@henryw374 look at deps.edn for running the tests - however if this is the same patch as the one in your fork there are failing tests I think?
a relatively straight forward fix https://github.com/clojure/clojurescript/commit/15f330f31a57ef7586b37a81d539f67f2a66b168
a pretty interesting bit of code for the compiler hackers
@mfikes ^ would like to see what Canary says about this
yes there are some failing tests. I can look at those. but it would be good to get some idea I'm headed in the right direction here. I mean it seems to work for the compilation scenarios I know of, but I've had to move the code that reads data_readers.cljc... put it in the analyzer ns for now.
@henryw374 oddly I can't preview the patch in JIRA
perhaps there's a formatting problem?
there's just something wrong w/ it - nothing in there for me
ok weird. I'll have a look...
@dnolen I've replaced that patch with a new one .. which now has some changes in it ;-)
@dnolen Looks like other cases might be cropping up. I'll see if I can write new ticket(s) as needed to cover them. https://travis-ci.org/github/mfikes/coal-mine/builds/771684236
@mfikes I'm a Travis newb - how do I see what's failing?
@dnolen That particular job is split across 5 "parts", and 2 of them are failing ^
Click on a failing one and then scroll down and you will see the output
@mfikes out of curiosity what are these? 4clojure problems? 🙂
Yes. It is a source of diverse code for which we know the expected behavior of.
ok wasn't really related to my fixes which did address the problem
rather we've never done generic walking and we were missing the :children
key on some binding AST nodes
so this fixes a bug, and now we have a decent test case that checks that generic walking works
@mfikes is the next build kicked off?
@dnolen I'll kick one off (normally they are run at night but it is easy enough to manually kick one off)
@mfikes already looking promising
the two that failed before now succeed
all green