gwtproject/gwt

GWT Open Source Project

flatlander

@flatlanderer_gitlab

maybe 10 years more? haha

Colin Alworth

@niloc132

the stated goal of "gwt 3" is that the ecosystem works on either gwt2 or j2cl, and it is up to the user to decide which they want, which solves their problems

if someone needs it to happen and is willing to work with me and my company to update (or do the work them selves), sure, why not?

going back 10 years and seeing what js tooling is still leading the way is an education ;)

flatlander

@flatlanderer_gitlab

yeah that makes sense, hopefully we call can build prosperous apps with GWT, and who knows, I think this is important that apps that sells will come like from what I know Google is just focused with Flutter and building apps that generates revenues (as well as other companies) thus it have this "money" in the ecosystem which makes it move really fast, last I heard there's flutter 2 for web and stable

Tested it and it doesn't feel good

Colin Alworth

@niloc132

i've heard nice things, aside from ios native performance of animations - dart didn't really impress me back in the day, and i think typescript has some wonderful and also terrible features, but i still find gwt to be more productive for the kinds of long-term projects i'm interested in enabling

flatlander

@flatlanderer_gitlab

yeah, the web is sluggish I tried, and its a canvas, a really big canvas

to be fair, btw, we use flutter for mobile ui, but for web, it doesn't work for us yet

but seriously, it got me thinking, is canvas the future of web? I mean like its like a browser inside a browser, it would be like inception

Colin Alworth

@niloc132

unless i'm very much mistaken, there is no way to provide good accessibility for canvas based impls, so i'm frankly terrified of such approaches

screenreaders, user stylesheets, etc

keyboard nav

flatlander

@flatlanderer_gitlab

yeah

you can basically build great UI with Tailwind CSS

instead

flatlander

@flatlanderer_gitlab

btw, found this while googling: http://gurselkoca.blogspot.com/2011/03/how-to-display-progressbar-during-gwt.html

Colin Alworth

@niloc132

mm, that's not enough

that just gets you a hook to know "hey it is uploading, draw a fake progress bar", not the actual one you want to count bytes and such

which ok, that works

flatlander

@flatlanderer_gitlab

you mentioned a hack right? when you have the idea please let me know

Dr. Lofi Dewanto

@lofidewanto_twitter

@ainslec Better late than never ;-) I put every update stuffs (articles, presentations) at this Padlet: https://padlet.com/lofidewanto/gwtintro

@All: If you have something, you could add into that Padlet by yourself and I'll check it before I approve...

Dr. Lofi Dewanto

@lofidewanto_twitter

@ainslec If you need to build a realworld webapp with IndexedDB (just like JPA / Hibernate), this article would be useful with all the Design Patterns in Java: https://bit.ly/GwtIndexedDBDesignPatterns

Michele Pollini

@MichelePollini

Hi! We have a old gwt application (v 2.8.2) and on chrome 89 version we have error 500 during some server calls. With version 88 is all good. Anyone have same issue? Thanks

Thomas Broyer

@tbroyer

What's the actual error behind that 500?

Michele Pollini

@MichelePollini

08-Mar-2021 12:37:07.989 SEVERE [http-nio-8090-exec-8] org.apache.catalina.core.ApplicationContext.log Exception while dispatching incoming RPC call
java.lang.NumberFormatException: Expected type 'int' but received a non-numerical value: �a�Ħ��9?�I��G1�,�!��$t��
�/s�Ħ��94��[2IM!œNԇS�5#��WT�J��V�o��>�5#��t�g҆�s��w��c�� ]��ѐ

��˻��5i�~�� .�Bӌ޾+��q��Z�jC�t9�����g��<CE��v�N��ov��!r���������ћދ�促�����}K��    Ѫ�v��-��׺�{6�{��̒v����M�Ƀ+��?)����K�4��fjx=��    ��R�50~ݪ��[���6�"g�{Y�ץ�*�fQܶ�0F�Y�u�K ��Hyn�=�ڶ�--���x��o

