I keep forgetting there's a slack channel for most anything in clojure
I'm curious what's the rationale for *out*
being in cljs.core
while *err*
and *in*
are in planck.core
?
Would imagine it’s because only self hosted cljs has in and err. “Normal” cljs runs in the browser which AFAIK only has console.log to do output.
Yeah. Some of those could end up in cljs.core
ultimately. An example of that kind of thing is *command-line-args*
which didn't make a whole lot of sense in cljs.core
initially.
hey @mfikes I'm trying to add to int-tests for the tty PR. the bash context turned out to be more useful than cljs, do you mind the PR expanding those tests? also, I used /usr/bin/script
which is ok on linux but mac os x needs some work, not sure what (I don't use that platform)
Yeah, some of the terminal-oriented script-based tests might be better for the tty?
testing. (I don't see how you'd really test things like that via the normal cljs.test
-style testing.)
macOS has the script
utility as well (https://unix.stackexchange.com/questions/25639/how-to-automatically-record-all-your-terminal-sessions-with-script-utility/317044).
@mike858 thanks, I didn't know. I'm using it to fake a tty for a subprocess, but CI reports different behavior on linux and mac os x (see https://github.com/planck-repl/planck/pull/911/commits/a9de737ba76764acff19b3887ecf192e3ef3f4e2#diff-9c7c4bea66c6d9d1f4ddf56049bf0f4a)
I've no ability to debug it, so if you have any ideas that would help. otherwise I'll drop the test
The f
switch operates differently (according to that SO page). Maybe that's the issue?
sounds plausible
I'm not on my Mac at the moment but can try later.
thanks. I'll take a close look at -f
meanwhile
Cool addition, by the way :)
necessity -> invention. we started using planck at work, already really great for quick scripting. there's always a wishlist of features though
well, turns out script
is way different on osx and linux. seems to work in CI now with a very different invocation for each. cc @mike858
@mail462: Oh, good to know!