when it should return the original text.
The bug occurs only when the second argument of replace-first is a pattern and the third argument is a function.
Hi @richhickey, thanks for your work on clojure and it's ecosystem, thanks for not being blind to community and for accepting PR!
Imo this flow is more community-friendly in comparison to old-school email-patch workflow (offered http://clojure.org/contributing). Some ecosystem like kernel/os developers could (must?) do this way, because there are only a few people (may be just one) who can fully understand influence of changes to the codebase and automated tests are hard, it may be also applicable to lang-core development, but not for libs, don't you think?
Maybe it's time to face this community problem and start to merge and accept more from pull requests, not to do in such way e.g. - clojure/tools.trace#4 (comment)
My thoughts may seem naive but they're genuine :)
I also suggest to +1 people to convert this topic to a petition ;)
@razum2um give up. This fight has been fought dozens of times before.
TL;DR: Cognitect likes it this way.
Just do a search for "contribution process" or similar on the mailing list. Some of the longest threads:
There were discussions of this on Clojure conferences, IRC, et cetera. Every single long time Clojure contributor I've talked to about this in the last couple of years has given up. So will you.
GitHub fabricated this 'pull request acceptance' from the application of a patch that had nothing to do with it.