sebastiw / edts Goto Github PK

In the pre-edts era I had my own functions to compile from the shell, which allowed me to compile with the option_all option whenever I needed, without including the infamous -compile(export_all) in the source.

Now edts keeps compiling all the time which makes that quite less useful. Any thoughts on how to instruct edts to compile with export_all (temporarily) for a module in a nice way? (or an alternative solution not implying modifications in EDTS)

... which is weird 'cause edts-find-local-function works just fine.

M-. for functions like process_chunk(State, <<Buffer/binary, Chunk/binary>>)
throws "ferl-search-function: function process_chunk/3 not found" error

Hello.

I'm new in Erlang. Trying setup Erlang environment with Emacs 24.2 and EDTS on Ubuntu 12.10. I install Erlang R16B.

When I open any "erl" file i get error message in Emacs:

EDTS[errors]: Could not register node "ERLANG"

Before that when i install EDTS from GIT and execute make command in shell, get error message:

lib/mochiweb/src/mochiweb_request.erl:6: parameterized modules are no longer supported

Any call to edts_code:refresh() with a system running mnesia causes a crash when mnesia_late_loader module is purged and reloaded:

=SUPERVISOR REPORT==== 12-Apr-2013::23:25:16 ===
     Supervisor: {local,mnesia_kernel_sup}
     Context:    child_terminated
     Reason:     killed
     Offender:   [{pid,<0.207.0>},
                  {name,mnesia_late_loader},
                  {mfargs,{mnesia_late_loader,start,[]}},
                  {restart_type,permanent},
                  {shutdown,3000},
                  {child_type,worker}]

which brings down the VM. We need a way to only reload modules specific to the project. Unfortunately the mnesia directory is not marked as sticky (http://www.erlang.org/doc/man/code.html#is_sticky-1) so the code is purged resulting in the above error

I have another issue, sometimes (maybe always, at least pretty often) when I save the file
my cursor jumps to the end of buffer after a while.

And it's really irritating since I hit save after each line I type..bad habit.
Since it happens after a while I wonder if have something to with result presentation from xref checks.

Emacs 23.3.1 (ubuntu 12.04)

When trying to edts-find-source-under-point while pointing onto one of my project modules (in this case it is l4m_db_bend):

(edts@damian)1> 09:39:07.528 [error] l4m@damian error loading module l4m_db_bend: {error,nofile}
(edts@damian)1> 09:39:07.600 [error] Error in remote call edts_code:get_function_info/3 on  4m@damian: 
{'EXIT',{{badmatch,error},[{edts_code,get_function_info,3,[{file,"src/edts_code.erl"},{line,157}]},{rpc,'-handle_call_call/6-fun-0-',5,[{file,"rpc.erl"},{line,203}]}]}}

which is result of not having project lib-dirs paths added to the code server path (which I figured out following the error message). I did not find any logs or messages reporting some difficulties about them when the project node is started (its console is fresh and clean ;) Maybe it's related to my project structure:

<root-dir>
    |
    +-node---------+
    |              |
    |              +--<otp-app1> (my app # 1)
    |              |
    |              +--<otp-app2> (my app # 2)
    |              |
    |              .
    |              .    (more apps ...)
    |              .
    |              +--.deps (rebar project dependencies)
    |
    +-lib (some C/C++ libraries here)

(I have "node" and ".deps" directories under my root dir (yes, there is DOT in the deps directory name) ;))

If all tests pass after saving a module of unit tests and I use C-x b to change to another module (another buffer?), the same lines remain highlighted. E.g. an ?_Assert on line 130 in the test module will mean that line 130 in the other module will be green after switching.

That seems like a reasonable way of opening a shell to mess with the code without using the project node that edts itself is using for other interactions

Missing closing bracket in the following line:

(otp-path   . "~/otp/r15b02"

Cheers,
syl20bnr

Faced with the problem trying to open source of the standard module "proplists" with "M-." command. The path of the src is supposed to be the one which is stored in beam. But on my pc erlang is not build from src, but is installed from the packages instead. And so the source code is located not in the place where edts is looking for.

