On Monday, July 6, 2015 at 2:42:04 PM UTC+2, Stephen Haberman wrote: > > Sounds like I'm late to party; I missed the 2015 meetup videos, does > someone have a link? (That will teach me to not login to G+ very > often...well played, Google...) >
https://www.youtube.com/playlist?list=PL1yReUCGwGvrqscLu1EAyYRPrr0ceEHLE (link to the slides is in the playlist description) > Not that my opinion matters very much (vs. the GWT team who's doing all > the actual work), but I'll +100 any plans for large/breaking > changes/rewrite for GWT 3.0/j2cl. > > Everyone that is concerned about backwards compatibility, you can stay on > GWT 2.x. What we need is foundations for a GWT that will be awesome 5 years > from now. For the next generation of applications. That's who we need to be > worried about. > > Yes, that means pain/being left out for the current generation of > applications, but that's how technology works. GWT 2.x will not suddenly be > taken off Maven central. Plenty of enterprise applications rely on > older/mature technology, e.g. jars/projects that aren't pushing out new > releases every 2 months with amazing new features. That's fine. If it's too > expensive to switch, then don't. > …with the caveat that “The web is not a platform. It's a continuum.” <https://adactio.com/journal/6692>, which in terms of browser bugs and browser sniffing means that libraries/frameworks have to be maintained if you expect them to continue to work with newer/future browsers and/or browser versions (remember the mess with every new IE version, requiring changes in GWT? has anyone tried GWT in MS Edge?) Hopefully this is going to stabilize with evergreen browsers, but there's still a limit in time before GWT 2.x "explodes" and everyone has no choice but migrating to something else (a newer version of GWT, or something else entirely). So saying "GWT 2.x is here to stay" really would depend on someone maintaining it: Sounds like a great opportunity for other companies to step-in and provide > enterprise support for 2.x as well. Google has never been interested in > that game anyway. > > Someone mentioned early access to j2cl for framework authors to start > porting; I'd throw my hat into that ring. It would be interesting to see > what Tessell looks like with it. > As I said elsewhere, you don't have to wait for an early access to j2cl to start porting. Proof (admittedly a simple/limited one): https://github.com/tbroyer/gwt-places > Tessell very heavily uses UiBinder to drive its MVP codegen, so I'll have > to either port Tessell to the next-generation of templates, or (more > likely) port UiBinder to j2cl (using APT/something; maybe not 100% > backwards compatible, but at least something that is not "throw away all > your templates and start from scratch"). > Re. UiBinder, maybe it could have a first generator pass creating Java source for the *.ui.xml, that could later be reused by an annotation processor working from the "owner Java class": what's specific to the *.ui.xml is generated by a first pass (before JavaC) –just like you do with Protocol Buffer and the like–, and the annotation processor (during JavaC) does only what's specific to the owner class (@UiField, @UiHandler, @UiFactory), calling into the previously-generated class for the *.ui.xml-specific things (that class could possibly be subclassed to override methods when the owner class has @UiField(provided=true) or @UiFactory vs. a simple @UiField, for instance) That unfortunately would call for a complete rewrite, and UiBinder is quite complex… But if the alternative is to write something else from scratch, then reusing the *.ui.xml and @UiField "contract" probably wouldn't be a bad starting point (and ease porting existsing code). But I think we should just stop speculating, wait a bit for Google to continue their experiment with j2cl and have a clearer view of how they want to migrate their existing apps, and then see which features Google would like to port themselves (because they need them anyway) and which ones they'd "abandon", and whether they'll be replaced or not. -- You received this message because you are subscribed to the Google Groups "GWT Contributors" group. To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit-contributors+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/google-web-toolkit-contributors/50244c3d-81b9-4a57-8fdb-2fc92ce2534c%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
