If I enable jade-interaction-mode in a buffer and then eval an expression, that buffer will be killed upon calling jade-quit. This is unexpected and undesirable. If no expressions are eval'd they are not killed.

I think it would be best to leave all jade buffers open, but disable jade-interaction-mode for source buffers and mark the other buffers as read-only.

Using version 20161210.804 from melpa.

I get the following when running indium-run-node:

Process indium-nodejs-process exited abnormally with code 9
node: bad option: --inspect

Running node from the terminal with node --inspect --debug-brk seems to work.

Hi.
when I try jade-connect-to-chrome I get

WS Error! on-open (void-function seq-map-indexed)

I'm using emacs 25.1.1 on Manjaro (Arch based) Linux. Any idea? Thanks

It would be awesome to have an ability to live reload the code.
Which can be implemented using https://chromedevtools.github.io/debugger-protocol-viewer/1-2/Debugger/#method-setScriptSource

Sometimes I do M-x indium-connect-to-chrome from a directory outside of the JS project. This is annoying because the debugger won't find my local files. It would be great if indium-connect-to-chrome could ask for a known workspace to bind to if it can't find any where indium-connect-to-chrome is started. That could for example be implemented with a defcustom remembering workspaces.

Sometimes, while debugging, I'm not able to inspect (C-c M-i or C-x C-e) objects. I always get Wrong type argument: arrayp, nil. Debugging Jade code reveals that jade-interaction-node-before-point calls js2-node-parent which return nil.

While in the debugger, I typed C-x M-i on a boolean variable.

Exception in websocket callback! (wrong-type-argument stringp :json-false)

1. Add a breakpoint
2. Make sure to hit it
3. The debugger prints a message in the echo area
4. Move the cursor to a expression and press C-x C-e
5. You won't see the result in the echo area as the debugger keeps printing there

When displaying the locals (l), I see

session: Object { settings: Object, redirectToPeriod: function, getPeriod: function, setPeriod: function, onPeriodChange: function }

But session contains other keys beyond these. I would like to get a ... at the end to show more keys are present but can't be displayed.

M-x indium-run-chrome
connect to a tab.
add throw "test" to a code.
On entering debugger press SPC.
After that indium will give me Can only perform operation while paused. on most of the requests.

If I choose "continue" everyting is fine though.

How to reproduce:

1. open a jade connection on a chrome tab
2. duplicate the tab on chrome
3. close the first one but keep the REPL buffer open
4. open a jade connection to the chrome tab

WS Error! on-open (wrong-type-argument stringp nil)
WS Error! on-message (wrong-type-argument integer-or-marker-p nil)

(setq websocket-debug t)
(setq websocket-callback-debug-on-error t)

I get this when I run jade-connect-to-chrome. Using chromium.

GNU Emacs 25.1.1 (x86_64-apple-darwin13.4.0, NS appkit-1265.21 Version 10.9.5 (Build 13F1911)) of 2016-09-20

Also, when I do (require 'jade) I get:

(file-error "Cannot open load file" "No such file or directory" "seq-25")

Ideas?

Sometimes Emacs stops on a line. The websocket buffer says it's because of a breakpoint but I killed all breakpoints (with C-c b k) and I don't see any anymore.

When typing C-x C-e on a variable name whose value is undefined. It happened right after the debugger opened.

Running M-x indium-run-node RET node results in

exec-path-from-shell-printf: Non-zero exit code from shell E:/eoksni-dir/portable/emacs/emacs-25.2-x86_64/libexec/emacs/25.2/x86_64-w64-mingw32/cmdproxy.exe invoked with args ("-l" "-i" "-c" "c:/Program Files (x86)/Gow/bin/printf.exe ’__RESULT\\000%s\\000%s\\000__RESULT’ \"${PATH-3ee3c256aae424afd6f79e5709ff36c2}\" \"${MANPATH-3ee3c256aae424afd6f79e5709ff36c2}\"").  Output was:
"’c:/Program’ is not recognized as an internal or external command,
operable program or batch file.

