sys/js: initial support of RIOT <-> javascript API mapping #7796

kaspar030 · 2017-10-23T15:35:11Z

This pull requests adds the foundation for accessing RIOT's API through javascript.

Currently available are timers and saul.

It uses the event system introduced in #6000 in order to asynchronously pass e.g., interrupts into javascript (whose state cannot be modified while it is running).

This is WIP as we need to agree on the concept, and licenses and documentation still missing.
~~Waiting for #7794, #6000.~~

kaspar030 · 2017-10-23T21:12:32Z

needs #7794, #6000

kaspar030 · 2017-10-23T21:14:23Z

needs #6000

miri64 · 2017-11-14T13:13:03Z

examples/javascript/main.js

@@ -1 +1,12 @@
-print("Hello from JerryScript!");
+print("Hello from JerryScript!!!!");


Real important change ^^!!!!

miri64 · 2017-11-14T13:16:53Z

Let me know if this isn't WIP anymore.

kaspar030 · 2017-11-14T13:18:25Z

Let me know if this isn't WIP anymore.

Actually, this is more "Request for comments". I don't know enough javascript to say if the API makes sense (from a javascript conventions perspective)...

jia200x · 2018-02-06T14:46:53Z

sys/js/timer.c

+
+    js_xtimer_t *js_xtimer = (js_xtimer_t *) native_p;
+    xtimer_remove(&js_xtimer->xtimer);
+    free(js_xtimer);


@kaspar030 isn't jerryscript freeing this js_xtimer block?

Well, jerryscript garbage-collects the javascript timer object. If that happens to be a "native" object, it calls that native object's free_cb, which happens to be this function.

yeah, I see. But why is it required to call free()? Shouldn't the js memory manager free the whole block?

No, when the native timer object is created (using js_object_native_create()), the js_xtimer_t is malloc'ed outside jerryscript's heap (so we still have control over it). Thus our binding code both allocateds and frees the object.

ok, I see ;)

jia200x · 2018-02-06T15:09:36Z

Very nice! I like it.
Just some minor comments:

Do you think writing lib.js functions directly in RIOT (and map them as js functions) could help to decrease memory consumption? Could also help to reuse code in case Lua or other scripting languages get into mainstream
Do you think it's possible to use TLSF instead of jerryscript heap implementation? So all processes can share the same heap.

On the other side, at some point could be nice to have a tiny RIOT-JS framework focused in IoT scenarios. But for now, I think the idea is ok and would be nice to have it merged. After that, it's easier to work in the a better API.

kaspar030 · 2018-02-06T15:24:01Z

Thanks a lot for the feedback!

Do you think writing lib.js functions directly in RIOT (and map them as js functions) could help to decrease memory consumption? Could also help to reuse code in case Lua or other scripting languages get into mainstream

Sure! It was just very convenient to write lib.js in Javascript. :)
It might not be worth the hassle for some things, but I guess we'll have to find out.

Do you think it's possible to use TLSF instead of jerryscript heap implementation? So all processes can share the same heap.

Difficult. Jerryscript uses non-trivial pointer compression in order to keep memory usage down, so this is not a matter of just mapping "js_malloc" to "tlsf_malloc", and might turn out counter-productive.

(TLSF also eats >3kB just for house keeping in its default configuration).

jia200x · 2018-02-07T10:32:22Z

Sure! It was just very convenient to write lib.js in Javascript. :) It might not be worth the hassle for some things, but I guess we'll have to find out.

I will give it a try and let you know ;)

Difficult. Jerryscript uses non-trivial pointer compression in order to keep memory usage down, so this is not a matter of just mapping "js_malloc" to "tlsf_malloc", and might turn out counter-productive.

I see

jia200x · 2018-02-07T17:47:49Z

@kaspar030 if you like the idea in #8533, there could be a make js command at some point :)

jia200x · 2018-03-20T15:18:05Z

just noticed something weird in native:

Executing lib.js...
Hello from lib.js
Executing main.js...
Hello from JerryScript!!!!
js_run(): Script Error!
Entering event loop...
^C
native: exiting

There's is a Script Error here.
Same code on samr21-xpro:

2018-03-20 16:16:52,922 - INFO # main(): This is RIOT! (Version: 2018.01-devel-280-gdf8da-archlinux-riot.js)
2018-03-20 16:16:52,926 - INFO # You are running RIOT on a(n) samr21-xpro board.
2018-03-20 16:16:52,929 - INFO # This board features a(n) samd21 MCU.
2018-03-20 16:16:52,932 - INFO # Initializing jerryscript...
2018-03-20 16:16:52,935 - INFO # Executing lib.js...
2018-03-20 16:16:52,965 - INFO # Hello from lib.js
2018-03-20 16:16:52,968 - INFO # Executing main.js...
2018-03-20 16:16:52,975 - INFO # Hello from JerryScript!!!!
2018-03-20 16:16:52,981 - INFO # Entering event loop...

jia200x · 2018-03-21T10:12:21Z

@kaspar030 have you checked JERRY_SYSTEM_ALLOCATOR? It forces js to use system's malloc, as seen here

jia200x · 2018-03-21T10:16:05Z

sys/js/callback.c

+
+static void js_callback_cleanup(js_callback_t *js_callback)
+{
+    DEBUG("release %lu\n", callback);


sed 's/callback/js_callback'

(and add ->callback.object?)

emmanuelsearch · 2018-05-04T11:58:34Z

what is WIP here?
what are steps towards having some version of this in master?

tcschmidt · 2018-06-06T12:09:07Z

@emmanuelsearch It looks like there are changes requested. Needs resolution, I suppose.

