Saturday, June 1, 2019

What is the advantage of GWT 3.0?

I love GWT 2.8 and I appreciate all the work that the community has put into it as well as the work towards GWT 3.0. I'm trying to understand what the advantage of GWT 3.0 is though. It seems like GWT 3.0 is a subset of GWT 2.8 with a different compiler under the covers. How is the closure compiler better than the GWT 2 compiler and is it really worth trying to switch when GWT 2.8 is mature and works well?

--
You received this message because you are subscribed to the Google Groups "GWT Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-web-toolkit/4d1411b2-4e90-4972-a59c-df3cf59ec017%40googlegroups.com.

Re: GWT - Still Active ?



On Sunday, June 2, 2019 at 1:19:40 AM UTC+3, Douglas de Oliveira Mendes wrote:
Ain't dead for me! Using it every day quite a lot. I still think it's a great option if you have tons of widgets, panels, tabs... I see GWT basically as an easier to distribute Java Web Start-like plataform.

I agree specially when you have lots of java developer in your workplace, it is easier for them to start with gwt than other framworks, 
 

I'm sitcking with RPC because it maximizes client-server code reuse. DTOs, Enums, Exceptions, utility classes... We have Java and Groovy on the server side. I believe RPC maximizes the strenghts of this setup. Some colleagues dislike the calls being serialized. But I think if that's the problem they'd just go after replacing the serialization of the RPC mecanism to a JSON based one (maybe this "jackson" thing someone mentioned?) instead of rewriting the RPCs into anything else. The server side is ours, we write it in Java so let's stick with RPC, don't you think?

I would like to use validation on the DTOs keeping them as POJOs (or something like, closest to pure Java), but the GWT website scared me off regarding that [1]. Another desire I have is avoiding interfaces having to extend "RemoteSevice". Should be enough to write some XML or whatever to tell which services are exposed, hence keeping the interfaces with no dependencies to GWT.

I think you should take a look at domino-rest and i promise you wont regret it, it is JaxRs compatible, works flawlessly in both gwt2 and gwt3/j2cl, you can use the same generated client in both client and server, interfaces are very clean, serialization is based on dominon-jackson so you can use DTOs with enums, collections, arrays ..etc. and i know that @Frank already used it to completely replace restyGwt from one of his apps. 
 

As for widgets, we mostly use Sencha GXT's components.

if you use lots of widgets i suggest you take a look at domino-ui you wont regret it i promise again:-). 


I'm not sure if default methods in interfaces are already available. Would be nice to have.

Java8 syntax is supported in gwt-2.8.2, i use default methods in interfaces all the time and they are pretty useful.
 

I also miss debugging but I'm not sure if GWT should take any blame or just my laziness for setting it back up.

[1] "CAUTION: Validation API is unmaintained and will be moved out of GWT SDK into a separate project after GWT 2.8."  http://www.gwtproject.org/doc/latest/DevGuideValidation.html

Next in my new libs queue is domino-validation stay tuned for it. :-)
 


Douglas


On Sat, Jun 1, 2019 at 4:52 PM Andrew Buck <anb...@gmail.com> wrote:
GWT is not dead! It's simply suffering from PR misunderstanding. People think that you have to use the old widget system to use GWT, but you don't. Just use Elemento instead of widgets and REST calls instead of RPC.  Regardless of what happens with GWT 3, using GWT 2.8 is future proof since it supports JsInterop. GWT remains the most robust system to develop web apps with. Also, combined with J2ObjC, it remains the only way to write 100% native apps for web, iOS, and Android and share 70% of the code between all platforms. Other systems that let you share code don't produce truly native apps.



On Friday, May 31, 2019 at 8:27:45 PM UTC-7, Peter Donald wrote:
On Fri, May 31, 2019 at 10:44 AM Craig Mitchell <ma...@craig-mitchell.com> wrote:
Off topic:  I do wonder how web assembly (WASM) is going to impact GWT, especially if it gets garbage collection, and therefore makes Java to WASM compilation possible.

That is the biggest risk IMO. When we did our analysis to decide on whether to commit to J2CL/GWT3.x for the next 10 years or not this was the only real risk that we found (or that Typescript gets a lot better backend).

WebAssembly is still a way off but projects like https://github.com/i-net-software/JWebAssembly do seem to be something to watch


--
Cheers,

Peter Donald

--
You received this message because you are subscribed to the Google Groups "GWT Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-web-toolkit/dfe83c33-bdee-4dd8-ba78-7db137a81333%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "GWT Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-web-toolkit/672a73e0-eb2a-4cee-a9ee-c028cf46c880%40googlegroups.com.

Re: GWT - Still Active ?



On Saturday, June 1, 2019 at 10:51:55 PM UTC+3, Andrew Buck wrote:
GWT is not dead! It's simply suffering from PR misunderstanding. People think that you have to use the old widget system to use GWT, but you don't. Just use Elemento instead of widgets and REST calls instead of RPC.  Regardless of what happens with GWT 3, using GWT 2.8 is future proof since it supports JsInterop. GWT remains the most robust system to develop web apps with. Also, combined with J2ObjC, it remains the only way to write 100% native apps for web, iOS, and Android and share 70% of the code between all platforms. Other systems that let you share code don't produce truly native apps.



