> Also, this bug report may be somehow remotely related: > https://github.com/kripken/emscripten/issues/5361 I don't think so, because in that bug undefined symbol is actually used in runtime. But in my case undefined symbols never used in runtime. But I think you right compiler does not understand that symbol is not used and erro ont it.
> I don't think this would work natively either, but linker would error out at > link time on a missing symbol You right again, compiler show this error, but as I said before I use flag "-Wl,--unresolved-symbols=ignore-all" to avoid this error. > You could try adding a single .cpp file to the build with that symbol defined My project is big game, that have two parts: game and editor. Editor is used by game devs, to make levels etc. Editor can be activated by special in-game flag. But for gaming editor is not needed, and it is hard task to compile it in web. So best solution is to drop editor's *.cpp and compile with "-Wl,--unresolved-symbols=ignore-all". In that case game works perfect, and it's much easier to compile. Btw, I add implementation for s_editor, and now I have same error but on next undefined symbol: LinkError: import object field '__ZTI11EditorScene' is not a Number Of course I can add mock implementation for all Editor's symbols, but may be there is a better solution? 2017-08-12 21:34 GMT+07:00 Jukka Jylänki <juj...@gmail.com>: > Also, this bug report may be somehow remotely related: > https://github.com/kripken/emscripten/issues/5361 > > > lauantai 12. elokuuta 2017 Jukka Jylänki <juj...@gmail.com> kirjoitti: >> >> I would expect unused imports to be optimized out. Perhaps this is a case >> where the compiler is not able to eliminate some function that is >> referencing the given object. I don't think this would work natively either, >> but linker would error out at link time on a missing symbol (similar to what >> you will get if linking with -s ERROR_ON_UNDEFINED_SYMBOLS=1 in Emscripten) >> >> 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. elokuuta 2017 Александр Гурьянов <caiiiy...@gmail.com> >> kirjoitti: >>> >>> Chrome error: >>> failed to asynchronously prepare wasm: LinkError: Import #3 >>> module="env" function="__ZN6Editor10s_instanceE" error: global import >>> must be a number >>> testing.html:160 LinkError: Import #3 module="env" >>> function="__ZN6Editor10s_instanceE" error: global import must be a >>> number >>> at http://localhost/bin.js?timestamp=1502360564253:5659:23 >>> at <anonymous> >>> testing.html:161 LinkError: Import #3 module="env" >>> function="__ZN6Editor10s_instanceE" error: global import must be a >>> number >>> at http://localhost/bin.js?timestamp=1502360564253:5659:23 >>> at <anonymous> >>> testing.html:1 Uncaught (in promise) abort({}) at Error >>> at jsStackTrace >>> (http://localhost/bin.js?timestamp=1502360564253:5206:12) >>> at stackTrace >>> (http://localhost/bin.js?timestamp=1502360564253:5220:11) >>> at abort (http://localhost/bin.js?timestamp=1502360564253:15934:43) >>> at http://localhost/bin.js?timestamp=1502360564253:5664:4 >>> at <anonymous> >>> 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) 1.37.18 (commit >>> > cb2260892ddc2c73aa40dbe0821f5a8ce3f79975) >>> > >>> > bin.js: >>> > function __ZN6Editor10s_instanceE() { >>> > Module["printErr"]("missing function: _ZN6Editor10s_instanceE"); >>> > abort(-1); >>> > } >>> > // ... >>> > getBinaryPromise().then((function(binary) { >>> > return WebAssembly.instantiate(binary, info); // <-- HERE >>> > })).then((function(output) { >>> > >>> > I don't know how to look inside wasm, it's looks like binary file: >>> > STACKTOP env STACK_MAX env __ZN6Editor10s_instanceE env >>> > >>> > c++: >>> > class Editor { >>> > private: >>> > static Editor* s_instance; >>> > public: >>> > static Editor* getInstance() { return s_instance; } >>> > >>> > } >>> > >>> > >>> > Yes it's a singleton pointer, but this class is not implemented, it >>> > have many undefined references, including to s_instance: >>> > 51: undefined reference to `Editor::s_instance' >>> > >>> > So I think problem in that place, but I don't want to implement it, it >>> > never calls. >>> > >>> > >>> > 2017-08-10 17:12 GMT+07:00 Jukka Jylänki <juj...@gmail.com>: >>> >> Is that on latest Emscripten 1.37.18? That looks like a new error that >>> >> is not known. What does the generated code around that location look >>> >> like? What should __ZN6Editor10s_instanceE be, is 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 run it due this error: >>> >>> failed to asynchronously prepare wasm: LinkError: import object field >>> >>> '__ZN6Editor10s_instanceE' is not a Number testing.html:161:40 >>> >>> LinkError: import object field '__ZN6Editor10s_instanceE' is not a >>> >>> Number testing.html:160:37 >>> >>> LinkError: import object field '__ZN6Editor10s_instanceE' is not a >>> >>> Number testing.html:161:40 >>> >>> >>> >>> I have some unresovled symbols in project, but this symbols never >>> >>> calls from code. I also build native version 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. >>> >>>> >>> >>>> 2017-08-09 19:04 GMT+07:00 Jukka Jylänki <juj...@gmail.com>: >>> >>>>> This looks like a performance bug in the browser. It may be >>> >>>>> possible >>> >>>>> to reduce the compilation time, but without a test case it's not >>> >>>>> possible to know what might be wrong. Is the project public >>> >>>>> somewhere? >>> >>>>> >>> >>>>> One known source of slow compilation times is if one has very large >>> >>>>> switch-case statements with tens of thousands of cases. >>> >>>>> >>> >>>>> Though FF 54 already supports WebAssembly. Do you have a >>> >>>>> WebAssembly >>> >>>>> build of the same, does that compile quickly on Firefox? >>> >>>>> >>> >>>>> 2017-08-09 14:52 GMT+03:00 caiiiycuk <caiiiy...@gmail.com>: >>> >>>>>> Hi. I use Firefox 54.0 to test my game that I compile from C++ to >>> >>>>>> js. Js >>> >>>>>> size is 6.3Mb, when I open game in FF, then it compiles asm.js for >>> >>>>>> very long >>> >>>>>> time: >>> >>>>>> Successfully compiled asm.js code (total compilation time 55703ms; >>> >>>>>> stored in >>> >>>>>> cache) >>> >>>>>> >>> >>>>>> As you see I wait one minute before game starts, all other >>> >>>>>> browsers start >>> >>>>>> game immediately. It's possible to reduce this compilation time? >>> >>>>>> >>> >>>>>> I have very powerfull CPU - i7, os is Ubuntu x64 >>> >>>>>> >>> >>>>>> -- >>> >>>>>> 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 emscripten-discuss+unsubscr...@googlegroups.com. >>> >>>>>> 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 emscripten-discuss+unsubscr...@googlegroups.com. >>> >>>>> 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 emscripten-discuss+unsubscr...@googlegroups.com. >>> >>> 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 emscripten-discuss+unsubscr...@googlegroups.com. >>> >> 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 emscripten-discuss+unsubscr...@googlegroups.com. >>> 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 emscripten-discuss+unsubscr...@googlegroups.com. > 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 emscripten-discuss+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
