Potential roadmap #4
Replies: 3 comments 11 replies
-
|
@kaby76 I haven't progressed much on the specs since we talked, because I was trying out WebAssembly stuff that is critical to our initiative. This significantly influences any roadmap because not only is it still too early to engage development, but it's also too early to pick a language for building antlr5. Fundamentally, since the tool requires a runtime, and in order to run the tool in the web, a required step is to produce a WebAssembly runtime with a wrapper for Typescript (wrappers for Java, C# and whatever languages we want to support can come later). A temporary option could be to use GraalVM to run the existing JS runtime until we have a WebAssembly one. To be explored... Another option is Dart, which seems to have mature support for WebAsssembly, and for which we already have a runtime. All the above only influences the technical side of things though, and I believe we could start working on the functional side starting from the current tech stack. It might actually be easier to implement new features in Java and migrate them later if required, rather than migrate the code base immediately and implement the new features using a less mature toolset (in the meantime, we might find that Java supports WebAssembly well enough to not migrate at all, and generate the WebAssembly from the Java code). From the above, a potential 1st phase technical roadmap could look as follows:
In parallel, we could progress on the functional front:
Let me know what you think |
Beta Was this translation helpful? Give feedback.
-
|
BTW, I've started work on code-base improving and optimisation in the mid of 2022: https://github.com/KvanTTT/antlr4/commits/dev/ It's back-portable, all test are passing. |
Beta Was this translation helpful? Give feedback.
-
Why can't we just continue work on the project in a separated branch since the core remains the same (at least at the beginning). @parrt opinion is required here. |
Beta Was this translation helpful? Give feedback.
-
|
How do you propose to deal with PRs and a completely different CI/CD ? What benefits do you expect from working in an antlr4 branch ? |
Beta Was this translation helpful? Give feedback.
-
|
It sounds like you guys are interested in a major change to the functionality and a complete rewrite so it makes sense to start a new repository |
Beta Was this translation helpful? Give feedback.
-
Just to continue development with the current code base CD and switch to a new one if it's impossible. But nevermind, let's clone git repository (with keeping history). |
Beta Was this translation helpful? Give feedback.
-
|
@ericvergnaud Thank you for taking the initiative here!! I think it's important to consider the next incarnation of Antlr. It's a wonderful tool but let's see where we can take it to address some of the problems. By the way, you might want to start a discussion about the problems, if for nothing else but a place to sort of vent! As this will take some time, it would be nice to have another minor or patch release of Antlr4. |
Beta Was this translation helpful? Give feedback.
-
|
Let's have those parallel discussions in Rants and keep this one for the antlr5 roadmap. |
Beta Was this translation helpful? Give feedback.
-
|
Also, are you sure you want to make a v5 in Java? I think Go would be a lot
better these days - simpler and faster (though I don't think performance is
that big a deal with the tool these days). But then again, I would promote
Go wouldn't I? ;)
Have you written up anything about WASM being the only target
anywhere Eric? I must admit that I am a bit skeptical about that, but I am
prepared to be convinced. I have seen various emails and so on flying
around.
Some grist for the mill here:
How about syntax directed lexing? Many compilers with hand written parsers
and lexers will do this. It is though one of the reasons that some
languages are hard to map formally into a specification like ANTLR. I was
thinking of an optional mode. It is a switch in the way the whole thing
goes together though.
Jim
…On Fri, Dec 29, 2023 at 9:47 PM Ivan Kochurkin ***@***.***> wrote:
copy antlr4 repo to an antlr5 repo
Why can't we just continue work on the project in a separated branch since
the core remains the same (at least at the beginning). @parrt
<https://github.com/parrt> opinion is required here.
—
Reply to this email directly, view it on GitHub
<#4 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJ7TME6OSYZOCTDTITHQALYL3CXHAVCNFSM6AAAAABAW6YMZCVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM3TSNZSGI3TS>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
Beta Was this translation helpful? Give feedback.
-
See #6 |
Beta Was this translation helpful? Give feedback.
-
Is that a different wording for #10 ? |
Beta Was this translation helpful? Give feedback.
-
|
Kotlin looks like a good compromise, see discussion #16 (comment) |
Beta Was this translation helpful? Give feedback.
-
I think this is part of the mission statement for ANTLR 5, described in the README: https://github.com/antlr/antlr5-specs/blob/master/README.md
I think in #16 is where we argue about the development language to use and Go is listed as one of the options considered |
Beta Was this translation helpful? Give feedback.
-
Phase 1:
Beta Was this translation helpful? Give feedback.
All reactions