cljr-rename-symbol doesn't play well with 'pretty-mode'

[JAremko]

Steps to reproduce the problem

Run cljr-rename-symbol with foo renamed to bar on

(defn foo [arg1 arg2])
(foo 1 2)
(partial foo 1)

Expected behavior

(defn bar [arg1 arg2])
(bar 1 2)
(partial bar 1)

Actual behavior

(defn bar [arg1 arg2])
(bar 1 2)
(partial foo 1)

Environment & Version information

CIDER version

;; CIDER 0.18.0snapshot (package: 20180802.1919), nREPL 0.2.13
;; Clojure 1.9.0, Java 10.0.1

clj-refactor.el version

2.4.0-SNAPSHOT

Lein version

Leiningen 2.8.1 on Java 10.0.1 OpenJDK 64-Bit Server VM

Emacs version

26.1

[expez]

Can't reproduce. From lein new test I got:

(ns test.core)

(defn foo
  "I don't do a whole lot."
  [x]
  (println x "Hello, World!"))

(foo 1)

(partial foo 1)

Rename foo to bar:

(ns test.core)

(defn bar
  "I don't do a whole lot."
  [x]
  (println x "Hello, World!"))

(bar 1)

(partial bar 1)

You already had a bar and wanted foo to have the same name or something? Need more details to reproduce.

[JAremko]

@expez bad copy-pasta in the first listing. fixed.

[JAremko]

This is what I'm getting

[expez]

Do you get the same if you load the code into the REPL before doing the refactoring?

E.g. by using C-c C-k in CIDER?

[benedekfazekas]

i can see that you are using prettifing symbols on partial. try the feature with prettifing switched off. it might confuse the column counts cljr works with.

…

On Sat, 18 Aug 2018, 13:48 Eugene Yaremenko, ***@***.***> wrote: This is what I'm getting [image: https://i.imgur.com/oXPZD7z.png] <https://camo.githubusercontent.com/8ec487650865bde5fe823adeda1762261fdeb314/68747470733a2f2f692e696d6775722e636f6d2f6f58505a44377a2e706e67> [image: https://i.imgur.com/xBNvQhC.png] <https://camo.githubusercontent.com/2cd4ccd9c6ad8bb7754b4e1273b282c293c9a60f/68747470733a2f2f692e696d6775722e636f6d2f78424e765168432e706e67> — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#429 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA9PnjRk-CtU_RY1mDKhfKddrwI7I2vpks5uSA0fgaJpZM4WCgnu> .

[JAremko]

@expez Doesn't help. Btw it reports Renamed 3 occurrences of foo so it sees all 3 occurrences.
Also sometime (after rename) I get:

@benedekfazekas OMG 🙀 it helped. I don't understand.

[JAremko]

So basically cljr-rename-symbol conflicts with clojure-enable-fancify-symbols in Spacemacs. I'm just not sure how font lock shenanigans can affect cljr-rename-symbol.

[JAremko]

it might confuse the column counts cljr works

@benedekfazekas Checked. You are right. It affects column number. It's strange that there was no error report or warning of any kind. Can be really nasty in big projects.

---

Now we just need to decide who's bug it is 😄
I vote for clj-refactor.el bug!

[benedekfazekas]

yeah. used to love prettify symbols too. but it messes with cljr and even more importantly with indentation. I still use it but very tuned down.

…

On Sun, 19 Aug 2018, 11:42 Eugene Yaremenko, ***@***.***> wrote: it might confuse the column counts cljr works @benedekfazekas <https://github.com/benedekfazekas> Just caked. You are right. It affects column number. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#429 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA9PniiNa3AES_JTWdWnQ087bfAvkPq0ks5uSUD8gaJpZM4WCgnu> .

[expez]

Is there a way to get the 'real' column number? If not, then I don't see how cljr could handle this...

[JAremko]

@expez Indirect buffers probably can be used for this. They can have different modes then the original one.

[JAremko]

@expez Also in this case composite.el has function find-composition that probably can help.

[JAremko]

Keep in mind that this also doesn't work:

(defn foo
  "I don't do a whole lot."
  [x]
  (println x "Hello, World!"))

(foo 1)

(partial map identity) (foo 1)

I don't know how clj-refactor.el actually does substitution but from what i gather the best way to solve it will be letf-ing prettyfication list around the substitution logic so nothing will be "prettyfied" when you visit files for substitution. To deal with files that already have visiting buffer you can create Indirect buffers of those without pretty-mode and substitute in them.

[expez]

@JAremko We can get issues at both ends of this:

1. At the beginning we send down line/col to the middleware to find occurrences of a symbol to work on .
2. After the middleware returns we use line/col to make edits to the right locations.

Both of these will report the wrong column number with pretty-mode, but sometimes it isn't fatal (because you're still inside the same symbol).

[benedekfazekas]

also the 2. is reported by the middleware (refactor-nrepl) which is clojure code working with the files in the filesystem. no way that will know about prettify symbols

…

On Mon, 20 Aug 2018, 10:20 Lars Andersen, ***@***.***> wrote: @JAremko <https://github.com/JAremko> We can get issues at both ends of this: 1. At the beginning we send down line/col to the middleware to find occurrences of a symbol to work on . 2. After the middleware returns we use line/col to make edits to the right locations. Both of these will report the wrong column number with pretty-mode, but sometimes it isn't fatal (because you're still inside the same symbol). — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#429 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA9PniS-ymA6l3OpiUdl2k1eKLIThJ0Kks5uSnE2gaJpZM4WCgnu> .

[expez]

no way that will know about prettify symbols

Thankfully it doesn't have to, because the pretty symbols only really exist in emacs. The regular text is what's stored in the file on disk.

[benedekfazekas]

right. sorry, what i meant is that the line/cols returned by the middleware are not emacs buffer related

…

On Mon, 20 Aug 2018, 12:13 Lars Andersen, ***@***.***> wrote: no way that will know about prettify symbols Thankfully it doesn't have to, because the pretty symbols only really exist in emacs. The regular text is what's stored in the file on disk. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#429 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA9PnqPk0jdgnafQepKPseFYqnY8oKWHks5uSou6gaJpZM4WCgnu> .

[Nimamoh]

has anyone found a workaround for this issue?

[JAremko]

@Nimamoh I ended up simply disabling fancy symbols primary because they also mess up long line highlight.