I agree with you, people should start thinking of gwt as just the compiler, they dont need to use widgets to call it GWT nor any of the gwt built in modules, there is a lot of modern libs out there that really competes, i built domino-ui without using anything related to widgets, all i am using is elemental2, elemento, and few of the ported gwt modules, and it works with both gwt2 and j2cl -at least i have a branch working with j2cl- not only this one but a lot more of good tool like my new domino-rest and what about Vue-gwt. gwt is really just the compiler and the ecosystem around it is really amazing and it is more amazing of what you can do with it.
 

On Friday, May 31, 2019 at 8:27:45 PM UTC-7, Peter Donald wrote:
On Fri, May 31, 2019 at 10:44 AM Craig Mitchell <ma...@craig-mitchell.com> wrote:
Off topic:  I do wonder how web assembly (WASM) is going to impact GWT, especially if it gets garbage collection, and therefore makes Java to WASM compilation possible.

That is the biggest risk IMO. When we did our analysis to decide on whether to commit to J2CL/GWT3.x for the next 10 years or not this was the only real risk that we found (or that Typescript gets a lot better backend).

WebAssembly is still a way off but projects like https://github.com/i-net-software/JWebAssembly do seem to be something to watch


--
Cheers,

Peter Donald

--
You received this message because you are subscribed to the Google Groups "GWT Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-web-toolkit/2581803b-09fb-4d2f-9dba-1267aa477662%40googlegroups.com.

Re: GWT - Still Active ?

On Sunday, June 2, 2019 at 1:22:09 AM UTC+3, Edson Richter wrote:
This is my opinion (from a person with more than 30 years of experience in software enginnering), and I respect every other persons opinion - I'll just not put my eggs on this basket for another 8 years "just to see if it will get better".
Last year I had big issue with dates, because GWT has outdated DST tables (backend had correct dates, front end show everything one hour earlier!!!). This is just one little big problem for enterprise apps. I've asked support for this issue here, without any result (if someone had pointed me how to fix, I'll contribute code back to the project!). Finally, I had to write my own DST table and use "creativity" to overcome this "bug".
Not having support is also a big issue.

I can understand your frustration regarding this, but i wont call GWT just because of this, specially that i know the CLDR in gwt have been updated to the recent data, we still need to generate an updated DateTimeZone constants file which is delayed because the generator is not found yet at google, but we didnt stop there, and i already started a new tool to generate this file, still not finished and needs more tests and review, but my point is that updating GWT was slow because we had to go through the gateways of google, but now as a community we can do things faster, all we ask is that the community should also help, it is not in the hands of google anymore, it is in our hands.
 
Not being able to debug the code anymore is another one.

Can we blame GWT on this, the old dev mode have been deprecated for a long time and is not even supported anymore by any of the browsers, but in the same time browsers are getting better at debugging javascript, and this is what we are actually do, i always debug in the browsers, and i dont think that any other Js framework will give you any better options in this regards. in my case i still can debug my code as java code in the IDE since all my business logic resides in presenters that has nothing to do with gwt ;-).
 
More and more, GWT will get those "pieces" failing... and finally, it won't be usable anymore.

I am not sure about this, instead of saying falling i would say migrated or ported, we are already working on the next gwt version and one really good thing we are doing is making sure that any migrated or ported module from gwt2 should work in both gwt2 and gwt3, and so for we have already ported may modules and there is no need to wait for next gwt version to start using them , as you can start using then right now and mix them with what ever you use from gwt2 and be ready for the next gwt version.
 

Regards,

Edson


Em sexta-feira, 31 de maio de 2019 11:41:27 UTC-3, Jamal Romero escreveu:
Out of curiosity, what would prevent someone still build projects based on current GWT 2.8.2 and keep using all the goodies? I think as of 2.8.2 it is future proof especially with a shift to jsinteop included in current release version? People even with current version took their own path and modernized part of GWT like the excellent gwt material & domino ui kit.

--
You received this message because you are subscribed to the Google Groups "GWT Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-web-toolkit/33322063-8e51-455e-b8d3-e139c4db1a0a%40googlegroups.com.

Re: GWT - Still Active ?

This is my opinion (from a person with more than 30 years of experience in software enginnering), and I respect every other persons opinion - I'll just not put my eggs on this basket for another 8 years "just to see if it will get better".
Last year I had big issue with dates, because GWT has outdated DST tables (backend had correct dates, front end show everything one hour earlier!!!). This is just one little big problem for enterprise apps. I've asked support for this issue here, without any result (if someone had pointed me how to fix, I'll contribute code back to the project!). Finally, I had to write my own DST table and use "creativity" to overcome this "bug".
Not having support is also a big issue.
Not being able to debug the code anymore is another one.
More and more, GWT will get those "pieces" failing... and finally, it won't be usable anymore.

