elixir-editors / emacs-elixir Goto Github PK

:test.call([dsadsa: :test.call("dsadsa"),
            dsadsa: :test.call("sada")])

:test.call([dsadsa: :test.call("dsadsa"),
                          dsadsa: :test.call("sada")])

a = "" <> "?"

[12:58:37] <betawaffle>  it seems to mark everything that comes after with the string coloring
[12:58:39] <antifuchs>   ah (:
[12:58:49] <antifuchs>   that seems dissatisfying
[12:58:54] <antifuchs>   can you give me an example?
[12:59:06] <antifuchs>   also, which emacs version are you on?
[12:59:07] <betawaffle>  sure
[12:59:41] <betawaffle>  24.3.50
[13:00:31] <betawaffle>  well, if i type ?x -> x
[13:00:42] <betawaffle>  both the -> and the next x are also "string" color
[13:01:11] <betawaffle>  i think it only fixes itself if i kill the buffer and re-open it
[13:02:02] <betawaffle>  so, i tried making ? a prefix character, and adding an entry to the keyword table thing
[13:02:09] <betawaffle>  to match those expressions
[13:02:38] <betawaffle>  and that worked, but it didn't fix the indentation after a ?\b
[13:03:08] <betawaffle>  anything with a \ in it messes up indentation, unless the \ is in a string or something
[13:06:53] <betawaffle>  any ideas?
[13:06:55] <antifuchs>   that sounds about right
[13:07:21] <antifuchs>   I think that behavior means that the syntax-propertize thing is buggy…
[13:07:23] <antifuchs>   let me test
[13:09:43] <antifuchs>   hm, yeah, that does look completely wrong.
[13:09:46] <antifuchs>   grump
[13:10:07] <antifuchs>   so it doesn't syntax-color everything after the ?x as a string
[13:10:20] <antifuchs>   but the case indent is completely bad
[13:10:50] <betawaffle>  well, it does, but only for a second, and then it doesn't get reset back to default color
[13:11:23] <betawaffle>  it's like the property is sticking to it
[13:11:28] <antifuchs>   hm, not on this end. hmmm. argh
[13:11:37] <antifuchs>   I was testing on my local syntax highlighting branch
[13:11:46] <betawaffle>  so you type ?xx and the second x is regular?
[13:11:49] <antifuchs>   so, yay, it's working in the new world! isn't that something? ):
[13:12:02] <betawaffle>  new world?
[13:12:15] <antifuchs>   hah, the one where elixir will be able to deal with %r{} and such
[13:12:20] <antifuchs>   elixir-mode that is
[13:12:24] <antifuchs>   anyway - not there yet
[13:12:36] <antifuchs>   ?xx is a syntax error anyway, I believe (:
[13:12:40] <antifuchs>   ?\x isn't though
[13:12:50] <betawaffle>  well yea, but does it highlight correctly?
[13:13:24] <betawaffle>  lemme take a screenshot
[13:13:31] <antifuchs>   it depends on when you add the ?, actually (:
[13:13:49] <antifuchs>   hard to know what's correct for syntax errors. I'd say nasal daemon advisory is in effect
[13:15:40] <betawaffle>  well, ?x -> 1 is not a syntax error
[13:15:47] <betawaffle>  and the -> 1 are red
[13:15:48] <antifuchs>   that's right
[13:16:00] <antifuchs>   that looks correct on my end, but there are local changes that affect it
[13:16:14] <betawaffle>  what are the local changes?
[13:16:25] <antifuchs>   let me put them on a branch real quick
[13:19:27] <antifuchs>   https://github.com/antifuchs/emacs-elixir/compare/support-more-sigils is it
[13:20:22] <betawaffle>  oh hey, this looks good
[13:21:28] <antifuchs>   I think the case indentation after ?b -> x is still messed up, but that might be a better starting point (:
[13:21:40] <betawaffle>  yea
[13:22:59] <betawaffle>  so, is there a good way to reload the mode stuff when I change something?
[13:23:12] <betawaffle>  i've just been restarting emacs :(
[13:23:40] <antifuchs>   M-x load-file RET RET should get you mooooost of the way there - but some things (old/renamed variables) may still linger and cause trouble
[13:24:09] <antifuchs>   most of the time I change expressions (and reload them with C-M-x) until stuff seems to work, then restart and test on a clean slate
[13:24:15] <antifuchs>   also, ert tests have helped me a lot
[13:24:42] <betawaffle>  http://cl.ly/image/3p1b3y3S3G0T
[13:25:06] <antifuchs>   yeah, that case is totally out of whack
[13:25:10] <betawaffle>  i typed the one where my cursor is after opening the file
[13:25:25] <antifuchs>   gruuuuuump
[13:25:41] <antifuchs>   yep, I see it too
[13:25:45] <betawaffle>  what version of emacs are you on?
[13:25:55] <antifuchs>   23.3.1
[13:26:12] <antifuchs>   yeah, I'm sure syntax-propertize must be buggy there
[13:26:27] <antifuchs>   you could edebug that function & step through it
[13:26:41] <betawaffle>  how do i do that?
[13:27:43] <antifuchs>   ah, with the point inside the function elixir-syntax-propertize, M-x edebug-defun RET, then trigger something that would call it (i.e. insert chars in that elixir buffer) - and a buffer will pop up where you can step through the code with SPC