emmanuelsearch · 2018-08-21T12:08:04Z

examples/javascript/main.js

+
+led.write(1);
+
+button.on_threshold(0, function(){led.write(0); return true;}, saul.op.EQ);


@kaspar030 how about showing some timer function too, for documentation purposes ;)
Something more or less like below, i.e. blink if you press the button?

blink = function () { while (1) { value += 1 % 2; this.led.write(value); t = timer.setTimeout(this.blink, 500000); } } button.on_threshold(0, this.blink}, saul.op.EQ);

jia200x · 2018-09-26T11:06:17Z

@kaspar030 may I open a rework of the event handler part, and then gradually add the remaining parts of the lib.js file?

jia200x · 2018-10-04T07:35:56Z

@kaspar030 may I open a rework of the event handler part, and then gradually add the remaining parts of the lib.js file?

ping @kaspar030

kaspar030 · 2018-10-04T09:00:01Z

@kaspar030 may I open a rework of the event handler part

Sure. What's wrong with this?

jia200x · 2018-10-04T09:02:29Z

Sure. What's wrong with this?

Nothing wrong. Just add Doxygen labels, patch pointed out errors and reduce the scope to speed up the merge. What do you think?

stale · 2019-08-10T05:08:15Z

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. If you want me to ignore this issue, please mark it with the "State: don't stale" label. Thank you for your contributions.

miri64 · 2020-06-23T12:13:20Z

@kaspar030 any plans to work on this in the near future?

Teufelchen1 · 2024-01-23T10:08:43Z

@kaspar030 👀

kaspar030 · 2024-01-23T10:46:38Z

I won't work on this anymore.

kaspar030 assigned emmanuelsearch Oct 23, 2017

kaspar030 added 17 commits November 7, 2017 14:34

sys: js: initial commit of javascript support module

069fcc5

examples/javascript: make use of js module

4fa2be0

sys/js: rename timer.setCallback() -> timer.setTimeout()

a409510

sys/js: add initial saul support

001938e

examples/javascript: split js into lib.js and main.js

8e361ef

lib.js: add saul.sample()

456bd0c

sys/js: keep list of native referecences to js objects

8894ecc

sys/js: refactor callback handling

8f2eaf5

sys/js: add saul.get_by_name() support

04c1d55

sys/js: lib.js: change setInterval() abort condindition to "!=false"

2e03d16

sys/js: lib.js: change on_threshold to event semantics

92895fa

sys/js: optionally execute event after js_shutdown()

1abe9a5

sys/js: don't set js_event_queue in js_init()

6c693d6

sys/js: refactor native reference mechanics

0ce89f1

sys/js: add some helper functions

639b17e

sys/js/callback: fix rebase error

ecb753a

sys/js: export js_add_external_handler()

df8dae9

kaspar030 removed the State: waiting for other PR State: The PR requires another PR to be merged first label Nov 7, 2017

kaspar030 force-pushed the riot.js branch from d671571 to df8dae9 Compare November 7, 2017 13:34

emmanuelsearch requested a review from miri64 November 14, 2017 13:09

miri64 reviewed Nov 14, 2017

View reviewed changes

danpetry mentioned this pull request Jan 25, 2018

Lua support #8434

Closed

jia200x reviewed Feb 6, 2018

View reviewed changes

jia200x mentioned this pull request Feb 7, 2018

core: add DEFAULT_APP_ENTRY to select main() function #8533

Closed

jia200x requested changes Mar 21, 2018

View reviewed changes

danpetry assigned danpetry and unassigned danpetry May 28, 2018

emmanuelsearch reviewed Aug 21, 2018

View reviewed changes

stale bot added the State: stale State: The issue / PR has no activity for >185 days label Aug 10, 2019

kaspar030 added the State: don't stale State: Tell state-bot to ignore this issue label Aug 19, 2019

stale bot removed the State: stale State: The issue / PR has no activity for >185 days label Aug 19, 2019

MrKevinWeiss added this to the Release 2021.07 milestone Jun 21, 2021

MrKevinWeiss removed this from the Release 2021.07 milestone Jul 15, 2021

kaspar030 closed this Jan 23, 2024

		@@ -1 +1,12 @@
		print("Hello from JerryScript!");
		print("Hello from JerryScript!!!!");


		led.write(1);

		button.on_threshold(0, function(){led.write(0); return true;}, saul.op.EQ);

sys/js: initial support of RIOT <-> javascript API mapping #7796

sys/js: initial support of RIOT <-> javascript API mapping #7796

Uh oh!

Conversation

kaspar030 commented Oct 23, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

kaspar030 commented Oct 23, 2017

Uh oh!

kaspar030 commented Oct 23, 2017

Uh oh!

Choose a reason for hiding this comment

Uh oh!

miri64 commented Nov 14, 2017

Uh oh!

kaspar030 commented Nov 14, 2017

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jia200x commented Feb 6, 2018

Uh oh!

kaspar030 commented Feb 6, 2018

Uh oh!

jia200x commented Feb 7, 2018

Uh oh!

jia200x commented Feb 7, 2018

Uh oh!

jia200x commented Mar 20, 2018

Uh oh!

jia200x commented Mar 21, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

emmanuelsearch commented May 4, 2018

Uh oh!

tcschmidt commented Jun 6, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jia200x commented Sep 26, 2018

Uh oh!

jia200x commented Oct 4, 2018

Uh oh!

kaspar030 commented Oct 4, 2018

Uh oh!

jia200x commented Oct 4, 2018

Uh oh!

stale bot commented Aug 10, 2019

Uh oh!

miri64 commented Jun 23, 2020

Uh oh!

Teufelchen1 commented Jan 23, 2024

Uh oh!

kaspar030 commented Jan 23, 2024

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

kaspar030 commented Oct 23, 2017 •

edited

Loading