The function to drop this error was indium--setup-exec-path and it seems to be optional (by the way, to me it seems like a terrible idea to manipulate global exec-path without warning or interacting with user in any way - if I understand correctly what this function does) so I executed (setq indium-exec-path-setup t) by myself to skip this function, then node process is started fine and I am even given the repl buffer:

But when I press RET, nothing happens. I looked up that RET is bound to indium-repl-return, so I M-x'ed it with same result - nothing happens at all, no errors in *Messages*.

I don't know if it is relevant, but I start indium-run-node with this file:

/**
 * adsfasdfasdfasfd
 * @param {*asdfasf} input 
 */
function qwer(input) {
    const aaa = "asdf";
    let bbb = "vcvzx";
    for (let i = 0; i < aaa.length; ++i) {
        let mmm = 10;
        let bbb = 111111;
    }
}
qwer()

By the way, I also tried M-x indium-scratch RET 5+5 C-x C-e also with no success, just nothing happens without any error messages.

Win 10, emacs 25.2.1, started it with emacs -Q, installed indium like so:

(require 'package)
(add-to-list 'package-archives
             '("melpa" . "https://melpa.org/packages/"))
(package-initialize)

M-x package-refresh-contents
M-x package-install RET indium

I accidentally evaluated the buffer in indium-scratch, which had an infinite loop as written. The REPL was stuck printing returned values.

After quitting indium -- not sure if this was the right course of action -- I saw a bunch of errors scrolling through the minibuffer, too quickly to read.

Then Emacs froze.

Again, not sure if there is a better command to run, or if this is really a node.js issue, but it would be nice if there were some way to stop evaluation of the loop.

I'm currently trying to implement the Firefox backend. However, given the major differences between the Firefox remote debugging protocol and the Chrome one, I think it could be nice to include a todo list here.

• Implement the Firefox Remote Debugger Protocol Backend
  • Implement the protocol API (we might look at this).
  • Implement a function to query tabs information
  • Since Firefox doesn't have a per tab protocol should we use only one process for all tabs? Also each times we make a new connection, Firefox display a prompt to accept remote debugging.
  • Implement the evaluation function
  • Implement the completion function
  • Implement the debugger functions
    - [ ] Implement the breakpoints functions
    • Implement the resume function
    • Implement the step-into function
    • Implement the step-out function
    • Implement the step-over function
    • Implement the continue function
        - [ ] Implement pausing on exceptions and get them
  • Implement the property inspector
  • Sync the console logs

That's all for now.

If I enter [{"one": true}] into the JS repl, then inspect the resulting array, the buffer starts with:

Array[1]

  0: Object { one: false }

If I enter [true] then the result is printed with preview as:

Array[1] [ false ]

In both cases, inspecting the value itself renders the boolean correctly. I have tracked this to indium-webkit--description where the case for true is never matched. If I change the case matched for to the string "true" then both cases reported above are correct.

Incidentally I tried entering {"key":true} at the repl to test that directly rather than embedding it in an array, but got:

SyntaxError: Unexpected token : {  }

I installed node-nightly, but even after that I can not see any listen option. I tried --inspect and --debug but it runs the program straight away and when I try to connect from emacs to that port it gives me failure..

sandric@sandric-mac ~> node-nightly -v
v7.0.0-nightly2016082451f96dfcfc
sandric@sandric-mac ~> node -v
v5.1.0
sandric@sandric-mac ~> node-nightly --listen test.js
/usr/local/lib/node_modules/node-nightly/node-nightly/bin/node: bad option: --listen

indium   : 20170504.117
gnu emacs: 25.1.1
chrome   : 58.0.3029.96 (64-bit)

App uses Oracle Jet 3.0.0 which uses require.js and knockout.

The top index file has

<script data-main="ui/js/main" src="jet-v3.0.0/js/libs/require/require.js"></script>

I start up chrome via:

open -n "/Applications/Google Chrome.app" --args --user-data-dir=/tmp/hcChromeDebugUserDataDir/  --remote-debugging-port=9222 http://localhost:8080/ui/

I then do indium-connect-to-chrome

It asks for Host/Port/Tab

I reload the page and it hits a breakpoint that I did not set. Chrome says Paused in Emacs debugger

Emacs shows the code it is stopped at in a buffer named *JS Debugger*

I press s and see:

* returnFromLogin <http://localhost:8080/ui/ui/js/utils.js>
  Closure <http://localhost:8080/ui/ui/js/main.js>

Everything seems to work at this breakpoint (e.g., locals, following links, etc).

