I found that it is because emcc does not pass sanity checks. Because latest
version is used fastcomp. When i compiled llvm-fastcom this error is
disappeared. But for now i have this message on emcc -v:
./emcc -v
emcc (Emscripten GCC-like replacement + linker emulating GNU ld ) 1.13.2
clang version
I think it is will be useful feature. Usually native code has not important
functions that cannot be easily ported, for example playing videos between
game levels or video/sound effects such as fadeout and fadein usually
implemented as uninterrupted loop. For simplicity i just disable this
Finally, I integrate regal gl into emscripten. You should call
RegalMakeCurrent((RegalSystemContext)1); to associate emscripten gl context
with regal gl context. Like this:
glutInit(, argv);
glutInitDisplayMode(GLUT_SINGLE | GLUT_RGB);
glutInitWindowPosition(0, 0);
No, this bug affects only Safari browser (IOS 10). Where I can learn more
about JS VM deopts?
As I understand in emscripten we use typed array as HEAP, so we allocate
this array on start and then use it (without grow/shrink). Why GC is
relevant in this case?
2016-11-10 1:47 GMT+03:00 Alon Zakai
Man, you're my hero!
2016-11-10 5:39 GMT+03:00 Danoon :
> I ported Dosbox to Java, so most of my knowledge of the emulation comes
> from the Dosbox project.
>
> After I got Dos games to work in Java, my goal became to get Win95/98 era
> games to run in a browser using Java.
We checked that completely disabling rendering does not helps, so I think
that problem is not FULL_ES emulation. Is there are other way to profile js
code, without builtin browser profiler.
2016-11-17 22:01 GMT+03:00 Alon Zakai :
> I'm not a GL expert so I'm not sure. @juj
perfomance
is near 30-40fps, but after some moment it increased to 50-60fps without
any "magic".
2016-11-19 0:17 GMT+03:00 Александр Гурьянов <caiiiy...@gmail.com>:
> We checked that completely disabling rendering does not helps, so I think
> that problem is not FULL_ES e
I use normalizeAsm and denormalizeAsm and all works, but only on O1 mode,
but not in O2 and O3. My impl:
function pgo(ast) {
var PATCHED_FUNCS = set();
traverseGeneratedFunctions(ast, function(func) {
var sig = func[1];
if (sig in PATCHED_FUNCS) return null;
var asmData =
Yes I mean optimizations.
"-O3"
Like "-O2", but with additional JavaScript optimizations that can
take a significant amount of compilation time and/or are relatively
new.
Note: This differs from "-O2" only during the bitcode to
JavaScript (final link and JavaScript generation)
I compile with -O0 -s SAFE_HEAP=1 -s STB_IMAGE=1 and does not have any
error in logs
2017-03-23 16:07 GMT+07:00 caiiiycuk :
> Hi! I've porting another game on cocos2d with emscripten, and for js
> version game hangs after two levels with out of memory error.
> memoryprofiler
Hi, what os do you use? I have same problem on Safari 10 in MacOs. In my
case switching from O3 to O2 helps.
2017-03-23 15:25 GMT+07:00 Alon Faraj :
> Im using emscripten and chrome 57.
> My application is workign as expected but when I try to use a
> specific feature that
For example I have class hierarchy:
Object <- MyObject
Object <- MenuItem
class MyObject: public Object {
void handler(MenuItem* item);
}
Is it safe to cast void(MyObject::*)(MenuItem*) to
void(Object::*)(Object*)? Or I should use MULATE_FUNCTION_POINTER_CASTS=1
in this case too?
I thinking
Works! Thank you :)
2017-04-20 23:21 GMT+07:00 Alon Zakai :
> I think you're missing a call to emscripten_sleep() in that loop,
>
> https://github.com/kripken/emscripten/wiki/Emterpreter#
> emterpreter-async-run-synchronous-code
>
> We don't automatically pause and resume
TypeError: asm.js type error: fixnum is not a subtype of double? or floatish
2017-08-01 15:25 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
> Hmm, after that manipultaions code can't be compiled in asm.js mode (
>
> 2017-08-01 14:06 GMT+07:00 caiiiycuk <caiiiy...@gmail.