�Н��5$B凧��4�m��}�7��;9K��
��U��K�[��uT"�m^��˙$�m{�p�E�W��~S$5cfD�zi��{��j��Uה��׆m�D�J��K�U��"��m���l릖<AڑL81BĸZu��y�>�[p:��o5�q�O�n��U\p��w��ǲ9�Ï�ˤ��^G��5e��m瞎N�}��#��K�qs��I�Qß�R��7��-zsmo��ۍ:�ڏA O K7
at com.google.gwt.user.server.rpc.impl.ServerSerializationStreamReader.getNumberFormatException(ServerSerializationStreamReader.java:1027)
at com.google.gwt.user.server.rpc.impl.ServerSerializationStreamReader.readInt(ServerSerializationStreamReader.java:551)
at com.google.gwt.user.client.rpc.impl.AbstractSerializationStreamReader.prepareToRead(AbstractSerializationStreamReader.java:96)
at com.google.gwt.user.server.rpc.impl.ServerSerializationStreamReader.prepareToRead(ServerSerializationStreamReader.java:477)
at com.google.gwt.user.server.rpc.RPC.decodeRequest(RPC.java:240)
...

We are currently investigate about it

Michele Pollini

@MichelePollini

It occur only with https

Thomas Broyer

@tbroyer

It looks like the request payload may be gzipped or something like that, and is not ungzipped before being handed to the servlet.

Colin Alworth

@niloc132

i dont think its gzip (unless somehow the first few bytes were removed?), but for sure something funny happened to just the payload and probably not the request line, headers for it to get this far. if you have the rest of the headers (see if chrome is saying what it is sending differently), rest of the stack trace (see which filters could be causing problems here) that might help pin it down

Colin Alworth

@niloc132

@flatlanderer_gitlab in short, once you have the Request object, use JSNI to access its xhr field, and then do the JS snippet you posted above to get the notification when it changes

Chris Ainsley

@ainslec

Hello, trying my best to understand how JsInterop works with JavaScript methods that return a particular complex data structure

For example, I want to be able to process the data from a call to this:

var arrayOfKeyValueArrays = Object.entries(localStorage)

This will return an array containing two element arrays, first element key, second element value

I just want to essentially map this to something like a String[][]

I don't know the correct way of mapping this using JsInterop

Colin Alworth

@niloc132

@ainslec this is a very good question that gets to a lot of different things, including how JS kinda plays fast and loose with things that java would get mad about - i mean, String[][] is okay if your object only has string values, but for a complex structure of an object passed in, or something else like Number or Boolean, you obviously can't use that, you'd need Object[][]

it is informative to read how elemental2-core doe sot: JsObject.entries returns a jsarray of a jsarray of a union type, and the union type is generic

you can ask each entry "are you a string, or are you a T" - but of course, you the reader don't need to check this, since you can just use index 0 or 1 and already know what the type is

in gwt2, JsArray<JsArray<String>> (or JsString) is technically more correct than String[][], esp if you end up casting the contents of the array (a String[] is instanceof CharSequence[] in java, but if is a js array, it has no type data on it)

but of course your use case of localStorage isnt terribly complex, all values are strings, so you really have your pick of approaches here

Chris Ainsley

@ainslec

So hypothetically, if using JSInterop, what would be the wrapper method for returning Object[][] ? Or is JSNI the way to go?

Colin Alworth

@niloc132

well, JsObject.entries exists, so consider using that

if you dont want to use that, making your own would be roughly

@JsMethod(name="Object.entries", namespace=JsPackage.GLOBAL)
static native String[][] objectEntries(Object value);

the annotation (and native modifier) says "this exists already at this path"

if you wanted Object[][] that would be fine too, and just cast things as they come out

anything* jsni can do, jsinterop can do better

(*there are exceptions, but almost all of them are already managed in the jsinterop-base jar, that way you dont have to think about them)

Where communities thrive

People

Repo info

Activity