PROBLEM:

If I visit either of those files (or any other javascript file in the project) and explicitly set a breakpoint, the app never stops at the breakpoint.

NOTE: when I use chrome's more tools / developer tools - breakpoints work as expected.

It would be convenient if I can set an option which will prevent REPL buffer from showing after connection to a chrome process.

I see a breakpoint in the fringe. I press C-c b b to remove it and get

Breakpoint at specified location already exists.

Currently if I log multiple things like this:

console.log("this", "is", "sparta");

What I get is:

"this"
"is"
"sparta"
undefined

It would be useful to merge short output into one line separated by comma.

Another thing which is extremely useful is a file:line of the message source. Currently there is no way to distinguish between

console.log("one", "two")

and

console.log("one");
console.log("two"):
```.

I think the print string could have ... at the end when not fitting in a line.

It's useful to run some code on the emacs/js side when script was updated.
For example I want to:

(indium-eval "window.dispatchEvent(new Event('patch'))")

in the hook and then listen to it in the js app to rerende the page.

Chrome dev tools has a convenient checkbox in the network tab called disable cache.
It's a very usefull feature and it would be great to have it in the Indium.

I think it should be something like this from a user's point of view:

(setq indium-network-disable-cache t)
;; and M-x
(indium-network-toggle-cache)

https://chromedevtools.github.io/debugger-protocol-viewer/1-2/Network/#method-setCacheDisabled

Unless a workspace is used, when stepping using the debugger the source buffer flickers.

I did not found the aswer on the web so I ask here: is JSX support planned ?

I experience a lot of weird behavior on Windows - can't even press RET in the repl buffer. So should I make a detailed bug report or Windows is simply not supported so I shouldn't bother?

I'm having problems setting breakpoints.

Hi, first off - thank you very much for this great package :)

After upgrading recently, I get these error messages when I try to connect to a running chrome instance:

WS Error! on-open (void-function seq-map-indexed)
WS Error! on-message (void-function seq-map-indexed) [4 times]

It used to work before, and I don't have devtools open.

I'm on Windows 8.1 using Emacs 25.1.

There is a useful checkbox in the devtools network tab -> disable cache which disable cache when devtools is opened. It would be nice to have this feature in the Indium.

https://chromedevtools.github.io/devtools-protocol/1-2/Network/#method-setCacheDisabled

In particular, it doesn't work with the following:

for (var i = 0; i < 9; i++) {
	console.log(i);
};

When trying to evaluate the following command (from Getting up and Running):

M-x indium-connect-to-nodejs RET 127.0.0.1 RET PORT RET PATH

I get the following error:

websocket-open: make client process failed: Connection refused, :name, websocket to ws://localhost:9229//c6face84-df0e-4af2-90b1-c3ce0b54b4fd, :buffer, nil, :host, localhost, :service, 9229, :nowait, nil

This is on:

• Emacs 25.1
• MacOS 10.12.4
• node 7.2.1

any ideas what's happening? project looks great btw...want to get it working so I can convert my team from Atom...

I tried following the documentation to get Indium running, but get stuck here:

"Start a process with M-x indium-run-node."

That command doesn't seem to exist.

Any ideas?

I get the following error when I try to connect to a nodejs process

WS Error! on-close (wrong-type-argument stringp nil)
error in process filter: websocket-outer-filter: Invalid HTTP header sent
error in process filter: Invalid HTTP header sent

This is what I got in *backtrace* when I enabled toggle-debug-on-error

Debugger entered--Lisp error: (websocket-invalid-header . "Invalid HTTP status line")
  signal(websocket-invalid-header "Invalid HTTP status line")
  websocket-outer-filter([cl-struct-websocket closed nil #[257 "\301\203��\301 \210\302�\300\"\207" ["ws://127.0.0.1:9229/node" nil jade-webkit--handle-ws-open] 4 "\n\n(fn WS)"] jade-webkit--handle-ws-message jade-webkit--handle-ws-closed jade-webkit--handle-ws-error nil nil nil "ws://127.0.0.1:9229/node" nil nil #<process websocket to ws://127.0.0.1:9229/node> nil "yfSHlJgu85t/P7Pku3qBnVAV5X4=" nil] "HTTP/1.0 400 Bad Request
