This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2018-01-25
Channels
- # arachne (1)
- # beginners (22)
- # boot (21)
- # cider (23)
- # cljs-dev (16)
- # cljsrn (9)
- # clojure (118)
- # clojure-dev (11)
- # clojure-greece (16)
- # clojure-italy (10)
- # clojure-losangeles (4)
- # clojure-russia (14)
- # clojure-serbia (4)
- # clojure-spec (58)
- # clojure-uk (33)
- # clojurescript (30)
- # cursive (17)
- # datomic (48)
- # docs (22)
- # events (1)
- # fulcro (24)
- # hoplon (3)
- # jobs (6)
- # jobs-discuss (4)
- # keechma (4)
- # leiningen (11)
- # luminus (4)
- # midje (1)
- # off-topic (107)
- # onyx (30)
- # other-languages (12)
- # pedestal (4)
- # re-frame (72)
- # reagent (6)
- # remote-jobs (1)
- # shadow-cljs (16)
- # spacemacs (3)
- # specter (9)
- # uncomplicate (4)
- # unrepl (40)
@bronsa Does your recent change to tools.reader now use normal (not TCO-optimized) recursion, and thus potentially a call stack that is as deep as long sequences of things to read in input files can be?
Pretty subtle bug, by the way. Thinking about ways that a linter like Eastwood might be able to warn about such a thing sounds pretty tricky, and like anything one would try to detect it would have lots of false positives and negatives.
@andy.fingerhut the recursion is only in whitespace paths
and # dispatch forms, looks like?
But I guess those would rarely nest nearly as deeply as whitespace.
@andy.fingerhut https://github.com/clojure/tools.reader/commit/5ec80af5729efad880724fa1c87fd7be2ef5608a
Cool. I wasn't even very deeply concerned after your explanation above, but that looks like one less Eastwood issue in the future. 🙂 Thanks.