Hmm, after that manipultaions code can't be compiled in asm.js mode (
2017-08-01 14:06 GMT+07:00 caiiiycuk :
> I think I found another way, I wrote uglifyjs script that remove Math_fround
> from everywhere except list of functions. This script just find calls to
> Math_fround
with flag
-Wl,--unresolved-symbols=ignore-all and it works fine, asm.js also
works.
2017-08-09 19:36 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
> The source is not public, but I think I can share result js if it's useful.
>
> I will try to build WebAssembly and check.
>
> 20
> You could try adding a single .cpp file to the build with that symbol
>> defined, or add a __EMSCRIPTEN__ specific path somewhere that defines it, or
>> hook into Wasm instantiation in the html file and add that entry in the
>> import object.
>>
>> torstai 10. elo
s that a singleton
> pointer?
>
> 2017-08-10 6:55 GMT+03:00 Александр Гурьянов <caiiiy...@gmail.com>:
>> Update on compile time:
>> -O2 - ~60 sec
>> -O3 - ~7 sec
>>
>> Also I try to build project with WASM=1, and it built successfully,
>> but I can't
at http://localhost/bin.js?timestamp=1502360564253:5664:4
at
If this abort() is unexpected, build with -s ASSERTIONS=1 which can
give more information.
2017-08-10 17:32 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
> emcc (Emscripten gcc/clang-like replacement) 1
I've checked, INLINING_LIMIT=1 wan't helps. May be I can see
something in intermediate object file (bc)?
2017-07-11 20:04 GMT+07:00 Jukka Jylänki :
> There have been reports about -llvm-opts being buggy in the past,
> especially in combination with inlining. You could try to
Just to be sure, I use emscripten-fastcomp/bin/clang, and
emscripten-fastcomp/bin/clang++ to compile my sources
2017-07-13 13:13 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
> *compiled
>
> 2017-07-13 13:12 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
>>
*compiled
2017-07-13 13:12 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
> Wow, great :) I don't know that I can compile native builds with
> fastcomp compiler.
>
> Btw, I've compilted my sources with same emscripten-fastcomp compiler,
> with -O3 flag and game just r
tively with the exact same
> compiler (i.e., use clang/clang++ from emscripten's fastcomp) and see if the
> problem occurs there. A pure LLVM issue would show up there, and you can
> then see if it's a known LLVM issue.
>
> On Wed, Jul 12, 2017 at 10:37 AM, Александр Гурьянов <caiiiy.
with ll files, and my source file.
https://drive.google.com/file/d/0B28AXjYMDNZsRi01Z2trUy1aU0U/view?usp=sharing
2017-07-12 12:16 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
> I've checked, INLINING_LIMIT=1 wan't helps. May be I can see
> something in intermediate object file (bc
Wow, great! Sorry for my inattention. Btw, I've checked that
profiling-funcs emits to mach code. I used O2 --emit-symbol-map, and
then restore names with uglifyjs, there are results:
-O2 --emit-symbol-map: 9.3Mb
-O2 --emit-symbol-map + uglifyjs: 9.8Mb
-O2 --profiling-funcs: 14.5Mb
2017-07-24
I tried this, unfortunately when I replace all floats with double even
native version does not work, physics is broken. Maybe I made mistake
somewhere, I will try again.
2017-07-27 22:53 GMT+07:00 Alon Zakai :
> There isn't an option for that currently. But in your specific
The source is not public, but I think I can share result js if it's useful.
I will try to build WebAssembly and check.
