planck 2016-11-04

anyone else seeing this?

Jonathans-MacBook-Pro:planck jonathan$ script/test-all-c 
Running unit tests...
Running integration tests...
Running integration tests...
462c462
< Only :as alias, :refer (names) and :rename {from to} options supported in :require; offending spec: [foo.bar :quu] at line 1 
---
> Only :as alias and :refer (names) options supported in :require; offending spec: [foo.bar :quu] at line 1 

Integration tests have failed.

@johanatan You may need to run script/clean so that it builds a new ClojureScript compiler

(I suspect that because the error message diff refers to some stuff present in the newer compiler releases.)

could failure to run clean also produce a planck binary that isn't kosher?

Yeah… I’d have to check but clean may use Planck itself, which is probably a bad idea in hindsight

after building HEAD on planck-c, i tried planck with my screencap script and it seemed to just hang indefinitely (without printing anything). Initial debugging indicated that the main thread was in block_until_timers_complete

which would imply that there was a SUCCESS returned back to main

[ahh, yea, I just brew installed --HEAD to get a working 'planck' for the clean lol 🙂 ]

Yes, I’ve done that. It is silly to have to install planck to clean itself. Ugh.

Perhaps I’ll revise some of Planck’s build scripts to not use Planck.

@mfikes do you think it is a good idea for processes started with sh to inherit the pgid of planck itself?

[i'm occassionally seeing runaway processes and it would be nice to be able to kill them all in one fell swoop]

No clue. I wonder if Clojure’s sh impl behaves that way.

What you suggests seems to make sense.

well, i can't see how it would hurt anything and having it that way would make the use case above nicer