Reza,
since I am new to the project and haven’t followed the previous discussion about archetype, I like to sort out one question for myself:
Will the final starter only provide a maven command to use the archetype and thereby create the project structure (which I think is generally a good idea,
as you can use it without UI) or will there be additionally the possibility to get it as a downloadable zip file?
I think the archetype should be fine for any experienced developer, but people new to the business might be bothered by the additional step (first install
maven, then copy the command from UI, then get the project structure), especially since all other starters out there, especially from the probably most competitive Spring community, work like that and which they see and evaluate in comparison.
Alex
Von: starter-dev <starter-dev-bounces@xxxxxxxxxxx>
Im Auftrag von Reza Rahman
Gesendet: Dienstag, 10. Januar 2023 00:41
An: starter developer discussions <starter-dev@xxxxxxxxxxx>
Betreff: Re: [starter-dev] Project structure change proposal voting
Just so we are all on the same page, the voting is: +1, -1 (optionally as a veto - note that vetoing is pretty extreme and rare at the Eclipse Foundation), and +0 (abstain/neutral). Both committers or anyone else can and should vote. Only the committer vote
is binding. The simple majority carries the day.
My vote is +1.
We have discussed this a lot already and I, Jeyvison, and Bazlur have analyzed our code extensively to try to come up with a good answer. A good answer that will satisfy everyone just doesn't exist and probably never will? We should move forward with a single
Archetype for now and see in the future if there is a better path. One possibility is that we simply have internal modularization for the vendor specific parts (pom.xml, Dockerfile and README.md). Trying to separate by Jakarta EE version is especially painful
right now as we don't even have the basic functionality yet and the code doesn't vary much by Jakarta EE version anyway. So we will find up with 80-90% code duplication, making every little change three times (forever more) and gain very little readability
in the end anyway since the code will still have many conditional renderings given the feature set we are looking to deliver. The code volume will explode from 2-3K+ lines to more like 6-9K+ lines without having added anything in terms of additional functionality.
You can see all this for yourself here if you don't want to take our analysis at face value:
https://github.com/m-reza-rahman/starter/tree/master/archetype.
On 1/9/2023 6:08 PM, Jeyvison Nascimento wrote:
Hello, folks. This is a follow up from the meeting we had today. As talked before and can be checked in
the meeting records,
we wanna have a final decision about the project structure so we can proceed with the modifications as soon as we can.
This proposal is about we keep the monolith approach and reevaluate when needed.
Since this decision will directly affect an on-going house cleaning PR made by Reza, I'd suggest Friday(January 13th) as an end date. Let us know your vote.
_______________________________________________
starter-dev mailing list
starter-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://dev.eclipse.org/mailman/listinfo/starter-dev
