Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.lang.java.programmer > #39928

Re: Java record

From Daniele Futtorovic <da.futt.news@laposte-dot-net.invalid>
Newsgroups comp.lang.java.programmer
Subject Re: Java record
Date 2023-03-08 21:44 +0100
Organization A noiseless patient Spider
Message-ID <tuas2m$12sbr$1@dont-email.me> (permalink)
References <906f5a76-929d-46cb-b5e0-fcc8e4ad4e0en@googlegroups.com> <tuanur$125a7$1@dont-email.me> <c9a3f6c1-be79-4d17-898a-fb1b3c5eb77dn@googlegroups.com>

Show all headers | View raw


On 08/03/2023 21:14, e.d.pro...@gmail.com wrote:
>>> In web, we use DTOs on the back end to pass data to and from the Javascript/Typescript on the front end. It looks like passing to the front end works, now how do we use it as a model to get the user input?
>> Note: if you do get an exception, you should at least state its message
>> in your post.
>>
>> According to the changelog, support for Java records has been added to
>> Gson in version 2.10:
>> <https://github.com/google/gson/releases?q=2.10&expanded=true>. Switch
>> to that and you should be good to go.
> 
> Aha! So my sample code does work. I hadn't updated my dependencies in a while on my test project. I need to finish that pom and add a site report with the dependency-updates plugin.
> I was referencing gson 2.8.6. I switched to 2.10.1 and it solved that error. Now, if we can switch to Java 17 (15+), we use record for all DTOs? This is much more efficient and/or secure than using classes with lombok?
> 
> All fields in a record are final, so it can only have the one constructor with parameters for every field? This sounds tedious to instantiate on the back end unless you're using a framework or convenience method that's going to populate, from a ResultSet? Front end data is simple if you can instantiate from JSON using gson. Requiring all values final on instantiation will require a very different DTO structure than the legacy apps I've seen.

I doubt it's going to be much different than Lombok in terms of 
performance or reliability. The only difference I can see is in terms of 
clarity of intent. If you start your class with `record`, then you make 
clear for all to see what you intend the class for. The heuristic value 
of that is non-zero.

I haven't had the chance to use records in production yet, so I can but 
mostly speculate. A quick lookup however suggests that you can use 
Lombok's @Builder on `record` classes.

-- 
DF.

Back to comp.lang.java.programmer | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread


Thread

Java record "e.d.pro...@gmail.com" <e.d.programmer@gmail.com> - 2023-03-01 11:18 -0800
  Re: Java record Arne Vajhøj <arne@vajhoej.dk> - 2023-03-01 20:23 -0500
    Re: Java record "e.d.pro...@gmail.com" <e.d.programmer@gmail.com> - 2023-03-02 04:01 -0800
  Re: Java record Daniele Futtorovic <da.futt.news@laposte-dot-net.invalid> - 2023-03-08 20:33 +0100
    Re: Java record "e.d.pro...@gmail.com" <e.d.programmer@gmail.com> - 2023-03-08 12:14 -0800
      Re: Java record Daniele Futtorovic <da.futt.news@laposte-dot-net.invalid> - 2023-03-08 21:44 +0100
        Re: Java record "e.d.pro...@gmail.com" <e.d.programmer@gmail.com> - 2023-03-08 12:54 -0800
      Re: Java record Arne Vajhøj <arne@vajhoej.dk> - 2023-03-13 10:55 -0400
        Re: Java record "e.d.pro...@gmail.com" <e.d.programmer@gmail.com> - 2023-03-13 08:25 -0700
  Re: Java record Jason H <DarthPiriteze@deathstar.gal> - 2023-03-21 18:02 +0000
    Re: Java record "e.d.pro...@gmail.com" <e.d.programmer@gmail.com> - 2023-03-22 04:04 -0700

csiph-web