Meanwhile I found another testcase that needs EXPORT_ALL, so I made
MAIN_MODULE automatically enable that.

There isn't a current plan to implement dlopen for fastcomp. I would like
to think some more and try to find a way to do it  that is optimized, but
that isn't a short-term thing.

- Alon



On Thu, Jun 5, 2014 at 4:50 AM, Marc Agbanch <[email protected]> wrote:

>
> My bad, I did a mistake when I wrote the commands in the text file, on my
> environment, EMCC_FORCE_STDLIBS was correctly written and defined. The -s
> EXPORT_ALL=1 works well, thank you for your help.
> Do you already know when you plan to add the dynamic linking compatibiltiy
> with fastcomp? I like the idea to build  a main part and add some plugins
> later, even if it isn't as fast as a one file js code.
>
> -Marc
>
> Le mardi 3 juin 2014 02:26:42 UTC+2, Alon Zakai a écrit :
>>
>> Ok, I ran this locally. There are two problems: First, you had
>> EMCC_FORCE_LIBS instead of EMCC_FORCE_STDLIBS in the environment options.
>> Second, due to how asm.js works, the symbols also need to be exported, so
>> this requires -s EXPORT_ALL=1 when building main. I honestly don't remember
>> why that is needed, though - might be fixable, but this code is mostly
>> deprecated as mentioned on the wiki.
>>
>> - Alon
>>
>>
>>
>> On Fri, May 30, 2014 at 7:45 AM, Marc Agbanch <[email protected]>
>> wrote:
>>
>>> I tried to add the LINKABLE option to the build, it didn't removed the
>>> reference error.
>>>
>>> Le mardi 27 mai 2014 20:09:47 UTC+2, Alon Zakai a écrit :
>>>>
>>>> Try to also build with    -s LINKABLE=1     , which will ensure that
>>>> LLVM dead code elimination does not remove the code.
>>>>
>>>> - Alon
>>>>
>>>>
>>>>
>>>> On Thu, May 22, 2014 at 8:48 AM, Marc Agbanch <[email protected]>
>>>> wrote:
>>>>
>>>>> I created the environment variable  EMCC_FORCE_STDLIBS and set It to
>>>>> 1. The compilation time emit
>>>>> including libcxx
>>>>> including libcextra
>>>>> including libcxxabi
>>>>> including gl
>>>>> to the compillations steps, but it didn't change the runtime output.
>>>>> libc was already included. I join the code which produced the error with
>>>>> that message, I think it is reproducible.
>>>>>
>>>>>
>>>>> Le mercredi 21 mai 2014 23:47:28 UTC+2, Alon Zakai a écrit :
>>>>>>
>>>>>> You may need to force-include a system library if the library uses it
>>>>>> but not the main file (see Linking wiki page). However, I am a little
>>>>>> surprised that would be the case for the new() operator, unless the main
>>>>>> file is not using libc (which includes malloc and new).
>>>>>>
>>>>>> - Alon
>>>>>>
>>>>>>
>>>>>>
>>>>>> On Wed, May 21, 2014 at 8:45 AM, Marc Agbanch <[email protected]>
>>>>>> wrote:
>>>>>>
>>>>>>> Thanks for your answer.
>>>>>>>
>>>>>>> Following yours advices, I build with -s ASSERTIONS=1. I also used
>>>>>>> -g4 for the source map. The  return error is "Error in loading
>>>>>>> dynamic library: ReferenceError: __ZdlPv is not defined". Thanks to
>>>>>>> the source map, I know that function is associated with the c++ operator
>>>>>>> delete in my side module. The difference between my simple library which
>>>>>>> worked perfectly and that one is the first one use only c code, and the
>>>>>>> second implement a use c++ classes. It seems to me that dlopen() have 
>>>>>>> some
>>>>>>> issues with c++ code.
>>>>>>> I'll try to join the necessary about my code tomorrow.
>>>>>>>
>>>>>>> Le vendredi 16 mai 2014 02:07:09 UTC+2, Alon Zakai a écrit :
>>>>>>>>
>>>>>>>> If you build with   -s ASSERTIONS=1   (on by default in -O0), you
>>>>>>>> should see a message saying why it failed to evaluate that file.
>>>>>>>>
>>>>>>>> - Alon
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> On Tue, May 13, 2014 at 6:24 AM, Marc Agbanch <[email protected]>
>>>>>>>> wrote:
>>>>>>>>
>>>>>>>>>
>>>>>>>>> Hi,
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> I succeeded in porting  a C ++ project of mine to html with
>>>>>>>>> embedded JS thanks to Emscripten.
>>>>>>>>>
>>>>>>>>>  Now, I'm currently trying  to use the dlopen support. I made a
>>>>>>>>> test with a simple library  and it ran perfectly.  I made another 
>>>>>>>>> test with
>>>>>>>>> a bigger  code and then, during the execution, the dlerror() folowing 
>>>>>>>>> my
>>>>>>>>> dlopen()  returned the folowing message:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> "Could not evaluate dynamic lib: /lib_note.js"
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> I checked the compiling options (DLOPEN_SUPPORT, MAIN_MODULE,
>>>>>>>>> SIDE_MODULE, EXPORTED_FUNCTIONS, --preload-file) ,my code, the wiki, 
>>>>>>>>> but I
>>>>>>>>> really don't have a clue about what could cause that error. Any idea ?
>>>>>>>>>
>>>>>>>>> --
>>>>>>>>> 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/d/optout.
>>>>>>>>>
>>>>>>>>
>>>>>>>>  --
>>>>>>> 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/d/optout.
>>>>>>>
>>>>>>
>>>>>>  --
>>>>> 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/d/optout.
>>>>>
>>>>
>>>>  --
>>> 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/d/optout.
>>>
>>
>>  --
> 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/d/optout.
>

-- 
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/d/optout.

Reply via email to