\nContent-Type: text/html; charset=UTF-8
\n
\nWebSockets request was expected
\n")
  #[(process output) "\303�\304\"�\305	\n\")\207" [process websocket output process-get :websocket websocket-outer-filter] 3](#<process websocket to ws://127.0.0.1:9229/node> "HTTP/1.0 400 Bad Request
\nContent-Type: text/html; charset=UTF-8
\n
\nWebSockets request was expected
\n")

I'm running emacs master build on macOS Sierra.

Hello,

I followed your instructions until the point

M-x jade-connect-to-chrome

But the command isn't available. What could be wrong?

I use Emacs 25.1 with prelude installed.

Sometimes, Chromium waits for Emacs and Emacs shows a debugger at a particular line but there is no clue why. Maybe it's an invisible breakpoint, maybe it's an error in the code. Would it be possible to be more explicit?

In case you weren't aware, there's a trademark issue here. See pugjs/pug/issues/2184 for more information...probably a wise idea to rename the project.

TL;DR: Jade was forced to switch to Pug. You'll face the same issues.

Currently simple object like

let foo = {bar: "Hello baz!"};
foo; // C-x C-e

is printed like

  __defineGetter__: function {  }
  __defineSetter__: function {  }
  __lookupGetter__: function {  }
  __lookupSetter__: function {  }
  __proto__: function {  }
  bar: "Hello baz!"
  constructor: function {  }
  hasOwnProperty: function {  }
  isPrototypeOf: function {  }
  propertyIsEnumerable: function {  }
  toLocaleString: function {  }
  toString: function {  }
  valueOf: function {  }

It would be awesome if default noisy methods could be hidden or at least moved down.

Some error messages in the REPL contain JS file names:

Error: TypeError: Cannot read property 'forEach' of undefined… {  }
From http://localhost:7002/Client/js/core/widgets/panelGroupWidget.js

I would like to be able to find this file within Emacs.

When opening a connection, Indium will often ask:

Close the current Indium connection? (y or n)

There is no clue what will happen if I answer No to this question. I propose this instead:

This requires closing the current connection. Are you sure? (y or n)

I can fix that if you agree on the problem and proposed solution.

I was trying to update a Homebrew formula for Indium, and when Travis runs buttercup -l test/test-helper.el -L . on a "Precise" Linux worker I see these test failures:

========================================
Repl output should display a prompt
error: (error "No buffer named *JS REPL*")
========================================
Repl output should display a welcome message
error: (error "No buffer named *JS REPL*")
========================================
Repl output should be able to clear all output
error: (error "No buffer named *JS REPL*")
========================================
Repl output should be able eval and print results
error: (error "No buffer named *JS REPL*")
========================================
Repl output should be able to inspect objects
error: (error "No buffer named *JS REPL*")
Ran 90 specs, 5 failed, in 28.7 seconds.

Full logs: https://travis-ci.org/dunn/homebrew-emacs/jobs/241658581#L1308

Node is installed, and the tests pass on the macOS worker. Any idea what might be going on? Thanks!

Hi there,

I can't seem to call the jade-scratch command from the repl.

Chrome's API has an setting to pause execution on uncaught exception, all exceptions, or never.

Hi. Please take a look at https://github.com/TatriX/kite-mini.el
I've added some basic sourcemap support for similiar to jade kite-mini project.
It's actually very primitive but I hope you can find it usefull because currently most of the projects use some kind of transformers like babel or typescript.

So to make it work I'm using the following:

(defun kite-mini-update-custom ()
  (interactive)
  (let ((default-directory "~/my-projects/"))
    (shell-command "npm run build" "*kite-mini-compilation*"))
  (kite-mini-call-rpc
   "Debugger.setScriptSource"
   (let ((script (--find (string-suffix-p "my-project/dist/application.js" (plist-get it :url))
                         kite-mini-rpc-scripts)))
     (plist-put script :source-map (cons 'lazy-source-map "~/www/my-project/dist/application.js.map"))
     (list :scriptId (plist-get script :id)
