FWIW, we're not currently planning to remove the current behavior of modularization that behaves like a factory.
I do agree that in most cases using a single instance and then having the lifecycle management within the module is probably more efficient (avoids duplicate wasm instances). On Fri, Dec 13, 2024 at 11:09 AM Brooke Vibber <br...@pobox.com> wrote: > Cool! Sounds like I will have to make changes to support a lifecycle > around a single instance for my codec instantiation case but arguably > that's superior in general. :) > > I'll prototype this work soon. > > -- b > > On Fri, Dec 13, 2024, 9:56 AM 'Brendan Dahl' via emscripten-discuss < > emscripten-discuss@googlegroups.com> wrote: > >> Update on modularization changes: >> >> A new experimental setting "-sMODULARIZE=instance" has landed >> <https://github.com/emscripten-core/emscripten/pull/22867> in version >> 3.1.73. The new output creates a single instance of the Wasm module and >> also uses ES module exports for the corresponding exported Wasm and runtime >> JS functions. The module also exports a default function for initializing >> the module that must be called before any of the named exports can be >> called. This style of module is hopefully more similar to what ES module >> users expect and should be easier for bundlers to reason about. >> >> An example of using the module: >> >> import init, { foo, bar } from "./my_module.mjs" >> await init(optionalArguments); >> foo(); >> bar(); >> >> Feedback on the new setting is welcome here or please file a github issue. >> >> On Monday, April 1, 2024 at 4:08:55 PM UTC-7 Brooke Vibber wrote: >> >>> I don't need internals in global scope, but I _do_ need multiple >>> instantiation if possible for the demuxers and codecs in ogv.js. >>> >>> If that's dropped, I'll have to rewrite some C and JavaScript code for >>> my internal APIs to support multiple active data streams in a long-lived >>> instance with explicit data structure lifetime management, instead of each >>> instance being single-use and garbage collected independently. It wouldn't >>> a hard blocker, but it would require some refactoring if there's only one >>> mode and it's single-instance. :) >>> >>> -- b >>> >>> >>> On Mon, Apr 1, 2024 at 12:52 PM 'Sam Clegg' via emscripten-discuss < >>> emscripte...@googlegroups.com> wrote: >>> >>>> A few of us on the core emscripten team have been tossing around ideas >>>> about how to improve/modernize the modularization of emscripten-generated >>>> code. >>>> >>>> I've created a google doc >>>> <https://docs.google.com/document/d/1ccH3NsRQgEc1HXWI_IFAnMbTM0Vn1lP_EhJmdLRzYvs> >>>> with >>>> comments open to anyone. Please feel free to comment on this document and >>>> raise any concerns you have either there or here on this thread. >>>> >>>> The first step along the way is to improve encapsulation by bringing >>>> back the old `-sMODULARIZE_INSTACE` settings (perhaps with a better name). >>>> All this setting really does is wrap the emscripten-generated code in a >>>> private function context. I'm hoping we can do this without affecting >>>> code size in any significant way. >>>> >>>> Ideally we would make this new mode the default, or even the only >>>> mode. However there will likely be some users who are currently depending >>>> on the emscripten internally being available in the global scope. So at >>>> least initially this will be optional, and perhaps enabled in `-sSTRICT` >>>> mode. >>>> >>>> It would be very useful to know how many folks currently depend on the >>>> internals of the generated code being available in the global scope? Do >>>> you use emscripten-generated functions or globals without accessing them >>>> via the `Module` object? (Note that if you use --closure=1 you already >>>> cannot access these internals due to minification, so you won't be >>>> affected by this change). >>>> >>>> cheers, >>>> sam >>>> >>>> -- >>>> 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-disc...@googlegroups.com. >>>> To view this discussion on the web visit >>>> https://groups.google.com/d/msgid/emscripten-discuss/CAL_va2_FWhBnRgKi_33Fw4nkhqFjV_uJKU4gGdSoBZjKf%3DOTwQ%40mail.gmail.com >>>> <https://groups.google.com/d/msgid/emscripten-discuss/CAL_va2_FWhBnRgKi_33Fw4nkhqFjV_uJKU4gGdSoBZjKf%3DOTwQ%40mail.gmail.com?utm_medium=email&utm_source=footer> >>>> . >>>> >>> -- >> 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. >> To view this discussion visit >> https://groups.google.com/d/msgid/emscripten-discuss/dc8c4405-60ae-4c02-9bf4-92c89408f48an%40googlegroups.com >> <https://groups.google.com/d/msgid/emscripten-discuss/dc8c4405-60ae-4c02-9bf4-92c89408f48an%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> > -- > 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/PCRw4ttc6Rc/unsubscribe > . > To unsubscribe from this group and all its topics, send an email to > emscripten-discuss+unsubscr...@googlegroups.com. > To view this discussion visit > https://groups.google.com/d/msgid/emscripten-discuss/CAFnWYTkVYvd1Ht4%3DKG2YVE%2BC4NHwMqSSQKdW%3DSyfZ4oGV3Ug0A%40mail.gmail.com > <https://groups.google.com/d/msgid/emscripten-discuss/CAFnWYTkVYvd1Ht4%3DKG2YVE%2BC4NHwMqSSQKdW%3DSyfZ4oGV3Ug0A%40mail.gmail.com?utm_medium=email&utm_source=footer> > . > -- 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. To view this discussion visit https://groups.google.com/d/msgid/emscripten-discuss/CAONg2ag3eAkNrmWzGHkhA76rva5V%2B%2BBaHYSV3Btqm5HC6VVuDg%40mail.gmail.com.