Maybe it's better to look for src in otp dir if "otp-path" variable is set?

Two problems with darker themes such as zenburn:
*edts-show-doc-under-point seems to be using default foreground color in box, but the background is always white, so a light foreground color is more or less invisible.
*Documentation in Auto-complete seems to be using the defined background for Popup-tip but not the foreground, and is instead using the default foreground color.

I'm using the newer deftheme and not Color-theme, if that could have anything to do with it.

It would be nice if autocompletion would complete a function name with arity if you're typing inside an export statement.

If I have the following code structure (with rebar):

lib/
    /app1
        /src/file1.erl
        /ebin/file1.beam
src/
    *.erl
ebin/

and I open file1.erl and modify it, edts_code will compile and load it, however it will store the beam file in root ebin rather that lib/app1/ebin. The relevant code is in edts_code

get_compile_outdir(File, Opts) ->
  case proplists:get_value(outdir, Opts) of
    undefined -> filename_to_outdir(File);
    OutDir    ->
      case filelib:is_dir(OutDir) of
        true  -> OutDir;
        false -> filename_to_outdir(File)
      end
  end.

my module does have outdir set in the compile options:

[{options,[{outdir,"ebin"},
           debug_info,
           {d,'DEV_ONLY'},
           {i,"include"}]},...

and it is a valid directory, however it is the wrong ebin directory. Could we assume that if OutDir =:= "ebin" it actually means filename_to_outdir(File)? I believe that will result in finding the correct ebin directory if code is compiled form src/.erl or lib/app1/src/.erl.

Of course this could be the wrong behaviour. Is the correct way to work this project structure to have all beam code in the root ebin folder?

Steps to reproduce:

1. Type opening double quote (").
2. Type any backslash-escaped character (\n, \t, or any other).

Actual behavior:
Useless auto-completion on function names appears.

Expected behavior:
Nothing happens. It is usually meaningless to include function names in string literals.

Here's a few warnings before hand and the actual error. I'm running arch linux with the latest Erlang.

/home/liam/.emacs.d/liam-custom/edts/lib/webmachine/src/webmachine_decision_core.erl:715: crypto:md5/1 is deprecated and will be removed in in a future release; use crypto:hash/2
/home/liam/.emacs.d/liam-custom/edts/lib/webmachine/src/webmachine_decision_core.erl:719: crypto:md5_init/0 is deprecated and will be removed in in a future release; use crypto:hash_init/1
/home/liam/.emacs.d/liam-custom/edts/lib/webmachine/src/webmachine_decision_core.erl:725: crypto:md5_final/1 is deprecated and will be removed in in a future release; use crypto:hash_final/2
/home/liam/.emacs.d/liam-custom/edts/lib/webmachine/src/webmachine_decision_core.erl:725: crypto:md5_update/2 is deprecated and will be removed in in a future release; use crypto:hash_update/3
/home/liam/.emacs.d/liam-custom/edts/lib/webmachine/src/webmachine_decision_core.erl:727: crypto:md5_update/2 is deprecated and will be removed in in a future release; use crypto:hash_update/3
ERROR: compile failed while processing /home/liam/.emacs.d/liam-custom/edts/lib/webmachine: rebar_abort

Hello.

I'm new in Erlang. Trying setup Erlang environment with Emacs 24.2 and EDTS on Ubuntu 12.10. I install Erlang R16B.

When I open any "erl" file i get error message in Emacs:

EDTS[errors]: Could not register node "ERLANG"

Before that when i install EDTS from GIT and execute make command in shell, get error message:

lib/mochiweb/src/mochiweb_request.erl:6: parameterized modules are no longer supported

Hi Thomas,

Thanks for a great project! I really enjoy flymake and the code completion, nice work. To get it to work the user has to add the lines from "setup on UNIX" from http://www.erlang.org/doc/apps/tools/erlang_mode_chapter.html to .emacs. That might be good to point out since some of them aren't needed for everyday usage of the erlang mode.