           :scriptSource (kite-mini--get-string-from-file "~/www/my-project/dist/application.js")))
   (lambda (result)
     (kite-mini-send-eval
      "window.dispatchEvent(new CustomEvent('change', {detail: {filename: 'application.js'}}))"
      (lambda (result) (message "Kite update done"))))))

Then in the buffer with kite-mini-mode I can use C-c C-k to compile the project and update it on the fly and still get sourcemaps, so I can see actual code lines in the repl buffer.

(use-package indium
  :ensure t
  :init
  (dolist (hook '(js-mode-hook
                  js2-mode-hook
                  js3-mode-hook
                  ))
    (add-hook hook #'indium-interaction-mode))
  :config
  (setq indium-chrome-executable "google-chrome-unstable")
  )

I enabled Indium like upper. And found that [C-x C-e] is bind to indium-eval-last-node.

It already happened a few times that Emacs freezes for around 10 seconds consuming 100% CPU after I politely asked M-x indium-connect-to-chrome. After the delay, the REPL gets refreshed and everything is back to normal.

Here's the output

Leaving directory ‘c:/Users/Julian/.emacs.d/elpa/indium-20170719.643’

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-backend.el at Thu Jul 20 00:54:25 2017
Entering directory ‘c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/’

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-breakpoint.el at Thu Jul 20 00:54:25 2017
indium-breakpoint.el:33:1:Error: Cannot open load file: No such file or directory, sourcemap

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-chrome.el at Thu Jul 20 00:54:25 2017
indium-chrome.el:36:1:Error: Cannot open load file: No such file or directory, sourcemap

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-debugger-frames.el at Thu Jul 20 00:54:25 2017
indium-debugger-frames.el:27:1:Error: Cannot open load file: No such file or directory, sourcemap

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-debugger-litable.el at Thu Jul 20 00:54:25 2017

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-debugger-locals.el at Thu Jul 20 00:54:26 2017

In end of data:
indium-debugger-locals.el:108:1:Warning: the following functions are not known to be defined:
    indium-debugger-get-current-scopes,
    indium-debugger-get-scopes-properties

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-debugger.el at Thu Jul 20 00:54:26 2017
indium-debugger.el:34:1:Error: Cannot open load file: No such file or directory, sourcemap

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-faces.el at Thu Jul 20 00:54:26 2017

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-inspector.el at Thu Jul 20 00:54:26 2017

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-interaction.el at Thu Jul 20 00:54:26 2017
indium-interaction.el:37:1:Error: Cannot open load file: No such file or directory, sourcemap

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-nodejs.el at Thu Jul 20 00:54:26 2017
indium-nodejs.el:45:1:Error: Cannot open load file: No such file or directory, sourcemap

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-render.el at Thu Jul 20 00:54:26 2017

In end of data:
indium-render.el:201:1:Warning: the function ‘seq-map-indexed’ is not known to
    be defined.

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-repl.el at Thu Jul 20 00:54:26 2017

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-scratch.el at Thu Jul 20 00:54:26 2017
indium-scratch.el:27:1:Error: Cannot open load file: No such file or directory, sourcemap

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-script.el at Thu Jul 20 00:54:26 2017
indium-script.el:32:1:Error: Cannot open load file: No such file or directory, sourcemap

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-structs.el at Thu Jul 20 00:54:26 2017

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-webkit.el at Thu Jul 20 00:54:27 2017
indium-webkit.el:40:1:Error: Cannot open load file: No such file or directory, sourcemap

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium-workspace.el at Thu Jul 20 00:54:27 2017

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/indium.el at Thu Jul 20 00:54:27 2017
indium.el:51:1:Error: Cannot open load file: No such file or directory, sourcemap

Compiling file c:/Users/Julian/.emacs.d/elpa/indium-20170719.643/run-lint.el at Thu Jul 20 00:54:27 2017