[13:27:51] <antifuchs>   press "c" to just continue
[13:27:54] <antifuchs>   edebug is amazing (:
[13:29:34] <betawaffle>  ok, so im in the debug after typing ?x SPC
[13:30:23] <antifuchs>   yay
[13:30:45] <antifuchs>   it should show you what it's doing in the other buffer (mostly, where point moves when it does its thing)
[13:32:53] <betawaffle>  what is supposed to be reseting the rest to normal color?
[13:33:24] <betawaffle>  even newlines aren't clearing it away now
[13:33:39] <antifuchs>   nothing - it's supposed to be normal already; the thing that adds the string fontification should be that last statement at the end
[13:34:30] <betawaffle>  ok, so, is it supposed to be colored like a string when I type the ?
[13:34:50] <betawaffle>  on a brand new line, i type ? and the ? is string-colored
[13:35:01] <betawaffle>  then i type the x and the x is string colored too
[13:35:11] <betawaffle>  then i type SPC and it goes into debug
[13:35:18] <antifuchs>   not necessarily, I believe. only after it finds a non-space char afterwards. Hrm.
[13:35:54] <betawaffle>  sounds like something else is causing it then
[13:36:06] <betawaffle>  your function isn't buggy!
[13:36:12] <antifuchs>   *phew!*
[13:36:18] <antifuchs>   yeah, possibly some conflicting font-lock regex maybe?
[13:36:19] <betawaffle>  but what's causing it :(!
[13:36:27] <antifuchs>   I'm suuuuper suspicious of those regexen
[13:36:33] <betawaffle>  how would I find it?
[13:36:50] <antifuchs>   that's what I've been asking myself also ):
[13:37:04] <antifuchs>   I guess regex-based font-locking isn't sustainable /:
[13:37:05] <betawaffle>  is there a way to get the whole mapping?
[13:37:41] <antifuchs>   elixir-mode-font-lock-defaults is it
[13:38:21] <antifuchs>   but together with the syntax properties, the only way to get the ultimate fontification is to describe-char the thing you are looking at
[13:38:59] <betawaffle>  oh, didn't know about describe-char
[13:39:11] <antifuchs>   it's pretty useful for this (:
[13:39:35] <betawaffle>  ok, how do i un-debug the function now?
[13:39:37] <antifuchs>   I believe there's a method for getting the ultimate char syntaxes for each char in a buffer, but not as a command (and I forget what it was)
[13:39:53] <antifuchs>   ah, press C-M-x with point inside it
[13:40:14] <antifuchs>   if there's still a debugger running, hit c to continue for the last time, and it'll not re-debug
[13:41:26] <betawaffle>  yea, the newline has string-face
[13:42:20] <antifuchs>   what the
[13:43:29] <betawaffle>  and when i type ? then a whole bunch of spaces, then try to do a new line, emacs freezes, and i have to C-g a whole bunch
[13:43:39] <antifuchs>   wow ):
[13:43:52] <antifuchs>   -that- would sound like a bug in that function
[13:48:03] <betawaffle>  ok, so more data
[13:48:08] <betawaffle>  if i type x x
[13:48:23] <betawaffle>  then go back between the xs
[13:48:37] <betawaffle>  right before the second x and type a ?
[13:48:45] <betawaffle>  the ? and x go red
[13:48:49] <betawaffle>  then if i delete the ?
[13:48:53] <betawaffle>  the x stays red
[13:49:07] <betawaffle>  if before, with the 2 white xs
[13:49:21] <betawaffle>  i go just after the first x, with a space before the second
[13:49:25] <antifuchs>   ugh, yep.
[13:49:25] <betawaffle>  and type the ?
[13:49:36] <betawaffle>  the ? goes red, but the second x stays fine
[13:49:43] <antifuchs>   that must be some incremental-syntactification thing in emacs
[13:49:58] <antifuchs>   oh god, we've built a house of cards on quicksand!
[13:50:28] <betawaffle>  oh sorry, the ? doesn't go red if im right after the first x
[13:50:33] <betawaffle>  only if i have a space after it
[13:50:41] <betawaffle>  and before the second
[13:51:05] <betawaffle>  and!
[13:51:14] <betawaffle>  if i do that thing with x ? x
[13:51:25] <betawaffle>  and start typing before the second x
[13:51:31] <betawaffle>  those chars are red like the ?
[13:51:45] <betawaffle>  but the original second x stays fine
[13:52:13] <betawaffle>  they are inheriting the text properties from the previous character
[13:52:42] <betawaffle>  but none of this happens without that function
[13:53:03] <betawaffle>  when i empty out the body, as i did in my branch, the ? things never get colored
[13:53:29] <antifuchs>   hmmm, wonder if there's some stickiness to it
[13:53:31] <antifuchs>   hm hm hm
[13:55:37] <antifuchs>   ok, I don't completely get it right now either - now just typing ?x in an elixir buffer font-locks as string forever after
[13:56:07] <betawaffle>  yea
[13:56:17] <antifuchs>   seems like I ran out of brain juice for now… can you add an issue about this? Feel free to just paste this convo (:
[13:56:27] <betawaffle>  ok

