Open bluenote10 opened 4 years ago
Hm, it's possible that this is a recent regression / change on devel. I'll take a look at it.
edit: the problem might just be that heatmap
is just a template and so mapIt
isn't found, because it's not imported in example_01.nim
. Can you check if it works if you import sequtils
there?
Also not sure if comments on commits show up in your notifications, so a link here: https://github.com/bluenote10/NimData/commit/8320f3c56ff9effd5f5ead1b4bfd1d2e88266042#commitcomment-40727862
edit: the problem might just be that heatmap is just a template and so mapIt isn't found, because it's not imported in example_01.nim
Yes manually importing on client side should be possible, but I've raised this as an issue, because that's not how such a problem should be worked around. It's normally the responsibility of the template author to bind necessary symbols properly. Otherwise client code would have to import lots of seemingly unrelated stuff just to get things compiled. In this case it would mean that import nimdata
is no longer suffcient and we'd have to document that whenever you import nimdata
you also have to import sequtils
because of some usage in nim-plotly templates. I'm not quite sure why it worked before though.
For now I have disabled the nim-plotly test parts in NimData to get a green build again (https://github.com/bluenote10/NimData/pull/68).
I vaguely remember that sometime in the last year some of the symbol lookup rules regarding templates were changed. So that's probably the reason.
I'll try to fix it with manual mixin
added.
I tried to get NimData to work on latest devel again, but I'm running into a weird problem coming from its
plotly
dependence. On travis I'm getting the error:I'm not quite sure why that is happening, because the code makes sense on first glance. Or do all these templates require to actually
bind
the symbols?