2017-08-09 19:04 GMT+07:00 Jukka Jylänki :
> This looks like a performance bug in the browser. It may be possible
> to reduce the compilation time, but without
Sorry, this is uglifyjs bug
2017-08-01 15:26 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
> TypeError: asm.js type error: fixnum is not a subtype of double? or floatish
>
>
> 2017-08-01 15:25 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
>> Hmm, after
subtype of intish
>>
>> 2017-08-22 11:33 GMT+07:00 Александр Гурьянов <caii...@gmail.com>:
>> > Roll a ball project:
>> > https://www.assetstore.unity3d.com/en/#!/content/77198
>> >
>> > 2017-08-22 11:32 GMT+07:00 Александр Гурьянов <caii...@gm
> >> As I see using SFE breaks asm.js validity? On my another project sfe
>> >> produce js less on 500Kb (from 6.5Mb), but FF says:
>> >> TypeError: asm.js type error: double is not a subtype of intish
>> >>
>> >> 2017-08-22 11:33 GMT+07:00 Александ
at Object.Module._extensions..js (module.js:417:10)
at Module.load (module.js:344:32)
at Function.Module._load (module.js:301:12)
at Function.Module.runMain (module.js:442:10)
2017-08-23 0:16 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
> While trying to reproduce a
Sorry, missed error line from prev message:
TypeError: asm.js type error: double is not a subtype of intish
2017-08-23 13:42 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
> Without -g2 I have this error:
> g[a+4>>2]=-$|0;
>
> In function:
> function Bg(a,b,$){a=a
TypeError: asm.js type error: double is not a subtype of intish
2017-08-23 13:42 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
> Sorry, missed error line from prev message:
> TypeError: asm.js type error: double is not a subtype of intish
>
> 2017-08-23 13:42 GMT+07:00
>>
>>> >> Great. I will try to reproduce asm.js bug. Btw, if you build
>>> >> roll-a-ball project in Unity, then output will be in Unity format
>>> >> (also js, but with Unity loader, Unity staff). You can find emscripten
>>> >> output in P
I prefer to use SDL_mixer_1.2 it works fast. Also it supports ogg, mp3
panning and dinstance.
2017-08-29 21:55 GMT+07:00 Goran Milovanovic :
> I would like to compile a project that's built on top of SDL2 libraries. One
> of those libraries is SDL2_mixer, which seems to
s,
>> Arnab
>>
>> On Tuesday, August 22, 2017 at 11:43:09 PM UTC-7, caiiiycuk wrote:
>>>
>>> TypeError: asm.js type error: double is not a subtype of intish
>>>
>>> 2017-08-23 13:42 GMT+07:00 Александр Гурьянов <caii...@gmail.com>:
>&g
Roll a ball project: https://www.assetstore.unity3d.com/en/#!/content/77198
2017-08-22 11:32 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
> Hi. I tried to use SFE on Roll-a-ball demo project, and get same
> results. Unity uses --separate-asm flag to build asm.js. Resulting js
cuk/sdk/Similar-Function-Elimination/src/similar_function_eliminator.js:35:20)
Maybe I doing something wrong. Can you try? Resulting js is here (build.asm.js):
https://drive.google.com/file/d/0B28AXjYMDNZscDFlZTdWMGMzQkU/view?usp=sharing
Thank you!
2017-08-22 1:35 GMT+07:00 Александр Г
)
at Module._compile (module.js:410:26)
at Object.Module._extensions..js (module.js:417:10)
at Module.load (module.js:344:32)
at Function.Module._load (module.js:301:12)
at Function.Module.runMain (module.js:442:10)
2017-08-20 12:44 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.
As I see using SFE breaks asm.js validity? On my another project sfe
produce js less on 500Kb (from 6.5Mb), but FF says:
TypeError: asm.js type error: double is not a subtype of intish
2017-08-22 11:33 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
> Roll a ball proje
I think that error can be reporudced on any Unity project, I will
check and back) Btw, thanks for nice stuff!
2017-08-22 1:27 GMT+07:00 arnab choudhury :
> Thanks Alon :).
>
> Caiiycuk, is the codebase you are trying to run SFE on available publicly?
> Can you post the
Hi guys! Sorry, I don't understand from now SFE step is builtin into
emscripten, or I should use it as post step after compilation? I want
to try it on my projects.
2017-08-19 22:12 GMT+07:00 Charles Vaughn :
> The biggest difference is SFE uses helper functions, since we need
Also, you defenitely should add this
https://github.com/achoudhury85/Similar-Function-Elimination to your
article. For big codebase it's work really cool, usually it reduce
code by 20% (in my case 2-3Mb).
