Hi Henry,

thanks for your reply. This makes sense, since when I call dissect
with a scalar value for x and y, it works as expected.

Regards,

Thomas

Am So., 13. März 2022 um 15:31 Uhr schrieb Henry Rich <[email protected]>:
>
> That message indicates an internal dissect error; an inconsistent
> condition was detected.
>
> IIRC dissect has not been upgraded to support the form [email protected] where v
> returns rank>0.
>
> That form executes as follows:
>
> 1. executes [x] v y to produce an array of selectors s which must be in
> the range [-#u,#u) .
> 2. infers the shape of a cell of [x]/y to be (-@#@$s)}.$[x]/y.
> 3. collects the argument cells of [x]/y to be operated on by each atom in u.
> 4. executes the verbs
> 5. rearranges the results to match the positions of the selectors
>
> Henry Rich
>
> On 3/13/2022 10:16 AM, Thomas Bulka wrote:
> > Hello forum,
> >
> > I wrote a small verb:
> >
> > r =: *`]@.(0: <: ])
> >
> > This works as expected:
> >
> >     0.1 0.01 r _1 2
> > _0.1 2
> >
> > However, it seems to be impossible to visualize the data flow of the
> > verb with dissect:
> >
> >     dissect '0.1 0.01 (*`]@.(0: <: ])) _1 2'
> > assertion failure
> > |assertion failure: inheritu
> > | 0 e.frame['u failed but u@v succeeded'
> >
> > Error during initial display
> >
> > I don't understand exactly, what assertion failed. Is there a problem
> > with my verb? Or are there currently limitations in dissect which make
> > it impossible to visualize the data flow?
> >
> > Thanks in advance for your help.
> >
> > Regards,
> >
> > Thomas
> > ----------------------------------------------------------------------
> > For information about J forums see http://www.jsoftware.com/forums.htm
>
>
> --
> This email has been checked for viruses by AVG.
> https://www.avg.com
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to