asjsc does not generate an index.html file, so you must be looking at the output of mxmlc.
Today, asjsc outputs a JS file that lists the dependencies, as I described. They're the exact same addDependency() calls that you'd see in mxmlc's index.html. - Josh On Tue, Jan 5, 2016 at 1:06 PM, OmPrakash Muppirala <bigosma...@gmail.com> wrote: > On Jan 5, 2016 9:36 AM, "Josh Tynjala" <joshtynj...@gmail.com> wrote: > > > > asjsc still generates all of the addDependency() calls required for a > debug > > build. No hand crafting needed. It generates them in a JS file named > > ProjectName-dependencies.js. Not coincidentally, that is the only JS file > > that you need to reference in a script tag in your HTML when running the > > debug build. > > I suppose it does not work that way today? I see all individual calls > loading each dependency in index.html. > > > > > - Josh > > On Jan 5, 2016 1:15 AM, "OmPrakash Muppirala" <bigosma...@gmail.com> > wrote: > > > > > On Mon, Jan 4, 2016 at 8:49 PM, Alex Harui <aha...@adobe.com> wrote: > > > > > > > Well, the compiler could be upgraded to process a template like Flash > > > > Builder currently does. I'm curious to know how many folks use Flash > > > > Builder and/or Ant tasks to process the html templates for SWFs vs > > > > plugging in some custom thing in their workflow. > > > > > > > > > > I remember in Flex 2 days where it was very cool to quickly write some > code > > > and hit the run button and see the browser pop-up with the app. That > made > > > for a very good first impression. If we don't have a quick way to > visually > > > see the JS code that got generated, there is not much of a first > > > impression, IMHO. > > > > > > Also, it is not very obvious that we need to create a html file with a > new > > > MainClass.start() on body load. And it seems like we are > > > using goog.addDependency calls to load the required Javascript files. > Do > > > we really expect the users to handcraft this everytime? That could be > tons > > > of JS files to be added by hand. Kind of defeats the purpose of having > a > > > transpiler. > > > > > > That said, it will become annoying very quickly when one realizes that > the > > > index.html cannot be changed. > > > > > > I think having a very simple html file as a default template is a happy > > > medium. It works for the instant gratification that new users would > seek > > > and more advanced users can dig in a bit deeper and swap out the > default > > > with a custom template. > > > > > > Thanks, > > > Om > > > > > > > > > > > > > > But IMO, the main reason to have an option is so folks can save a > step in > > > > getting the SDK and trying it out. > > > > > > > > -Alex > > > > > > > > On 1/4/16, 7:56 PM, "Josh Tynjala" <joshtynj...@gmail.com> wrote: > > > > > > > > >I should add that I'm not opposed to adding some kind of optional > flag > > > to > > > > >asjsc that tells it to generate an HTML file similar to how mxmlc > does > > > it. > > > > >That HTML file just doesn't seem especially useful to me, as I > consider > > > > >what it would be like to use asjsc in a real-world project. So I'm > > > trying > > > > >to get a better understanding of your perspective. > > > > > > > > > >- Josh > > > > > > > > > >On Mon, Jan 4, 2016 at 7:49 PM, Josh Tynjala <joshtynj...@gmail.com > > > > > > >wrote: > > > > > > > > > >> Is it actually necessary for the compiler to create some kind of > > > > >> boilerplate HTML for you? It may be a little useful for quick > demos, > > > > >>I'll > > > > >> concede, but many real world projects will need highly customized > HTML > > > > >> files. Many need things like analytics, CSS, and other static HTML > > > > >>content > > > > >> that isn't purely generated by JavaScript (for SEO and things). > > > > >> > > > > >> In fact, the compiler isn't really set up for customizing the HTML > > > that > > > > >>it > > > > >> currently generates with mxmlc. You can see it is mostly > hard-coded in > > > > >> JSGoogPublisher.java. It's actually very simple markup. Probably > too > > > > >>simple > > > > >> to use in production for most people, especially if they want to > use > > > > >>asjsc > > > > >> and integrate it into the rest of their web development workflow. > > > > >> > > > > >> - Josh > > > > >> > > > > >> On Mon, Jan 4, 2016 at 5:14 PM, OmPrakash Muppirala > > > > >><bigosma...@gmail.com> > > > > >> wrote: > > > > >> > > > > >>> On Mon, Jan 4, 2016 at 4:47 PM, Alex Harui <aha...