2017-10-24 0:49 GMT+07:00 Floh :
> I don't know why the .mem files have
Thank you :) I see that asm.js do not produce much overhead. Floh,
thank you for great article, btw why mem file have mem extension, not
mem.js? For my projects I rename this file to mem.js (and replace all
occurrences in js) to get auto gzip features, etc.
2017-10-19 0:03 GMT+07:00 Charles
Hi. This discussion is very interesting, but what do you think about
solving this problem from other side. I talk about solution like
proguard for java, is it possible to create such tool that drop unused
functions from generated js? Maybe it's very hard to collect what
functions from emscripten
But if there is no functions to delete, how generated js can be
shrinked, I don't understand. BTW for unity project, I write custom
tool for collecting functions that used in runtime, and near 70% of
function is never called. I can reduce size of unity project from 26Mb
to 12Mb, without having in
I've use linux sockets with emscripten. They work as in native code except
one thing - they should be async. So if you app use async socket then it
should work out the box. Ofcourse you should wrap server socket with
websockify it's work like a charm.
5 дек. 2017 г. 8:21 ДП пользователь "Mark
Hmm, I have an project (Dune 2) that uses emterpreter. The Dune 2
project, contains a lot of cycles like:
while (true) {
// do some animation stuff
// usually: shift pallete + render
}
or
while (true) {
// do read input
// for example: save game dialog (enter file name)
}
With
I can confirm that I also have same issue with some games on safari
browser. Game runs slow until you minimize/maximize browser or attach
debugger. I made a lot of profiling, but does not have success, profiler
shows that overall performance changes after "magic" not single function.
My theory is
(or did when I last tried). Could you please check if chrome is an
> issue also on your devices?
>
> 2017-11-08 6:59 GMT+02:00 Александр Гурьянов <caiiiy...@gmail.com>:
>>
>> I can confirm that I also have same issue with some games on safari
>> browser. Game
return false;
}
}
return;
}
abort();
2017-12-06 12:21 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
> I've use common internet address like '192.168.0.104:7014' and connect
> to socket in this way:
>
> int flags;
> if ((flags = fcntl(socketId, F_GETFL, 0)
I've use common internet address like '192.168.0.104:7014' and connect
to socket in this way:
int flags;
if ((flags = fcntl(socketId, F_GETFL, 0)) < 0) {
abort();
}
if (fcntl(socketId, F_SETFL, flags | O_NONBLOCK) < 0) {
abort();
}
if (socketId >= 0) {
if (connect(socketId,
Hi. I have experience with android development, and can answer from
android side.
- is it possible to control the mobile onscreen keyboard (show, hide
and get key events) without an HTML input element?
Unfortunately no. It can be made only in custom webview, if you create
android app with webview
t;
>> I'm not sure offhand if there's a way to get a long long result from EM_ASM
>> but double will work there.
>>
>> -- brion
>>
>> On Tue, Feb 20, 2018 at 1:16 AM Александр Гурьянов <caiiiy...@gmail.com>
>> wrote:
>>>
>>> Hi,
Hi, I found that simplest function getCurrentTimeMs() work different
in browser environment (tested on origin/incoming). Code is simple:
#include
#include
int main(int, char**) {
static struct timeval tp;
gettimeofday(, 0);
printf("%ld %ld %ld\n", tp.tv_sec, tp.tv_usec, tp.tv_sec *
Sorry, this is my fault. Native & browser output is identical.
2018-02-16 18:34 GMT+07:00 Александр Гурьянов <caiiiy...@gmail.com>:
> Hi, I am working on project that uses custom String class, this string
> class have methods:
>
> std::string String::getStdString() cons
Hi. I have recently switched to new version of emcc, and faced with
the problem that I should add functions to -s
EXTRA_EXPORTED_RUNTIME_METHODS to use them through
Module['function_name']
Most of errors that I have came from file_packager.py, I mean I used
file_packager.py to create preloadable
the file packager is run standalone. (We already had
> this in the docs and at runtime in -O0 and ASSERTIONS builds, but even
> better to print it from the tool itself, where it is most likely to be
> seen.)
>
> On Tue, Dec 26, 2017 at 1:17 AM, Александр Гурьянов <caiiiy...@g
What browser do you use? For me Chrome is dramatically slow while
compiling wasm, asm.js is faster at least 3 times.