(Fyi, to be able to get the erl-find-source-under-point et al. in edts to work for modules from erlang/otp I hade to build erlang/otp from the git repo. The compile module info contained bad paths when building from the source packages on erlang.org.)

After building erlang from the git repo I can use erl-find-source-under-point to go to eg gen_server. To be able to continue down into gen.erl from gen_server.erl l I had to add the erlang repo as one of my projects in the edts config.

I saw that you want to add a nice way to see the documentation. One way to do that is to add the erlang man pages to the man path in /etc/manpath.config and use the man-function in the standard erlang mode.

Node name always otp-usr@localhost

At some point, edts-refactor-extract-function stopped working correctly. Now when you create a function, free variables are not added as arguments to it, which makes edts-refactor-extract-function pretty much useless.

Frankly speeking I think that settings should not always be dependent on the project. Instead it would be better to set some general settings and to give a chance to change them if needed locally to the project. For example default otp_dir for all projects and etc.(like it is done now with edts-man-set-root)

The thing that confused me was that autocomplition is not working at all for standard modules (expect built in functions) if the project dir is not set properly. I mean sometimes you just download or clone the project on a github, and just want to modify it quickly, or experiment with it. You don't want to specify any specific project settings for it, but autocomplition would not work for it even for standatd modules until the module you are edditing is situated inside some project dir. It's confusing.

Also some thoughts about rebar which I mention in twitter. It seems to me that now it is one of the basic tools which many erlang programmers use. There is much information about dependencies on other applications, prefered otp_version and other stuff which can be picked from rebar.config files. So sometimes it's simplier to specify just one main rebar.config file, than specifiing lib-dir directories and other settings manually

Is there any package manager support for edts? Otherwise, adding support for melpa might be easy and help faster installation.

Question rather than an issue - I've got some projects where everything is working very well, and others (of a seemingly similar structure) where the include paths aren't being set correctly, so records etc can't be found and hence there are compiler errors when saving.

So the question is how does edts determine the include path, and where is this being stored / set? My elisp skills are next-to non-existant, but I'll happily hack around inside the erlang if you can give me some pointers where to look...

I have the following trace when opening an erlang file:

[yas] Loading for `erlang-mode', just-in-time: (yas--load-directory-1 /home/sbenner/.emacs.d/elpa/yasnippet-20121225.430/snippets/erlang-mode (quote erlang-mode) (quote (text-mode)))!
[yas] Loading compiled snippets from /home/sbenner/.emacs.d/elpa/yasnippet-20121225.430/snippets/erlang-mode
[yas] Loading for `erlang-mode', just-in-time: (yas--load-directory-1 /home/sbenner/.emacs.d/snippets/erlang-mode (quote erlang-mode) (quote (text-mode)))!
[yas] Loading compiled snippets from /home/sbenner/.emacs.d/snippets/erlang-mode
EDTS [debug]: Waiting to register node, (retries 5)
EDTS [debug]: Registering node erltests
EDTS [debug]: Sending async POST-request to http://localhost:4587/nodes/erltests?project_root=/home/sbenner/dev/projects/erlang&lib_dirs=lib
Making url-show-status local to  *http proxy.mdc.xxxxxx.org:3128* while let-bound!
EDTS [debug]: Reply (200 OK) received for request to http://localhost:4587/nodes/erltests?project_root=/home/sbenner/dev/projects/erlang&lib_dirs=lib
EDTS [error]: Unexpected reply: (200 OK)
EDTS [debug]: Waiting to register node, (retries 4)
EDTS [debug]: Registering node erltests
EDTS [debug]: Sending async POST-request to http://localhost:4587/nodes/erltests?project_root=/home/sbenner/dev/projects/erlang&lib_dirs=lib
EDTS [debug]: Reply (503 Service Unavailable) received for request to http://localhost:4587/nodes/erltests?project_root=/home/sbenner/dev/projects/erlang&lib_dirs=lib
EDTS [error]: Unexpected reply: (503 Service Unavailable)
EDTS [debug]: Waiting to register node, (retries 3)
EDTS [debug]: Registering node erltests
EDTS [debug]: Sending async POST-request to http://localhost:4587/nodes/erltests?project_root=/home/sbenner/dev/projects/erlang&lib_dirs=lib
EDTS [debug]: Reply (200 OK) received for request to http://localhost:4587/nodes/erltests?project_root=/home/sbenner/dev/projects/erlang&lib_dirs=lib
EDTS [error]: Unexpected reply: (200 OK)
EDTS [debug]: Waiting to register node, (retries 2)
EDTS [debug]: Registering node erltests
EDTS [debug]: Sending async POST-request to http://localhost:4587/nodes/erltests?project_root=/home/sbenner/dev/projects/erlang&lib_dirs=lib
EDTS [debug]: Reply (200 OK) received for request to http://localhost:4587/nodes/erltests?project_root=/home/sbenner/dev/projects/erlang&lib_dirs=lib
EDTS [error]: Unexpected reply: (200 OK)
EDTS [debug]: Waiting to register node, (retries 1)
EDTS [debug]: Registering node erltests
EDTS [debug]: Sending async POST-request to http://localhost:4587/nodes/erltests?project_root=/home/sbenner/dev/projects/erlang&lib_dirs=lib
EDTS [debug]: Reply (200 OK) received for request to http://localhost:4587/nodes/erltests?project_root=/home/sbenner/dev/projects/erlang&lib_dirs=lib
EDTS [error]: Unexpected reply: (200 OK)
EDTS [debug]: Waiting to register node, (retries 0)
EDTS [error]: Could not register node 'erltests'

You could see that:

• I changed the host name to localhost (0 does not work in my environment)
• At the first http call there is something about the corporate proxy
• The corporate proxy always returned 200 which could have side effects elsewhere.

Do you have some ideas about this ?

Thank you.

Cheers,
syl20bnr

if a calling function is defined in several cases, one should end up in the place that the function is called, not in the first branch of the calling function.

It would be great if you could move your cursor to a behaviour usage declaration line and be able to navigate into the module defining the behaviour.

Ex for clarity:

Cursor on
-behaviour(foo).

M-. would navigate into foo.erl

Cheers!

I am pretty sure it was introduced by 7aab9a7

A call like this is now reported as an error:

init_per_group(Group, Config) ->
  ?MODULE:Group({init, Config}).

Error reason is:

Call to undefined function fred_reserve_amount_SUITE:'$F_EXPR'/1

edts-show-doc-under-point is amazing!

It'd be even cooler if it did line breaks where the doc has line breaks and wrapped nicely when the doc is just very wide.

Even I should be able to do this but I don't expect to have time soon so here's the idea until I have time.

Currently yasnippet is loaded with erlang-mode but EDTS seems to not make it available through auto-complete.

When I set the variable url-proxy-services, EDTS won't start the main process.

Edts doesn't start project for following dir structure

project
├── erlang/
│   ├── apps/
│   ├── deps/
│   ├── docs/
│   ├── priv/
│   ├── rel/
│   ├── configure.sh*
│   ├── .edts
│   ├── .gitignore
│   ├── Makefile
│   ├── readme
│   └── rebar.config
├── .git/
│   ├── branches/
│   ├── hooks/
│   ├── info/
│   ├── logs/
│   ├── objects/
│   ├── refs/
│   ├── config
│   ├── description
│   ├── HEAD
│   ├── index
│   └── packed-refs
├── node/
│   └── .gitignore
├── static/
│   ├── css/
│   ├── images/
│   └── js/
├── www/
│   └── .gitignore
└── Makefile

