Issue with Elixir highlighting #707

whatyouhide · 2015-01-11T21:25:28Z

This HTML snippet:

<html>
  <head>
    <link rel="stylesheet" href="highlight/styles/default.css">
    <script src="highlight/highlight.pack.js"></script>
    <script>hljs.initHighlightingOnLoad();</script>
  </head>

  <body>
      <pre><code class="elixir">defmodule Foo do
  def foo, do: :foo
  def bar, do: :bar
end</code></pre>
  </body>
</html>

produces this buggy result:

because of the def f, do: # code syntax (as an alternative to the usual do/end block, which works just fine). I think @knewter took care of the Elixir highlighting, so if he has a quick fix that would be great! Otherwise, I'll try to start working on it as soon as I can. I never contributed to highlight.js though, so it could take me a while to get familiar with it 😞.

The text was updated successfully, but these errors were encountered:

Sannis · 2015-01-12T13:01:29Z

Am I right that with this syntax function body ends with new line? (\n)?

whatyouhide · 2015-01-12T14:08:54Z

@Sannis I don't think so. The do: is followed by an expression. Some examples of valid Elixir code:

# This reproduces the bug
def foo, do:
  3 + 5

or

# This is highlighted perfectly fine
def foo, do: 3 +
             4 +
             5

tonini · 2015-01-28T05:32:10Z

@Sannis @whatyouhide What is the status about that? we really would like to fix this because we love to use it in ex_doc :-)

Before I'm starting looking into it by myself, maybe you guys can have a quick look before me. :)

@knewter I know you have a lot to do, but if you got a few minutes, please have look. ❤️ :-)

whatyouhide · 2015-01-30T15:05:56Z

Any news on this? As @tonini said, we would really really love to fix this since it's screwing up a lot of examples in Elixir's documentation (as well as in the documentation of every library using ex_doc, i.e., every library 😄).

knewter · 2015-01-30T15:16:12Z

Urk, what's the situation with the tests these days? src/test.html isn't
a thing anymore, and test/index.html doesn't really trigger anything
(don't know that it should). Is there documentation on how to add new
language support re: testing/verifying? Because I was going to fix this
today and realize it's changed too much for me to get it done in 15 minutes
or so :(

On Fri, Jan 30, 2015 at 9:06 AM, Andrea Leopardi notifications@github.com
wrote:

Any news on this? As @tonini https://github.com/tonini said, we would
really really love to fix this since it's screwing up a lot of examples
in Elixir's documentation (as well as in the documentation of every library
using ex_doc, i.e., every library [image: 😄]).

—
Reply to this email directly or view it on GitHub
#707 (comment)
.

Josh Adams
CTO | isotope|eleven http://www.isotope11.com
cell 215-3957
work 476-8671 x201

isagalaev · 2015-01-30T17:07:15Z

@knewter the docs on testing are here: http://highlightjs.readthedocs.org/en/latest/building-testing.html (in short, we now have tools/developer.html to play with markup).

I'm not sure when any one of us will be able to look into it (for me it's probably next Tuesday), so feel free to pick it up.

knewter · 2015-02-04T14:57:27Z

I'm looking at it now, we'll see if anything comes of it /cc @tonini @whatyouhide

knewter · 2015-02-04T15:08:40Z

so @tonini @whatyouhide @josevalim I have a few questions about this one. In this case, isn't do an atom in an optionlist? That's the core problem here - it's matching "thing: other" such that thing is an atom in the list. That's matching do. I can change the regex such that do isn't matched, in which case it would become a keyword...but then you would always get that behaviour if you used do: in an optionlist as well, which is wrong (but better?) Basically, I need to do this work when @josevalim or someone is around to help explain to me my misunderstandings like this in the parser. Sadly the highlight.js syntax checking is in no way presently easy to 'port' from the actual tokenizer in elixir, primarily due to my lack of experience here (only parsers I've really written that were complicated were PEGs)

whatyouhide · 2015-02-04T15:12:29Z

@knewter I think the problem is not that do: isn't highlighted as a keyword (it's not, it's just an atom and that's fine). The problem is that all newlines after a do: get messed up.

Note that this is valid Elixir syntax:

%{a:
  3}

or

[do:
  something]

knewter · 2015-02-04T15:15:45Z

Aha I missed the point entirely :)

isagalaev · 2015-02-19T19:58:46Z

Am I getting this right that in the case of

def foo, do: :foo

The comma , actually signals the end of the function header and that do: is something else entirely than in the case of

def foo do
  # ...
end

?

eksperimental · 2015-03-07T04:18:13Z

@isagalaev what comes after , do: should be treated the same way as what's in between do and end, with the exception that that , do: cannot accept new lines, and ; and ( ) are used instead.
I don't know if that answers your question.

def foo, do: (IO.puts "hello"; IO.puts " world")

produces internally the exact same code as:

def foo do
  IO.puts "hello"
  IO.puts " world"
end

isagalaev · 2015-03-18T19:11:38Z

Ooookay, I think I've fixed it, finally:

I even had to introduce a new functionality in the core for that :-) Which I hope is going to be needed for another bug elsewhere.

josevalim · 2015-03-18T19:14:01Z

Awesome, thank you!

eksperimental · 2015-03-18T19:15:05Z

much appreciated @isagalaev !

whatyouhide · 2015-03-19T00:07:11Z

Awesome @isagalaev, thanks so much! ❤️

knewter · 2015-03-19T09:59:14Z

This is great. I can stop feeling terrible now! ❤️

Sannis added the bug label Jan 12, 2015

whatyouhide mentioned this issue Feb 3, 2015

update highlightjs to support more languages elixir-lang/ex_doc#185

Merged

isagalaev added this to the 8.5 milestone Feb 19, 2015

eksperimental mentioned this issue Mar 7, 2015

(elixir) interpolation, sigils and / functions #730

Closed

3 tasks

isagalaev closed this as completed in eaab7b3 Mar 18, 2015

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Issue with Elixir highlighting #707

Issue with Elixir highlighting #707

whatyouhide commented Jan 11, 2015

Sannis commented Jan 12, 2015

whatyouhide commented Jan 12, 2015

tonini commented Jan 28, 2015

whatyouhide commented Jan 30, 2015

knewter commented Jan 30, 2015

isagalaev commented Jan 30, 2015

knewter commented Feb 4, 2015

knewter commented Feb 4, 2015

whatyouhide commented Feb 4, 2015

knewter commented Feb 4, 2015 via email

isagalaev commented Feb 19, 2015

eksperimental commented Mar 7, 2015

isagalaev commented Mar 18, 2015

josevalim commented Mar 18, 2015

eksperimental commented Mar 18, 2015

whatyouhide commented Mar 19, 2015

knewter commented Mar 19, 2015 via email

Issue with Elixir highlighting #707

Issue with Elixir highlighting #707

Comments

whatyouhide commented Jan 11, 2015

Sannis commented Jan 12, 2015

whatyouhide commented Jan 12, 2015

tonini commented Jan 28, 2015

whatyouhide commented Jan 30, 2015

knewter commented Jan 30, 2015

isagalaev commented Jan 30, 2015

knewter commented Feb 4, 2015

knewter commented Feb 4, 2015

whatyouhide commented Feb 4, 2015

knewter commented Feb 4, 2015 via email

isagalaev commented Feb 19, 2015

eksperimental commented Mar 7, 2015

isagalaev commented Mar 18, 2015

josevalim commented Mar 18, 2015

eksperimental commented Mar 18, 2015

whatyouhide commented Mar 19, 2015

knewter commented Mar 19, 2015 via email