Warning: could not run smart terminal, falling back to dumb one
Eshell V6.1  (abort with ^G)

Selector: "^elixir-ert-.*$"
Passed:  0
Failed:  21 (17 unexpected)
Skipped: 0
Total:   21/21

Started at:   2014-06-29 16:25:29-0700
Finished.
Finished at:  2014-06-29 16:25:29-0700

FFfFFFffFFFFfFFFFFFFF

F elixir-ert-indents-after-empty-line

    Buffer: *Test buffer (elixir-ert-indents-after-empty-line): (Expected)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-indents-continuation-lines

    Buffer: *Test buffer (elixir-ert-indents-continuation-lines): (Expected)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-indents-continuation-lines-with-comments/2

    Buffer: *Test buffer (elixir-ert-indents-continuation-lines-with-comments/2): (Expected)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-indents-do-blocks

    Buffer: *Test buffer (elixir-ert-indents-do-blocks): (Expected)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-indents-do-blocks-after-linebreak

    Buffer: *Test buffer (elixir-ert-indents-do-blocks-after-linebreak): (Expected)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-indents-function

    Buffer: *Test buffer (elixir-ert-indents-function): (Expected)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-indents-function-calls-without-parens

    Buffer: *Test buffer (elixir-ert-indents-function-calls-without-parens): (Expected)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-indents-if

    Buffer: *Test buffer (elixir-ert-indents-if): (Expected)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-indents-if-else

    Buffer: *Test buffer (elixir-ert-indents-if-else): (Expected)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-indents-list-argument-continuation-lines-nicely

    Buffer: *Test buffer (elixir-ert-indents-list-argument-continuation-lines-nicely): (Expected)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-indents-records-correctly

    Buffer: *Test buffer (elixir-ert-indents-records-correctly): (Expected)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-indents-try

    Buffer: *Test buffer (elixir-ert-indents-try): (Expected)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-indents-try/after

    Buffer: *Test buffer (elixir-ert-indents-try/after): (Expected)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-indents-try/catch/after

    Buffer: *Test buffer (elixir-ert-indents-try/catch/after): (Expected)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-skip-backwards-over-comments

    Buffer: *Test buffer (elixir-ert-skip-backwards-over-comments)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-skip-forwards-over-comments

    Buffer: *Test buffer (elixir-ert-skip-forwards-over-comments)*<2>
    (void-function flymake-elixir-load)

F elixir-ert-treat-comments-the-same-as-whitespace-backwards

    Buffer: *Test buffer (elixir-ert-treat-comments-the-same-as-whitespace-backwards): With comment*<2>
    (void-function flymake-elixir-load)

defmodule Foo
  def bar do
    2
  end

  # last line
end

defmodule Foo
  def bar do
    2
  end

# last line
end

defmodule Unless do^
end

Debugger entered--returning value: nil
  looking-at("^[        ]*end$")
  elixir-mode-indent-line()
  indent-according-to-mode()