I had to patch edts with the following to get it up and running,
i.e. I had old beam files lying around in my ERL_LIBS path.
/Dan

diff --git a/lib/edts/src/edts_code.erl b/lib/edts/src/edts_code.erl
index 9809ad4..0c6d016 100644
--- a/lib/edts/src/edts_code.erl
+++ b/lib/edts/src/edts_code.erl
@@ -378,12 +378,22 @@ ensure_loaded(File) ->
M = list_to_atom(filename:basename(LoadFileName)),
Loaded =
case code:is_loaded(M) of

•  {file, File} -> false;
  

•  {file, _}    -> code:purge(M),
  

•                  {module, M} = code:load_abs(LoadFileName),
  

•                  true;
  

•  false        -> {module, M} = code:load_abs(LoadFileName),
  

•                  true
  

•  {file, File}  -> false;
  

•  {file, Orig} ->
  

•   code:purge(M),
  

•   case code:load_abs(LoadFileName) of
  

•     {module, M} -> true;
  

•     {error, Error} ->
  

•       io:format("Loading Error: ~p ~s => ~s~n", [Error, Orig, File]),
  

•       false
  

•   end;
  

•  false ->
  

•   case code:load_abs(LoadFileName) of
  

•     {module, M} -> true;
  

•     {error, Reason} ->
  

•       io:format("Loading Error: ~p ~s~n", [Reason, File]),
  

•       false
  

•   end
  

  end,
  {Loaded, M}.

I think I have problems with my .edts file because I'm experiencing all kinds of weirdness with emacs/edts , I'd like some help with an .edts file that suits.
I'm not running any of the releases, just cloned github.

The directory structure I have is

lib/app1
lib/app2

lib/a/b/app3
lib/a/b/app4

otp/installed/bin/erl
otp/installed/bin/erlc ... etc

All apps above are std erlang "apps"

edts cannot start the "project node" that has to be started outside and it's not possible to define a :start-command that works in all required circumstances, different flags to start command are used at various places, thus edts cannot start my project. The "erl" in otp/installed/bin/erl is the "erl" found first in the PATH

I tried various incantations of .edts , I'm not sure it's interesting here to attempt
to describe what happened,

Thanks

/klacke

I faced we an error using last branch from master, while trying to inspect standard module from otp:

13:26:35.442 [error] 'edts-0@vd' initialization crashed with error:{badmatch,{badrpc,{'EXIT',{{badmatch,{error,sticky_directory}},[{edts_code,ensure_loaded,1,[{file,"/home/me/.emacs.d/edts/lib/edts/src/edts_code.erl"},{line,384}]},{lists,map,2,[{file,"lists.erl"},{line,1173}]},{lists,map,2,[{file,"lists.erl"},{line,1173}]},{edts_code,load_all_in_dir,1,[{file,"/home/me/.emacs.d/edts/lib/edts/src/edts_code.erl"},{line,374}]},{lists,flatmap,2,[{file,"lists.erl"},{line,1184}]},{edts_code,load_all,0,[{file,"/home/me/.emacs.d/edts/lib/edts/src/edts_code.erl"},{line,71}]},{rpc,'-handle_call_call/6-fun-0-',5,[{file,"rpc.erl"},{line,203}]}]}}}}
Stacktrace:

Stacktrace:
[{edts_server,do_init_node,3,[{file,"src/edts_server.erl"},{line,289}]},{edts_server,handle_call,3,[{file,"src/edts_server.erl"},{line,177}]},{gen_server,handle_msg,5,[{file,"gen_server.erl"},{line,588}]},{proc_lib,init_p_do_apply,3,[{file,"proc_lib.erl"},{line,227}]}]

And it seems to me, that I know why.
When edts opens the module from Otp, it also find all modules in that directory and calls enshure_loaded in edts_code for them. And as the path is relative, but not absolute, it tries to purge them all and load due to pattern {file, _}. So we got {error,sticky_directory}.