2018-04-08 1:19 GMT+07:00 Goran Milovanovic :
> After all the files are downloaded, it takes ~5 seconds for my program to
> actually start. Since my cpu
Wow, is it mean that if changes will be made, then we can have a
little performance boost if use float32 samples?
2018-04-05 10:36 GMT+07:00 Alon Zakai :
> Yeah, looks like the current code just supports U8 and S16. Probably because
> the first things ported with it used
Hmm, looks link time optimizations id broken --llvm-lto 2 produce not
worked js, but game stuck forever on loading. If I found little test
case then I post it.
Anyway, is it correct that I can run my "SAFE" optimization on step
(2) and as result get correct js with no performance lost against
O3 is used to determine how
> much to optimize in other things than LLVM (asm.js/wasm optimizations).
>
>
> On Mon, Mar 26, 2018 at 6:44 PM, Александр Гурьянов <caiiiy...@gmail.com>
> wrote:
>>
>> Hmm, looks link time optimizations id broken --llvm-lto
+1, I've also use some hacks to compress FS better then DXT.
2018-04-04 0:46 GMT+07:00 Floh :
> Huh, didn't know this exists :) In my case it wouldn't make sense to use
> though because I'm not using the emscripten filesystem but instead load data
> directly.
>
> -Floh.
>
>
> On
I used freetype (USE_FREETYPE=1) many times. It works fine. But if you
start your project from scratch I suggest to use browser capabilities to
render font it will be much much lighter. Anyway it is hard to tell why
your code did not work, please share your code. Did you try FULL_ES=2? As I
Hi. I tried it couple of times. Problem is in undefined behaviour my
code works correctly most of time, but sometimes unpredictable erros
are happen. I decide to not use them, because it is impossible to
prove that output is working correctly.
ср, 10 окт. 2018 г. в 0:09, Corey Lucier :
>
> We
2. I'm not sure about your second
> problem, but perhaps it is related.
>
> On Mon, Jan 14, 2019 at 2:32 AM Александр Гурьянов
> wrote:
>>
>> For bigger project, in EMCC_DEBUG=1 mode (only in debug mode) I had this
>> error:
>> ```
>> emscripte
Hi guys. Sorry for offtopic but maybe its good time to improve. Can you add
some localisation mechanism for docs. I want to translate it to Russian.
пт, 21 дек. 2018 г., 19:30 Alon Zakai alonza...@gmail.com:
>
>
> On Fri, Dec 21, 2018 at 2:35 AM Beuc wrote:
>
>> The GitHub->GitLab migration
+1
чт, 29 нояб. 2018 г. в 08:40, 越闽魏 :
>
>
> Interesting!
>
> Could you share your script that split up the large js file? @Arnab
>
> Thank you!
>
> On Tuesday, December 29, 2015 at 4:48:48 AM UTC+8, Arnab Choudhury wrote:
>>
>> FYI - In order to work around the problem, I went ahead and wrote a
Amazing job! Congrats!
пт, 4 янв. 2019 г., 1:40 'Thomas Lively' via emscripten-discuss
emscripten-discuss@googlegroups.com:
> This is super cool! 10/10 will play more when I can. It's really great
> seeing the WebAssembly work paying off with such a low barrier to start
> playing such a polished
ngine, source_map, extra_info, just_split, just_concat))
File "/home/caiiiycuk/emscripten/sdk/emscripten/tools/js_optimizer.py",
line 470, in run_on_js
filenames = pool.map(run_on_chunk, commands, chunksize=1)
TypeError: map() got an unexpected keyword argument 'chunksize'
ninja: build
Hi! What is '--mvp-features'? After updating to latest version
(1.38.23/incoming, and
binaryen/master:5b5789495a97602869f18d552b2a9e1814edefae) I can't
compile anything:
main.cpp:
int main() {
return 0;
}
--
~/emscripten/sdk/emscripten/emcc -Oz main.cpp -o main.html
Unknown option
Hi guys. I'm really curious about why emscripten is written both on
python and javascript. Tests are written in python but main sources in
javascript (not only libraries, but also some optimizers). Why you do
in that way? This choice is due to some environment limitions or just
because it's easier
Agree, I never debug emscripten output. Source map never works for big
code bases. So all bug fixing done by playing with C++ native build,
in hope that it will be ported 1:1, actually it's almost true.
