java Long ? — Reply to this email directly, view it on GitHub <#2600 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA4QSWHF7GFQGIHR3KCTLXLZRY2LBAVCNFSM6AAAAABMUAMYS2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEOJTHEYDCNZXGE>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[ptrthomas]

@bouncysteve good to know, thanks. I'll keep this open for a while anyway - to see if others have similar issues

[ptrthomas]

closing as no others have reported this