I'm surprised that this hasn't been reported previously, because we can get
a crash even without atan or any division:
sin( x*(x+1) - x^2 - x ) # this crashes sage
This also crashes with sinh or cos or tan in the place of sin, but not with
exp or log.
PS I verified on CoCalc that this is not a new bug: this crashes version
9.1.
On Monday, May 29, 2023 at 5:51:28 PM UTC-6 Nils Bruin wrote:
> Localized a bit further:
>
> var("q A")
> p = A*(1+1/A)-A-1
> V=(q^p)._maxima_().rectform()
>
> after this, V._sage_() crashes, and I think it's the same crash as above.
>
> Transcribing what V is in maxima, we get:
>
>
> I*sin(atan2(0,q)*((1/A+1)*A-A-1))*abs(q)^((1/A+1)*A-A-1)+cos(atan2(0,q)*((1/A+1)*A-A-1))*abs(q)^((1/A+1)*A-A-1)
>
> which indeed causes a crash when input directly into sage. More simply:
>
> var("q A")
> U=atan2(0,q)*((1/A+1)*A-A-1)
> sin(U) #this crashes
>
> It looks like an infinite recursion, so probably the constructor (of sin?)
> tries to do some simplification, which ends up trying to construct the same
> expression again. The traceback implicates Pynac.
>
>
> On Monday, 29 May 2023 at 13:07:50 UTC-7 dwb...@gmail.com wrote:
>
>> I was shown the following way of getting a segmentation fault in Sage.
>>
>> sage: var("q A")
>> sage: p = A*(1+1/A)-A-1
>> sage: (q^p).full_simplify()
>>
>> This consistently causes a crash. The person who found it was doing some
>> actual work, got a crash, and boiled it down to a minimal example.
>>
>> Daniel Bump
>>
>
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/sage-devel/a308677c-a2b7-4391-b77a-78225a480419n%40googlegroups.com.