пт, 5 апр. 2019 г. в 03:28, Brian Gavin :
>
> My experience with source maps is similar to
eve, that this is our problem:
>> https://bugs.webkit.org/show_bug.cgi?id=181723
>>
>> пн, 11 мар. 2019 г. в 22:31, Александр Гурьянов :
>> >
>> > Nothing special. Just switch to -Oz flag, and drop some 3rd party
>> > libraries. Tools like bloaty is very use
I think to add ios into blacklist for wasm, because this bug is
>>>> completely random. Test from Brion
>>>> (https://github.com/brion/min-wasm-fail/blob/master/min-wasm-fail.js)
>>>> always succeed.
>>>> вт, 31 июл. 2018 г. в 3:57, Alon Z
s, how did you find a way to reduce your wasm size? We're bringing
> games to the web, and any way we can optimize the file sizes would make all
> the difference.
>
> On Mon, Mar 11, 2019 at 12:56 AM Александр Гурьянов
> wrote:
>>
>> No, I found way to reduce wasm size,
I believe, that this is our problem:
https://bugs.webkit.org/show_bug.cgi?id=181723
пн, 11 мар. 2019 г. в 22:31, Александр Гурьянов :
>
> Nothing special. Just switch to -Oz flag, and drop some 3rd party
> libraries. Tools like bloaty is very usefull. In my case I need too
> decre
m-fail.js)
>> always succeed.
>> вт, 31 июл. 2018 г. в 3:57, Alon Zakai :
>> >
>> > How big is TOTAL_MEMORY in this project? It could be either the binary
>> size itself, or the memory size, that it doesn't have enough memory for.
>> >
>>
Hi. I think that best way to do it is to use wasm-opt from binaryen
(thanks to Alon, hi told me how to do this):
bin/wasm-opt --func-metrics test/hello_world.wast
==
global
[funcs]: 1
func: add
[binary-bytes] : 7
[total]: 3
binary : 1
local.get : 2
export: add
HI guys! I trying to implemnent wasm2wasm pass that remove specific
function from wasm. Inspired by ExtractFunction pass I wrote this
code:
for (auto& func : module->functions) {
if (names.find(func->name) != names.end()) {
// wipe out all the
Maybe another option is to use optimized build with -Oz (to minimize
startup time).
пт, 1 февр. 2019 г. в 06:51, Floh :
>
> As long as you're on fairly recent Android/iOS versions where the browsers
> have WASM support, it should "just work".
>
> If WASM is not an option you'd need to fallback
subset of the wasm text
> format, the s-expression part).
>
> Overall, binaryen focuses on the wasm binary format and optimizing and
> analyzing it, while wabt focuses on reading and writing of the full wasm
> standard, both binary and text.
>
> On Mon, Feb 4, 2019 at 9:31 AM
Hi guys! I want to write some tool to analyze wasm files. My first
target is to create tool like bloaty to analyze size of contribution
of each function in resulting size. But bloatly is universal tool, I
want to concentrate on wasm format and maybe create something like
"code explorer" tool (like
the bigger missing pieces here. This
>> will hopefully improve soon in toolchains and browsers, there is ongoing
>> work on it. Meanwhile debugging a parallel native build is often best, as
>> mentioned.
>>
>> On Thu, Apr 4, 2019 at 8:56 PM Александр Гурьянов wrote:
>&g
Hi guys. I am not familar with MacOs, so maybe my question is really
simple. I have a lot of projects based on OpenGL variations (GL1, ES,
ES2). Emscripten have brilliant support for it. My main OS is linux
that also provide fantastic capabilities to building OpenGL projects.
Because of that
Thanks for feedback. Now I am saving and restoring all wasm mutable
globals. But seems it does not help, I still have unstable behaviour.
I forgot to say that I use emterpeter too, can it produce additional
problem for save/restore wasm state?
