Sebastien Binet

Cool, sounds like a plan. Keep us posted.

I don't think it would be too hard (famous last words). I mean, delve does it for debugging, right ? the thing is, ATM, it would be a bit useless...

still interested in having a go interpreter. actually, a few months ago, I teamed up with Elliott (@elliott5) to draw some plans and try to accrete a critical mass: https://docs.google.com/document/d/1Hvxf6NMPaCUd-1iqm_968SuHN1Vf8dLZQyHjvPyVE0Q/edit...

Sure, why not. Let's make go-interpreter less of a ghost town :)

and I have started some more detailed leg work wrt bytecodes and VMs here: - https://sbinet.github.io/posts/2016-09-07-start-interpreter/ - https://sbinet.github.io/posts/2016-09-09-tiny-interpreter/

like `h5py` did (and probably `PyTables` too) one needs to implement a hdf5 converter for `Go` strings, so they are part of (and properly handled by) the hdf5 type system....

(apologies for the belated answer) `go-hdf5` should definitely handle this better. could you attach a little `h5py` test script as a reproducer?

**untested**, but one way would be to retrieve the matching `go` type from the `Datatype`: ``` go dtype, err := dset.Datatype() gotype := dtype.GoType() switch gotype.Kind() { case reflect.Uint8: container[i]...

@zchee feel free to steal parts or all of my initial PR as you deem necessary. I have very few cycles for go-clang these days, unfortunately...

I'd be willing to review those :) migrating `gonum/plot` to drop freetype (and use `font.Face`) has been on my want-to list for the longest time...