nodemap is virtually untested

[zoffixznet]

I see a single test for nodality and no other tests at all.

We also appear to have this method fully documented: https://docs.perl6.org/routine/nodemap

[zoffixznet]

On second glance, this routine is actually used under the hood by the hyper nodal method calls. So the routine is tested as far as weeding out its bugs goes, but perhaps underspecced in a sense that the roast won't cover implementations that don't implement their hyper nodal methcalls using this routine.

[lizmat]

I guess lack of certainty on the final semantics are the reason why it is undertested and underdocumented. One could consider it more or less an internal method at this stage, imo.

…

On 20 Dec 2017, at 16:30, Zoffix Znet ***@***.***> wrote: On second glance, this routine is actually used under the hood by the hyper nodal method calls. So the routine is tested as far as weeding out its bugs goes, but perhaps underspecced in a sense that the roast won't cover implementations that don't implement their hyper nodal methcalls using this routine. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

[zoffixznet]

One could consider it more or less an internal method at this stage, imo

Do you think it should be removed from the docs then? Seems to be unused in the ecosystem

(somewhat related Issue on writing extra docs for XXXmap methods: Raku/doc#1720 )

[lizmat]

I’ll leave that decision up to jnthn TimToady. I would be in favour of removing it from the doc for now.

…

On 20 Dec 2017, at 16:38, Zoffix Znet ***@***.***> wrote: One could consider it more or less an internal method at this stage, imo Do you think it should be removed from the docs then? Seems to be unused in the ecosystem (somewhat related Issue on writing extra docs for XXXmap methods: Raku/doc#1720 ) — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.

[zoffixznet]

I’ll leave that decision up to jnthn TimToady.

@jnthn @TimToady ping: should nodemap be considered an internal method and removed from docs (and also likely moved into Rakudo::Internal to handle the nodal hyper method calls)?

[zoffixznet]

https://irclog.perlgeek.de/perl6-dev/2017-12-20#i_15604510

jnthn | My recollection in this area is that some years back (pre-6.c), TimToady looked at the various meta-op guts impl functions we had and concluded that they should be public and given public names. produce (the function form of the [...] metaop) appeared at that time as a result of this
| iirc, nodemap/hypermap/deepmap came around this time
| And so were indeed intended as public