It's good idea to reload user modules, but we need to ignore {error,sticky_directory} on otp modules.

System info

• Ubuntu 13.04 32bit
• GNU Emacs 24.3.50.1 (i686-pc-linux-gnu, GTK+ Version 3.6.4)
  of 2013-05-10 on actinium, modified by Debian
• R16B
• edts just pulled (however the error had appeared before)

Error description

In *edts* buffer the following appeared:

(edts@damian)1> 09:17:03.544 [error] webmachine error: path="/code/parsed_expressions/mfa"
{error,{error,function_clause,[{lists,map,[#Fun<edts_resource_parse.0.94976884>,null],
[{file,"lists.erl"},{line,1223}]},{edts_resource_parse,to_json,2,
[{file,"src/edts_resource_parse.erl"},{line,64}]},{webmachine_resource,resource_call,3,
[{file,"src/webmachine_resource.erl"},{line,183}]},{webmachine_resource,do,3,[{file,"src/webmachine_resource.erl"},{line,141}]},
{webmachine_decision_core,resource_call,1,[{file,"src/webmachine_decision_core.erl"},
{line,48}]},{webmachine_decision_core,decision,1,
[{file,"src/webmachine_decision_core.erl"},{line,546}]},
{webmachine_decision_core,handle_request,2,[{file,"src/webmachine_decision_core.erl"},
{line,33}]},{webmachine_mochiweb,loop,2,[{file,"src/webmachine_mochiweb.erl"},
{line,69}]}]}}

and:

09:32:43.227 [error] webmachine error: path="/nodes/l4m/dialyzer_analysis"
{error,{error,function_clause,[{edts_resource_dialyzer,'-to_json/2-lc$^0/1-0-',
[{dialyzer_error,
[78,111,116,32,97,32,114,101,103,117,108,97,114,32,102,105,108,101,58,32,
"/home/damian/src/l4m/node/node/.deps/edown/ebin/edown_doclet.beam",10]}],
[{file,"src/edts_resource_dialyzer.erl"},{line,96}]},{edts_resource_dialyzer,to_json,2,
[{file,"src/edts_resource_dialyzer.erl"},{line,96}]},
{webmachine_resource,resource_call,3,[{file,"src/webmachine_resource.erl"},
{line,183}]},{webmachine_resource,do,3,[{file,"src/webmachine_resource.erl"},
{line,141}]},{webmachine_decision_core,resource_call,1,
[{file,"src/webmachine_decision_core.erl"},{line,48}]},
{webmachine_decision_core,decision,1,[{file,"src/webmachine_decision_core.erl"},
{line,546}]},{webmachine_decision_core,handle_request,2,
[{file,"src/webmachine_decision_core.erl"},{line,33}]},{webmachine_mochiweb,loop,2,
[{file,"src/webmachine_mochiweb.erl"},{line,69}]}]}}

lib/mochiweb/src/mochiweb_request.erl:6: parameterized modules are no longer supported
You know you waited for this.
Get rid of webmachine, I guess?

P.S. Oh, just noticed you're using tagged branch of webmachine, see if its fixed in master by now, it should be. There's pull request to webmachine fixing this and its merged.

when trying to go to lists:reverse edts fails as it tries to find lists:reverse/3.

-module(stupid_macro).
-export([foo/0]).
-define(why_on_earth(A, B, C), []).
foo() ->
  lists:reverse(?why_on_earth(
                   %% )
                   foo, bar, baz)).

It seems all the components are needed to reproduce the bug, The comment is needed and must contain a closing parenthesis, and trying to do the same with a 2 argument macro works. With a opening parenthesis I get a Scan error: "Unbalanced parenthesis" so I bet something goes funny when EDTS tries to parse the macro body

If it doesn't ring a bell to you I could debug it eventually :)

Not shure wherether it is a bug with edts or something is missing in my erlang configuration, but flymake will always highlight the paths to include files if they are not relative.