@adobe.com> > wrote: > > > > >>> > > > > >>> > If you diff asjsc vs mxmlc you'll see the difference. > > > > >>> > > > > > >>> > > > > >>> This is the difference I see: > > > > >>> > > > > >>> asjsc: -js-output-type=jsc > > > > >>> -external-library-path="$SCRIPT_HOME/../libs/js.swc" > > > > >>> mxmlc: -js-output-type=FLEXJS > > > > >>> -sdk-js-lib="$FLEX_HOME/frameworks/js/FlexJS/src" > > > > >>> > > > > >>> > > > > >>> So, -js-output-type=FLEXJS instead of jsc should do the trick of > > > > >>>creating > > > > >>> the index.html file? > > > > >>> > > > > >>> > > > > >>> > IMO, I wouldn't call a new script mxmlcnpm because others may > want > > > an > > > > >>> auto > > > > >>> > generated hmtl as well. Give it a more generic name. > > > > >>> > > > > > >>> > > > > >>> Here are the current use cases: > > > > >>> > > > > >>> 1. Convert AS3 (targeting HTML DOM) to JS -> use asjsc > > > > >>> 2. Convert AS3 + MXML (targeting FlexJS) to JS + HTML > use > mxmlc > > > > >>> > > > > >>> The use case we need to add is > > > > >>> Convert AS3 (targeting HTML DOM) to JS + HTML > > > > >>> > > > > >>> Something like asjshtmlc? In that case, shouldn't mxmlc be > renamed > > > to > > > > >>> mxmlcjshtmlc as well, for the sake of consistency? > > > > >>> > > > > >>> Or am I overthinking this? What would you suggest? > > > > >>> > > > > >>> Thanks, > > > > >>> Om > > > > >>> > > > > >>> > > > > >>> > > > > > >>> > -Alex > > > > >>> > > > > > >>> > On 1/4/16, 4:28 PM, "omup...@gmail.com on behalf of OmPrakash > > > > >>> Muppirala" > > > > >>> > <omup...@gmail.com on behalf of bigosma...@gmail.com> wrote: > > > > >>> > > > > > >>> > >I think I get it. > > > > >>> > > > > > > >>> > >I thought that the source code for js.swc was in > > > > >>> > >$FLEX_HOME/frameworks/js/FlexJS/src. > > > > >>> > >I guess that is not true? > > > > >>> > > > > > > >>> > >The original problem was that asjsc does not create the > index.html > > > > >>> file. > > > > >>> > >I > > > > >>> > >was asked to use mxmlc for that. (Refer to the npm install > flexjs > > > > >>> thread) > > > > >>> > > > > > > >>> > >When I used the script in {installed_flexjs}/js/bin/mxmlc, it > blew > > > > >>>up > > > > >>> > >because it could not find the definitions for HTMLElement, > > > > >>>SVGElement > > > > >>> etc. > > > > >>> > >because they are in js.swc. I don't think it blew up because > of > > > the > > > > >>> > >missing /frameworks/js/FlexJS/src folder. Adding the external > > > > >>>library > > > > >>> > >path > > > > >>> > >to js.swc fixed this issue. > > > > >>> > > > > > > >>> > >The way I did this was to create a new mxmlcnpm script and add > > > this > > > > >>> js.swc > > > > >>> > >library path in that. Is that okay? > > > > >>> > > > > > > >>> > >I guess another question is: what would be the best way to add > > > > >>>ability > > > > >>> to > > > > >>> > >create index.html capability to asjsc? > > > > >>> > > > > > > >>> > >Thanks, > > > > >>> > >Om > > > > >>> > > > > > > >>> > > > > > > >>> > >On Mon, Jan 4, 2016 at 4:16 PM, Alex Harui <aha...@adobe.com> > > > > wrote: > > > > >>> > > > > > > >>> > >> > > > > >>> > >> > > > > >>> > >> On 1/4/16, 4:09 PM, "omup...@gmail.com on behalf of > OmPrakash > > > > >>> > Muppirala" > > > > >>> > >> <omup...@gmail.com on behalf of bigosma...@gmail.com> > wrote: > > > > >>> > >> > > > > >>> > >> >In the flexjs/js/bin/mxmlc script, I see that we are > > > referencing > > > > >>> the ' > > > > >>> > >> >*/frameworks/js/FlexJS/src*' folder. > > > > >>> > >> > > > > >>> > >> This folder is intended as the place for folks to put > > > > >>>monkey-patched > > > > >>> JS > > > > >>> > >> files so they can override the JS in the SWCs if they need > to > > > > >>> > >>workaround a > > > > >>> > >> bug. > > > > >>> > >> > > > > >>> > >> What code blew up? Maybe we should create an empty folder > there > > > > >>>or > > > > >>> make > > > > >>> > >> the compiler tolerant of it not being there. > > > > >>> > >> > > > > >>> > >> Trying to use js.swc with MXMLC is not currently the common > > > > >>> > >>configuration > > > > >>> > >> for FlexJS. Most folks who are using MXML and AS to build a > > > > >>>FlexJS > > > > >>> app > > > > >>> > >> shouldn't need to write directly the the JS API especially > if > > > they > > > > >>> want > > > > >>> > >>to > > > > >>> > >> use a SWF version for testing and/or deployment. > > > > >>> > >> > > > > >>> > >> If you want to build out a different script for folks to use > to > > > > >>>build > > > > >>> > >> native apps, feel free to do that. > > > > >>> > >> > > > > >>> > >> -Aleex > > > > >>> > >> > > > > >>> > >> > > > > >>> > > > > > >>> > > > > > >>> > > > > >> > > > > >> > > > > > > > > > > > >