2017-01-10
| 22:11 | timsgardner | I'm getting obnoxiously huge errors from cljs.spec.test/instrument when large data fails the spec, is there any easy way to truncate that? |
| 22:14 | tolstoy- | What's the reason to use reducers over regular map/reduce? |
| 22:22 | justin_smith | tolstoy-: they can evaluate in parallel, for one |
| 22:22 | justin_smith | they have logic for efficiently using the CPU |
| 22:22 | tolstoy- | You'd use them if you had huge data sets, for instance? |
| 22:23 | justin_smith | huge data sets and a task that was dominated by CPU time that was parallelizable |
| 22:31 | miatomi | exi |
| 22:43 | tolstoy- | Browsing the clojure.org site: it's amazing how much of Clojure you really don't have to take advantage of to get a lot done. Protocols. Hierarchy. Reducers. Transducers. Metadata. Transients. Deftype/Defrecord. Agents. Refs. Even defmulti if a case statement does the job. |
| 22:46 | technomancy | tolstoy-: boring doesn't sell well |
| 22:47 | tolstoy- | Eh? |
| 22:48 | justin_smith | tolstoy-: in the case of protocols an metadata, you might not use them explicitly but you won't get much done in clojure without using code that uses them |
| 22:48 | technomancy | I mean, the stuff you actually use day-to-day doesn't get front-page billing on clojure.org to convince you to try clojure |
| 22:49 | tolstoy- | Yeah, it's all great stuff, but the core is pretty solid. ;) |
| 23:25 | TimMc | TEttinger: I should clarify that I wasn't up 'til 2 AM making a patch; last night I just found and reported it, and weavejester woke up to an email prompting him to do so today. |
| 23:26 | TimMc | haven't even seen the patch yet |