So this is the edts config:

(add-to-list 'load-path "~/.emacs.d/edts")
(require 'edts-start)

(edts-man-set-root "/usr/lib/erlang/man")

(setq edts-projects
      '(
        (;; 
         (name       . "feed_adapter")
         (root       . "/media/files/dev/erlang-feed_adapter")
         (node-name  . "feed_adapter")
         (lib-dirs   . ("deps"))
         (otp-path   . "/usr/lib/erlang")
        ))

Where erlang-feed_adapter is a true otp application with rebar.config inside.

I also copied stuff from here http://blog.erlware.org/2012/05/15/getting-flymake-and-rebar-to-play-nice/ in my .emacs, but it was long ago before I switched to edts from distelm and it used to work for some time =)

So this examples don't work properly:

-include("some.hrl").
-include_lib("some_lib/include/other.hrl").

But this works fine:

-include("../include/some.hrl").

May be you know how to fix it?
My emacs version is 24.2.1 and I use edts from master.

Would be nice to be able to navigate type specs as we do with functions and macros

Example of error message in *edts* console:

(edts@damian)1> 09:28:18.562 [error] Error in remote call edts_code:get_function_info/3 on l4m@damian: 
{'EXIT',{undef,[{edts_code,get_function_info,[gen_server,reply,2],[]},{rpc,'-handle_call_call/6-fun-0-',5,[{file,"rpc.erl"},{line,203}]}]}}

This happens both on Emacs 24.1 and 24.3 (ubuntu 12.04 and 12.10, OTP R15B02). The bug affects almost everything valuable in the tool.

I would like to be able to not run xref when I save a file.

A project directory with a "." (such as foo.bar) in it is causing an error when saving a file:

"EDTS [error]: Unexpected reply: (500 Internal Server Error)

Renaming the project directory to remove the "." fixes the problem

When EDTS fails to register node for some reason, it synchronously waits to retry several times with the following message:

EDTS [error]: Failed to register node, nodename. Retrying (N attempts left).

While it's retrying, emacs is frozen. To make matters worse, it will try register node again when some other file is opened right afterwards and freeze emacs again.
It would be nice to improve usability by somehow making this operation asynchronous and not freezing emacs.

See http://www.gnu.org/software/emacs/manual/html_node/elisp/Library-Headers.html

.. and nothing happens since then.

*Messages* sayes:

call-interactively: Wrong number of arguments: 
(lambda (result) "Runs xref-checks for current buffer on the node related to that
buffer's project." (interactive) (if (string= "erl" (file-name-extension (buffer-file-name))) 
(progn (edts-face-remove-overlays (quote ("edts-code-xref"))) (if (and edts-code-xref-checks 
(not (eq result (quote error)))) (progn (let ((module (ferl-get-module))) 
(edts-get-module-xref-analysis-async module edts-code-xref-checks 
(function edts-code-handle-xref-analysis-result) (current-buffer)))))))), 0

When autocompleting a function exported from another module through the list shown after pressing colon, EDTS will sometimes insert a number of blank lines at the end of the buffer equal (or close enough) to the number of available functions.

For instance, take a module foo with exported functions a, b, c. Typing "foo:" and waiting for the pop-up list is going to insert 3 blank lines at the end of the buffer.

This issue doesn't seem to occur all the time, and I haven't been able to check if auto completing a module name will trigger it as well.

When you expect to see docs for a function you get wrong and probably uncutted text, like here: http://en.zimagez.com/zimage/screenshot-21082013-131252.php

Auto completion and syntax highlighting in the emacs erlang shell would be really cool.

How to reproduce:

Do C-c C-d w on any function definition, you will get this error:

Symbol's function definition is void: edts-mfa-at-point