вс, 12 мая 2019 г. в 22:31, Liam Wilson :
>
> A
ug
> somehow. If you build with -s VERBOSE=1 it should print the decisions it
> makes when adding symbols, and then it should say which ends up requiring
> PATH_FS. Then hopefully we can figure it out from there. (Or, if you can
> provide a testcase I can look at that directly.)
>
&g
Hi is it possible to make snapshot of runtime, serelize it to file and
then restore runtime on other client. I understand that snapshotting
have a lot of pitfalls. But for now I want to implement simpliest
case. My target is wasm, main loop is executed by requestAnimation
frame.
So stack always
Hi guys! Just updated to latest incoming
(39158e5e344460d5c0a5da8913f87e4cb0db6e4a) and now have this error
that does not exists on previous version:
error: undefined symbol: PATH_FS
warning: To disable errors for undefined symbols use `-s
ERROR_ON_UNDEFINED_SYMBOLS=0`
Have no idea what is
Hmm, will try to bisect, my code definitely uses filesystem.
ср, 8 мая 2019 г. в 21:19, Floh :
>
> I just checked with my code, and everything still works, *but* I'm using
> NO_FILESYSTEM everywhere (there was a similar problem a few weeks ago too
> though:
Hi. You can replace it with -s WASM=0. That means to produce asm.js as
output, if you set -s WASM=1 then output will be wasm.
пт, 5 июл. 2019 г. в 08:57, juef :
>
> Greetings,
>
> I am trying to compile an old project from GitHub which uses emscripten, and
> the makefile calls emcc with a -s
ring an async event though.
>> >
>> > To debug this kind of thing, the DETERMINISTIC option may be useful. It
>> > makes all timing and random numbers deterministic, which for many programs
>> > (without user input, and single-threaded) means it is determini
Hi. I am trying to set TOTAL_MEMORY to 32mb, but compiler told me:
shared:ERROR: Memory is not large enough for static data (30099056)
plus the stack (5242880), please increase TOTAL_MEMORY (33554432) to
at least 35342960
Is there any way to find what and where this data located?
Thanks
--
HI. I used wasm-opt manually to optimize wasm binary. On last step I
used Oz to finalize optimizations. My problem is that after Oz
function indexes is changed and I can't get symbols map to restore
function names. How I can map new function indexes to old one (from
symbols file).
In other words:
wdosbox.wasm.js - is wasm binary
wdosbox.js.symbols - symbols
вт, 23 июл. 2019 г. в 12:03, Александр Гурьянов :
>
> I am sorry guys, my estimation of sizes/preformance was wrong.
>
> Ubuntu 19.04 / Chrome 72.0.3626.121
>
> O3:
> DEFAULT 2.0Mb 60 FPS /
ere
пн, 22 июл. 2019 г. в 18:15, Александр Гурьянов :
>
> Even digger works very poor here http://js-dos.com/ (in chrome). I
> just changed emscripten_sleep_with_yield with emscripten_sleep and
> that's all. Don't know why Chrome goes crazy.
>
> пн, 22 июл. 2019 г. в 18:0
Hi. Trying to compile dosbox with new latest-upstream. It's compiles
without error, but when I run it, I have this error
missing function: emscripten_sleep_with_yield
I just replaced all emterpreter related flags with
-s ASYNCIFY
--
You received this message because you are subscribed to the
июл. 2019 г. в 14:47, Александр Гурьянов :
>
> Hi. Trying to compile dosbox with new latest-upstream. It's compiles
> without error, but when I run it, I have this error
>
> missing function: emscripten_sleep_with_yield
>
> I just replaced all emterpreter related flags with
&g
h_yield). Module.setAsync is an internal API of old Asyncify - if you
> have custom JS code using it, you need to update it to the new Asyncify's API
> (see notes on handleSleep in the blogpost:
> https://kripken.github.io/blog/wasm/2019/07/16/asyncify.html ).
>
> On Thu, Jul 18, 20
Perhaps the main interpreter
> loop? It's hard to see more optimization opportunities there. I think that
> means we need a whitelist like in emterpreter. Btw, did you use the whitelist
> or blacklist there - which would be better?
>
> - Alon
>
>
> On Mon, Jul 22, 2019 at
1 - 100 of 182 matches
Mail list logo