
I can reproduce it with a small-ish example. This looks like a bug - will open an issue on github.


Does anybody know if Carl Eastlund from mischief hangs around here? It seems that mischief stopped working on racket HEAD since it started providing stream-take
which clashes with mischief stream-take
.

@pocmatos he’s not on slack, I would either email or report a github issue

@samth thanks, I did report an issue but I got no response. I will wait longer. Maybe he’s on hols.

@pocmatos he no longer works on racket (he’s at Jane Street) so an email might be necessary

Will send him an email then. Thanks.

@samth not sure I understood your comment, do you want me to create a PR based for qemu testing on azure or travis? currently I have some work for this on travis but feels a waste of my time to continue if are moving the CI to azure.

@pocmatos as I think I’ve said, we’re not “moving” the CI

I created that PR to explore adding an additional CI, which I’m particularly interested in for Windows

because AppVeyor is slow enough that we can’t run lots of the tests

ah… got it. So you would keep the current tests on travis and use azure for windows?

makes sense.

but if there’s something that would benefit from Azure on linux then obviously that would be worthwhile too

ok, let me look into migrating my work from travis to azure then. I will then open a pr.

Ah, okay, that makes sense. Thanks.

@mflatt Do you think it would make sense to make racket/private/serialize-structs
be a cross phase module?

(That way the serializable struct property lasts across phases.

you’d have to manually write all that code in ’#%kernel

@samth Ya, @michael.ballantyne and I did that. =

@samth The code for it is actually not that complex, and mostly just requires you do most of the define-struct
stuff yourself: https://gist.github.com/LeifAndersen/51aab9801caa5cf76542e32b7b958566


note that that misses the static struct info (I don’t know if that’s used though)

You are correct.

However, it is not used in the racket/serialize library.

And raco setup
works fine without it.

(As well as other programs I have that make heavy use of serialize.)

It does mean though that the racket library couldn’t, say, make a serialize
struct match expander.

Sounds ok

I was hoping to get some help figuring out why I haven’t been able to trigger an on-event
callback for my image-snip%
. I’m probably missing something simple, but when I drag the snip, “on-event” isn’t printed and the snip still drags behind others. Does anyone know the right pattern for snip events?

@david.alkire, the way you call set-flags removes all existing flags from the snip. You need to do something like:


@alexharsanyi thanks. it’s printing the message now. getting those flags right is tricky. i’ll keep working on it

You will find that with all these flags, your snip will handle more events that you need, and the editor will miss out on some events. You can fix that by passing back to the editor all events that you don’t handle.