Thanks Thomas. Good point on getting started experimenting before the compiler is made available. I've been building a few of libraries using APT for code generation recently, making use of the wonderful google & square libraries like auto and javapoet. I think it'd be pretty straightforward to build an annotation processor for a few existing libraries e.g. i18n. However, I'm really hoping to make use of GWT compiler level solutions to overcome some current APT challenges; in particular, invoking the processor when non-java files are changed such as ui.xml files. Daniel mentioned in his talk that there might be a general solution for this which would be pretty awesome to use.
I was probably too specific in asking for early access to the compiler. Early access to elemental 2.0 and jsinterop 2.0 would be really helpful to try and figure out how to port JSNI/JSO code over to jsinterop. In general my experience with GWT + APT for my own libraries has been really good, so I'm quite positive about the plans for GWT 3.0. Hopefully, with a clear migration story, the community will be able to smooth any rough edges that the core GWT team can't take on. On Sun, Jun 21, 2015 at 11:18 AM Thomas Broyer <[email protected]> wrote: > > > On Sunday, June 21, 2015 at 11:17:50 AM UTC+2, James Horsley wrote: >> >> Also worth noting that, outside of GWT-RPC, I think that many of the >> current major GWT features (widgets, uibinder, etc.) could be ported by the >> community such that only minor code changes are necessary. I would even >> hazard a guess that companies like Vaadin might be interested in supporting >> or at least contributing to a 3.0 based port things like the widget library. >> > > I'd rather bet on Sencha for widgets. > (note: I have absolutely no information about whether this will happen and > who would make it happen) > > >> As I've said in other emails though, I think the migration/transition >> story from 2.8 -> 3.0 needs work and open communication. Based on what's in >> the videos from the meetup, I don't think it has to be as radical of a >> change as is feared. To make it smoother, hopefully there will be early >> access to the compiler API such that the community can start thinking about >> building 3.0 versions of libraries we want available. >> > > You don't need access to the new compiler for that, you can start > experimenting right now: given that the idea is that the GWT.create() magic > will be removed and things would have to be generated upfront using JavaC > and annotation processors, that means the generated code has no GWT "magic" > (besides possibly JSNI, which would have to eventually change to JsInterop > or "JSNI 2.0") and can still be compiled using the current compiler (or > 2.8.0-SNAPSHOT if you need System.getProperty("user.agent")). > Disclaimer: I haven't yet tried it though. > Note that if you'd like to start experimenting with annotation processors, > I highly suggest you use auto-common's BasicAnnotationProcessor and other > utilities: https://github.com/google/auto/tree/master/common > > -- > You received this message because you are subscribed to the Google Groups > "Google Web Toolkit" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > Visit this group at http://groups.google.com/group/google-web-toolkit. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "Google Web Toolkit" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/google-web-toolkit. For more options, visit https://groups.google.com/d/optout.
