|
From: | John Snow |
Subject: | Re: [PATCH] PoC: Rust binding for QAPI (qemu-ga only, for now) |
Date: | Fri, 11 Sep 2020 10:19:53 -0400 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.11.0 |
On 9/11/20 7:28 AM, Paolo Bonzini wrote:
On 11/09/20 12:46, Daniel P. Berrangé wrote:Do we actually need/want it to be in the same monolithic repo as qemu, as opposed to a qemu-qapi-rust repo ?I think QAPI and qemu-ga should be separate repos altogether. QAPI should be included as a submodule in both qemu and qemu-ga. qemu-ga instead has absolutely no dependency on QEMU and viceversa, and is a prime candidate for removing all traces of the configure script and being a pure meson project. Paolo
I am still actively invested in moving ./scripts/qapi to ./python/qapi and turning it into its own installable module as I am doing with ./python/qemu.
The benefit is that it could be used for the build system in other projects; or possibly new backend plugins can be developed and maintained separately for it.
(I am working on a JSON-SCHEMA backend.)((Also, I want to enforce a single flake8/pylint/mypy regime for all of our critical Python code. Considered critical: testing infrastructure, anything used for the build process. This means everything in ./python/qemu, the tracetool, and the QAPI parser.))
I have a 20-30 patch series doing the usual mypy/flake8/pylint thing to the QAPI parser and it will be able to stand alone in short time.
--js
[Prev in Thread] | Current Thread | [Next in Thread] |