Regards,

Edson


Em sexta-feira, 31 de maio de 2019 11:41:27 UTC-3, Jamal Romero escreveu:
Out of curiosity, what would prevent someone still build projects based on current GWT 2.8.2 and keep using all the goodies? I think as of 2.8.2 it is future proof especially with a shift to jsinteop included in current release version? People even with current version took their own path and modernized part of GWT like the excellent gwt material & domino ui kit.

--
You received this message because you are subscribed to the Google Groups "GWT Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-web-toolkit/7c39c7bb-4619-45aa-9fc3-6069ce1fd345%40googlegroups.com.

Re: GWT - Still Active ?

Ain't dead for me! Using it every day quite a lot. I still think it's a great option if you have tons of widgets, panels, tabs... I see GWT basically as an easier to distribute Java Web Start-like plataform.

I'm sitcking with RPC because it maximizes client-server code reuse. DTOs, Enums, Exceptions, utility classes... We have Java and Groovy on the server side. I believe RPC maximizes the strenghts of this setup. Some colleagues dislike the calls being serialized. But I think if that's the problem they'd just go after replacing the serialization of the RPC mecanism to a JSON based one (maybe this "jackson" thing someone mentioned?) instead of rewriting the RPCs into anything else. The server side is ours, we write it in Java so let's stick with RPC, don't you think?

I would like to use validation on the DTOs keeping them as POJOs (or something like, closest to pure Java), but the GWT website scared me off regarding that [1]. Another desire I have is avoiding interfaces having to extend "RemoteSevice". Should be enough to write some XML or whatever to tell which services are exposed, hence keeping the interfaces with no dependencies to GWT.

As for widgets, we mostly use Sencha GXT's components.

I'm not sure if default methods in interfaces are already available. Would be nice to have.

I also miss debugging but I'm not sure if GWT should take any blame or just my laziness for setting it back up.

[1] "CAUTION: Validation API is unmaintained and will be moved out of GWT SDK into a separate project after GWT 2.8."  http://www.gwtproject.org/doc/latest/DevGuideValidation.html

Douglas


On Sat, Jun 1, 2019 at 4:52 PM Andrew Buck <anbuck@gmail.com> wrote:
GWT is not dead! It's simply suffering from PR misunderstanding. People think that you have to use the old widget system to use GWT, but you don't. Just use Elemento instead of widgets and REST calls instead of RPC.  Regardless of what happens with GWT 3, using GWT 2.8 is future proof since it supports JsInterop. GWT remains the most robust system to develop web apps with. Also, combined with J2ObjC, it remains the only way to write 100% native apps for web, iOS, and Android and share 70% of the code between all platforms. Other systems that let you share code don't produce truly native apps.



On Friday, May 31, 2019 at 8:27:45 PM UTC-7, Peter Donald wrote:
On Fri, May 31, 2019 at 10:44 AM Craig Mitchell <ma...@craig-mitchell.com> wrote:
Off topic:  I do wonder how web assembly (WASM) is going to impact GWT, especially if it gets garbage collection, and therefore makes Java to WASM compilation possible.

That is the biggest risk IMO. When we did our analysis to decide on whether to commit to J2CL/GWT3.x for the next 10 years or not this was the only real risk that we found (or that Typescript gets a lot better backend).

WebAssembly is still a way off but projects like https://github.com/i-net-software/JWebAssembly do seem to be something to watch


--
Cheers,

Peter Donald

--
You received this message because you are subscribed to the Google Groups "GWT Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-web-toolkit/dfe83c33-bdee-4dd8-ba78-7db137a81333%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "GWT Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-web-toolkit/CADkeT6q8RwLwS%3DkN-Nz5A6jy8TRuas3VCTq3aCvSrj%3Dm4a3zyQ%40mail.gmail.com.

Re: GWT - Still Active ?

GWT is not dead! It's simply suffering from PR misunderstanding. People think that you have to use the old widget system to use GWT, but you don't. Just use Elemento instead of widgets and REST calls instead of RPC.  Regardless of what happens with GWT 3, using GWT 2.8 is future proof since it supports JsInterop. GWT remains the most robust system to develop web apps with. Also, combined with J2ObjC, it remains the only way to write 100% native apps for web, iOS, and Android and share 70% of the code between all platforms. Other systems that let you share code don't produce truly native apps.



On Friday, May 31, 2019 at 8:27:45 PM UTC-7, Peter Donald wrote:
On Fri, May 31, 2019 at 10:44 AM Craig Mitchell <ma...@craig-mitchell.com> wrote:
Off topic:  I do wonder how web assembly (WASM) is going to impact GWT, especially if it gets garbage collection, and therefore makes Java to WASM compilation possible.

That is the biggest risk IMO. When we did our analysis to decide on whether to commit to J2CL/GWT3.x for the next 10 years or not this was the only real risk that we found (or that Typescript gets a lot better backend).

WebAssembly is still a way off but projects like https://github.com/i-net-software/JWebAssembly do seem to be something to watch


--
Cheers,

Peter Donald

--
You received this message because you are subscribed to the Google Groups "GWT Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-web-toolkit/dfe83c33-bdee-4dd8-ba78-7db137a81333%40googlegroups.com.