[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: proper way to resolve dependency version conflicts of rust crates?

From: Ignas Lapėnas
Subject: Re: proper way to resolve dependency version conflicts of rust crates?
Date: Mon, 08 Jan 2024 12:02:29 +0200
User-agent: mu4e 1.10.8; emacs 30.0.50


> 1. Sent a patch to the upstream of the crate as well as to guix updating
>   all crate dependencies to most-recent.
Sending a patch upstream updating package versions is probably desired by the
guix rust-team, can’t actually answer that. Contributing is another beast to
tackle for me some time in the future. Best guess, guix-devel mailing list
should provide an answer for that.

> 2. Define a package variant of said dependency, let’s say foo-0.13-1 for
>   v0.13.1 of foo.
I doubt that this is necessary if the package in question follows semver rules.

> 3. Substitute the required version in Cargo.toml of a affected crate to
>   match the package-version of said dependency we’ve packaged in Guix.
> If I had to guess 3. would be the most appropriate way forward on this?
I would go with option 3.

Currently most packages I see drops the ’=’ altogether in favor of ’^’.
For example the rust-base64 package:

│ (define-public rust-base64-0.13
│   (package
│     (inherit rust-base64-0.21)
│     (name "rust-base64")
│     (version "0.13.1")
│     (source
│      (origin
│        (method url-fetch)
│        (uri (crate-uri "base64" version))
│        (file-name
│         (string-append name "-" version ".tar.gz"))
│        (sha256
│         (base32 "1s494mqmzjb766fy1kqlccgfg2sdcjb6hzbvzqv2jw65fdi5h6wy"))
│        (modules '((guix build utils)))
│        (snippet
│         '(begin (substitute* "Cargo.toml"
│                   (("=0\\.3\\.2") "^0.3.2"))))))
│     (arguments
│      `(#:cargo-development-inputs
│        (("rust-criterion" ,rust-criterion-0.3)
│         ("rust-rand" ,rust-rand-0.6)
│         ("rust-structopt" ,rust-structopt-0.3))))))

Hope my blabering helps you mate.

Best Regards,
Ignas Lapėnas

reply via email to

[Prev in Thread] Current Thread [Next in Thread]