You know what? Let me humbly apologize for the noise here.
I first tried this with the ubuntu-default version (1.4.9) _without_ the
HEADLESS=1 option, and that failed at runtime due to a slew of things not
being defined (i vaguely recall things like "print", "read", etc, not being
defined), which lead me to use the HEADLESS=1 option. In this old version,
that got things to not fail at runtime pretty quickly, but exhibited the
behaviour i described above (not getting in to loop).
After upgrading to head (1.9.0), i failed to retry the test without the
HEADLESS=1 option. Had I done so (as i have now), i'd have seen that
everything works perfectly under nodejs as expected.
Thanks, and sorry again for the noise here. =)
-dave-
On Tue, Jan 28, 2014 at 11:40 PM, Dave Nicponski
<[email protected]>wrote:
>
>
>
> On Tue, Jan 28, 2014 at 10:45 PM, Dave Nicponski <[email protected]
> > wrote:
>
>> Hello,
>>
>> Thanks for the response.
>>
>>
>> On Tue, Jan 28, 2014 at 7:28 PM, Alon Zakai <[email protected]> wrote:
>>
>>> The HEADLESS option is not much tested, and used for testing purposes
>>> mainly. But I think that you need to call window.runEventLoop() manually.
>>> Even then I imagine there would need to be some minor fixes.
>>>
>>>
>> Hrm... OK i can do this, but doing this conditionally for nodejs from
>> C++ code seems strange.
>>
>
> Yeah, this doesn't work out-of-the-box.
> Adding:
>
> emscripten_run_script("window.runEventLoop()");
>
> at the end of main fails because headless.js 's runEventLoop triggers:
>
> exception thrown: TypeError: Object function Date() { [native code] } has
> no method 'realNow',TypeError: Object function Date() { [native code] } has
> no method 'realNow'
> at Object.window.runEventLoop
> (/home/daven/src/jsminer/jsminer/jsminer/build/trash/poc/headless_main_loop.js:5671:24)
>
> [...]
>
> I'll go with a different approach for the moment, and try to loop back to
> help fixing this when i have some free time. This is a pretty simple repro
> / use case which should probably be supported, and i can't imagine this
> test case would be too hard to fix.
>
> Thanks for looking in to this.
>
> -dave-
>
>
>
>>
>>> If there is a use case that we should support here though, we can
>>> consider that - why do you want to run headless in node?
>>>
>>
>> Basically, i've got a C++ application, some of which is compiled to
>> javascript and run in a client browser. However, much of what's being
>> executed isn't browser-related. Being able to factor out the browser
>> element for code testing is the idea. It's easier to make assertions about
>> code correctness when the application components are run in nodejs headless
>> than when they are executed in a browser, or so it seems to me.
>>
>> I don't want an entire browser to be emulated in headless mode obviously,
>> but basic infrastructure functionality like "reasonably correctly emulate
>> main event loops and web worker threads in headless mode" would be very
>> useful to have.
>>
>> Thanks,
>> -dave-
>>
>>
>>
>>> - Alon
>>>
>>>
>>>
>>> On Tue, Jan 28, 2014 at 1:51 PM, Dave Nicponski <
>>> [email protected]> wrote:
>>>
>>>> Sorry for the noob question, but here goes.
>>>>
>>>> # Source code
>>>> $ cat headless_main_loop.cc
>>>>
>>>> #include <stdio.h>
>>>>
>>>> #include "emscripten/emscripten.h"
>>>>
>>>> int foo = 0;
>>>>
>>>> void dummy_loop() {
>>>> printf("dummy loop: %d\n", foo++);
>>>> if (foo >= 5)
>>>> emscripten_cancel_main_loop();
>>>> }
>>>>
>>>> int main(int argc, char **argv) {
>>>> printf("Start\n");
>>>>
>>>> emscripten_set_main_loop(dummy_loop, 0, true); // XXX
>>>>
>>>> printf("Main done.\n");
>>>> return 0;
>>>> }
>>>>
>>>> # Build command for nodejs
>>>> $ ~/src/emscripen/emscripten/emcc -Qunused-arguments -Wall -Wextra
>>>> -Werror -Wno-unused-parameter -fexceptions -std=c++11 -x c++ -iquote ./src
>>>> -iquote ~/src/emscripen/emscripten/system/include/ -I . -DNDEBUG -O2
>>>> src/trash/poc/headless_main_loop.cc -o
>>>> build/trash/poc/headless_main_loop.js -s HEADLESS=1
>>>>
>>>> # Try with nodejs
>>>> $ (dev-test_pb) nodejs trash/poc/headless_main_loop.js
>>>> Start
>>>>
>>>> # Change line XXX from true to false (infinite loop), rebuild
>>>> # Try again
>>>> $ (dev-test_pb) nodejs trash/poc/headless_main_loop.js
>>>> Start
>>>> Main done.
>>>>
>>>>
>>>> In both cases, it works as expected if i compile to html and run in a
>>>> browser (obviously, without the HEADLESS=1 option).
>>>> Specifically, I see this in the console:
>>>>
>>>> Start
>>>> Main done.
>>>> dummy loop: 0
>>>> dummy loop: 1
>>>> dummy loop: 2
>>>> dummy loop: 3
>>>> dummy loop: 4
>>>>
>>>> ...which is exactly what i'd expect from both node and browser.
>>>>
>>>> Am I crazy, or shouldn't this work?
>>>>
>>>> Note that this is both with the default ubuntu install (1.4.9, old) and
>>>> the version i'm actually using from head (1.9.0 (commit
>>>> 7dc8c2ff08e46c9d9a88ba44bf221a404eeb1e5e))
>>>>
>>>> Thanks,
>>>> -dave-
>>>>
>>>> --
>>>> You received this message because you are subscribed to the Google
>>>> Groups "emscripten-discuss" group.
>>>> To unsubscribe from this group and stop receiving emails from it, send
>>>> an email to [email protected].
>>>>
>>>> For more options, visit https://groups.google.com/groups/opt_out.
>>>>
>>>
>>> --
>>> You received this message because you are subscribed to a topic in the
>>> Google Groups "emscripten-discuss" group.
>>> To unsubscribe from this topic, visit
>>> https://groups.google.com/d/topic/emscripten-discuss/2-rxY9ob3aY/unsubscribe
>>> .
>>> To unsubscribe from this group and all its topics, send an email to
>>> [email protected].
>>> For more options, visit https://groups.google.com/groups/opt_out.
>>>
>>
>>
>
--
You received this message because you are subscribed to the Google Groups
"emscripten-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.
