Missing font-locking

[manuel-uberti]

Expected behavior

Clojure own functions/forms correctly highlighted.

Actual behavior

As of now, if, let, loop and recur (for instance) are correctly highlighted, but select-keys, map,
get-in and concat (the first ones I catch) are not.

Steps to reproduce the problem

• open a .clj file with clojure-mode installed
• write (map inc [1 2 3])

Environment & Version information

clojure-mode version information

clojure-mode (version 5.7.0-snapshot)

Emacs version

GNU Emacs 27.0.50 (build 1, x86_64-debian-linux-gnu, GTK+ Version 3.18.9) of 2018-03-14

Operating system

Ubuntu 16.04

[manuel-uberti]

Namespaces in :require have problems as well.

[bbatsov]

This is some regression from #462

@Bost Mind checking what went wrong?

[Olical]

And namespaced keywords only highlight the second colon, the first is left white.

[Bost]

@Olical the highlighting of namespaced keywords is a regression from #462 indeed. (I'll fix it)

@manuel-uberti @bbatsov font-locking of select-keys map get-in concat hasn't changed since 5.6.1. Anyway it can be easily changed by applying this patch:

index 073e287..0d28c63 100644
--- a/clojure-mode.el
+++ b/clojure-mode.el
@@ -789,6 +789,7 @@ any number of matches of `clojure--sym-forbidden-rest-chars'."))
          "("
          (regexp-opt
           '("def" "do" "if" "let" "let*" "var" "fn" "fn*" "loop" "loop*"
+            "select-keys" "map" "get-in" "concat"
             "recur" "throw" "try" "catch" "finally"
             "set!" "new" "."
             "monitor-enter" "monitor-exit" "quote") t)

[manuel-uberti]

I see, but even assoc, assoc-in, contains? are not highlighted now. Should every function/macro/form be added then?

[manuel-uberti]

Also, user-defined macros are not highlighted either.

[bbatsov]

Those are getting highlighted by CIDER, not clojure-mode. Perhaps @Bost's changes to the font-locking regexps affected CIDER's font-locking.

[bbatsov]

@manuel-uberti @bbatsov font-locking of select-keys map get-in concat hasn't changed since 5.6.1. Anyway it can be easily changed by applying this patch:

@Bost Seems I was unclear. Sorry about this. Have you tried whether the dynamic font-locking in CIDER was affected by your changes?

[manuel-uberti]

Those are getting highlighted by CIDER, not clojure-mode. Perhaps @Bost's changes to the font-locking regexps affected CIDER's font-locking.

FWIW, the screenshot with test-macro comes from a .clj file with CIDER jacked-in.

[manuel-uberti]

@Bost is there anything I can do to help fixing this?

[Bost]

@Bost Seems I was unclear. Sorry about this. Have you tried whether the dynamic font-locking in CIDER was affected by your changes?

Uhm, how that works? I mean I copy-pasted the content of clojure-mode/test.clj to cider-scratch and it looks fine (Changes from my latest PR #475 were included).

[Bost]

Unfortunately the spacemacs//find-ert-test-buffer doesn't work for my any more so I let the travis do the testing for me. That' why you see here the PR retries. But now everything looks fine, so please have a look.

[bbatsov]

@Bost See http://cider.readthedocs.io/en/latest/configuration/#dynamic-syntax-highlighting

[Bost]

@bbatsov I can't get the dynamic-syntax-highlighting working... hmm.
I compile my emacs from the source. I tried branches origin/emacs-26, origin/emacs-25, clojure-mode master and also 5.6.1 and I also tried to run .src/emacs --no-init-file and then installed cider and clojure-mode from melpa-stable. And nothing...

[bbatsov]

Seems to work just fine for me with the old version of clojure-mode. I just start CIDER and it's there. For you there's 0 difference before and after?

[bbatsov]

Btw, I wonder if your color theme makes it harder to spot the differences or something. Also - which CIDER are you using?

[Bost]

I was launching emacs with --no-init-file and using melpa-stable (and then M-x cider-jack-in and M-x cider-scratch) so I was using defaults pretty much everywhere.... but(!) the dynamic-syntax-highlighting doesn't appear until you actually EVALUATE some sexp containing something dynamically highlightable!!! Grrr! Ok. Now I can do the actual problem analysis...

[Bost]

... and now I'm confirming this issue is a regression from #462 indeed.

[bbatsov]

@Bost Any progress with solving it? Breaking the dynamic font-locking is much bigger than fixing ns font-locking, so without some solution for this soon I'll have to revert the previous commits.

[Bost]

I think it's better to revert it for the moment. (Although I'll be having some time this week to work on it.)

[Bost]

Huh, I think the dynamic syntax highlighting is fixed now. (It was easier than expected). Please have a look at the PR 475.
@bbatsov Regarding the revert of #462 - since the problem appears to be a simple one and I do have time, please consider postponing the decision until the next week. Thx.

[bbatsov]

Fixed in #475.