Thank you, Gabriele! I agree that the functionality is useful, however the Apache policy is clear =/
Regards, _____________ Alex Porcelli http://porcelli.me On Thu, Nov 28, 2024 at 7:50 AM Gabriele Cardosi <gabriele.card...@gmail.com> wrote: > Hi Alex, > that's fine for me, no problem. > Reading some of the comments, I had the feeling that maybe something was > not completely clear. > I personally think it would be good to have that demo, because the jit > executor is used by itself, in some cases. > But, again, this can't be a blocker. > I'll try to take a look in my spare time... > > Best > > Gabriele > > Il giorno gio 28 nov 2024 alle ore 13:20 Alex Porcelli <a...@porcelli.me> > ha scritto: > > > Thank you for bringing better visibility of the consequences, Gabriele. > > > > However, as pointed out, unfortunately the way the demo is currently > setup > > - it has important restrictions with Apache policies. > > > > In the end of the day, the general decision has been to allow the whole > > move forward, we accept small sacrifices. > > > > Of course, this could be solved if a different library/license is used to > > accomplish same results. > > > > > > On Thu, Nov 28, 2024 at 3:41 AM Gabriele Cardosi < > > gabriele.card...@gmail.com> > > wrote: > > > > > Hi all. > > > I see multiple approvals on this removal proposal (and I'm not even > sure > > if > > > it is still active). > > > While I have no strong opinion on that, I would like to be sure that it > > is > > > fully clear, for everyone, that removing that bundle.js means, > actually, > > > removing the "demo" itself. > > > The scope of that demo was to provide an easy and OOTB way to: > > > 1 pick any arbitrary dmn file > > > 2 send payload to jitdmn to validate it > > > 3 send payload to jitdmn to execute it with different values, providing > > an > > > automatically-generated form > > > > > > This can't simply be replaced by curl, first because it would be > awfully > > > cumbersome to manually create the required payloads, and second because > > the > > > potential user of that demo could be non-techy users, maybe BA, or > anyone > > > else that anyway does not have time or way to go through the console. > > > And, IMO, this is confirmed by the fact that the demo has not much of a > > > usage for us, as developers: I worked a lot on the jitdmn, and I used > > that > > > only once, because I, as "tech", do not need it. > > > > > > m2c > > > > > > Best > > > > > > Il giorno gio 28 nov 2024 alle ore 09:13 Yeser Amer <ya...@apache.org> > > ha > > > scritto: > > > > > > > +1 > > > > > > > > On 2024/11/26 15:32:31 Jozef Marko wrote: > > > > > Hi everyone, we have an issue with licensing the 'bundle.js' file > for > > > > the the Apache KIE 10 release, see the discussion here [1]. > > > > > > > > > > All this discussion brought me to an idea, do we really need this > > file? > > > > After my investigation, it is used only as the 'jitexecutor-runner' > > demo > > > > app available as 'localhost:8080/index.html' once you start 'java > -jar > > > > jitexecutor-runner/target/quarkus-app/quarkus-run.jar'. > > > > > > > > > > There, you will have a simple webpage with two buttons: > > > > > > > > > > 'Choose file' button: > > > > > This button is for opening a model file. Internally, it calls > > > > 'fetchSchemaForm' from the 'bundle.js' autogenerates form fields for > > the > > > > selected model. > > > > > > > > > > 'Submit' button: > > > > > This is for model evaluation on the 'localhost:8080/jitdmn' > endpoint. > > > It > > > > internally calls 'handleForm' from the bundle.js > > > > > > > > > > Possible actions: > > > > > > > > > > * > > > > > If we check the 'jitexecutor/README' we can do the same demo using > > > > 'curl' command, so if there is no heavy usage of the jitexecutor demo > > UI, > > > > we could remove it, and remove also 'bundle.js' from the codebase > > > > > * > > > > > We could re-implement 'Choose file' and 'Submit' button handlers > > > without > > > > the 'bundle.js' probably, as it is basically invoking a rest endpoint > > > with > > > > some payload, however for sure it is not so straight forward to > > > > autogenerate fields in the webpage. We do such autogeneration in > > > > 'kie-tools' project, what gives as some chance, it is doable as > > > 'kie-tools' > > > > does not use 'bundle.js' > > > > > * > > > > > We could simplify the option above, we could refactor the page, the > > > > 'Choose file' would load 'model' part of the payload for 'jitdmn' > > > endpoint, > > > > but the 'context' payload would user need to provide manually - ie. > no > > > > schema reading and generating form fields by the demo app > > > > > * > > > > > maybe some other option ... > > > > > > > > > > What are your thoughts? Is it possible to remove 'bundle.js' from > the > > > > codebase to solve the licensing issue? If yes, who to do it? Do some > of > > > the > > > > options mentioned sound good for you? > > > > > > > > > > [1] > > > > > > > > > > > > > > > https://github.com/apache/incubator-kie-kogito-apps/pull/2111/files#r1853279114 > > > > > > > > > > > > > > > Jozef Marko > > > > > > > > > > Software Developer > > > > > > > > > > jozef.ma...@ibm.com > > > > > > > > > > > > > > > > > > > > > > > --------------------------------------------------------------------- > > > > To unsubscribe, e-mail: dev-unsubscr...@kie.apache.org > > > > For additional commands, e-mail: dev-h...@kie.apache.org > > > > > > > > > > > > > >
