Re: Strange GWT-RPC bug upgrading from 2.7 to 2.8+

2020-10-07 Thread Jens
Well ok that devirtual function is clearly broken. Normally these kind of functions do instanceof dispatches to delegate an abstract method call, e.g. Number.intValue(), to a concrete implementation, e.g. Integer.intValue. Below an example how it should look (GWT master branch with an example ap

Re: Strange GWT-RPC bug upgrading from 2.7 to 2.8+

2020-10-07 Thread Aaron Davidson
Thanks for the tip Lofi -- I don't know about vertispan, but looks like a great option for us to check out! On Tuesday, October 6, 2020 at 11:15:19 AM UTC-7 lofid...@gmail.com wrote: > If you need a fast solution I would recommend to use professional help > like Vertispan: https://www.vertispan

Re: Strange GWT-RPC bug upgrading from 2.7 to 2.8+

2020-10-06 Thread Michael Conrad
I don't know if this applies to your situation, but we had an issue similar to this (a long time ago) that was caused by shadowing properties from parent classes in child classes. On Tue, Oct 6, 2020 at 2:15 PM lofid...@gmail.com wrote: > If you need a fast solution I would recommend to use prof

Re: Strange GWT-RPC bug upgrading from 2.7 to 2.8+

2020-10-06 Thread lofid...@gmail.com
If you need a fast solution I would recommend to use professional help like Vertispan: https://www.vertispan.com Colin knows the transpiler inside out 😉 @Jens: thanks for your debugging tips, I would add your tips to my GWT deck. pha...@gmail.com schrieb am Sonntag, 4. Oktober 2020 um 05:16:11 U

Re: Strange GWT-RPC bug upgrading from 2.7 to 2.8+

2020-10-02 Thread Jens
> > And there you have it... absolutely no idea why it's malfunctioning in > this way. Any help is greatly appreciated! > Ok in sourcemaps it looks indeed a bit weird. However there will be additional JS code executed to convert java.lang.Integer into a primitive int which is represented by JS

Re: Strange GWT-RPC bug upgrading from 2.7 to 2.8+

2020-10-01 Thread Aaron Davidson
Hmm, we've got client and server code in separate packages, but I don't know exactly what the configuration would look like for separate project classpaths. If you feel like having a peek with me over a zoom call to look further, let me know! On Wednesday, September 30, 2020 at 1:06:10 PM UTC-

Re: Strange GWT-RPC bug upgrading from 2.7 to 2.8+

2020-10-01 Thread Jens
First of all, when updating GWT SDK you should delete your gwtUnitCache directory. Historically there has always been some hiccups with the cache when upgrading GWT. By deleting it you make sure to start a fresh compile without any cached information. If that does not help you should take the s

Re: Strange GWT-RPC bug upgrading from 2.7 to 2.8+

2020-09-30 Thread lofid...@gmail.com
Did you separate the modules in: - client - shared - server ... with its own classpath? Or did you put them together in *one project with one classpath*? The latter is not good, because you may get "classpath hell" so GWT transpiler uses wrong libs to transpile the Java code. I defin

Re: Strange GWT-RPC bug upgrading from 2.7 to 2.8+

2020-09-30 Thread Joker Joker
As an experiment, I would try changing the "offset" parameter type from "Integer" to "int" in both versions of the method. You can also try deleting all gwt-unitCache files. понедельник, 28 сентября 2020 г. в 12:43:11 UTC+4, pha...@gmail.com: > I have a large GWT app and we've been stuck on GWT

Strange GWT-RPC bug upgrading from 2.7 to 2.8+

2020-09-28 Thread Aaron Davidson
I have a large GWT app and we've been stuck on GWT 2.7 for years. If I try to update to 2.8 or 2.9, a random selection of our GWT-RPC calls are not encoded properly by the client. Certain fields in the request become 'undefined' (example below), and cannot be decoded. We *REALLY* want to upgrad