update packages to rustls v0.23.46 to address many security advisories #43

Open
opened 2026-10-14 20:42:00 +00:00 by decathorpe · 9 comments
Owner

The latest release of rustls, version 0.23.46, addresses four low-to-medium security vulnerabilities:

This is in addition to many other security fixes that have only landed in the v0.23 branch of rustls. Applications that still depend on rustls v0.22 or even v0.21 will need to be ported to newer versions of their dependencies.

Dependent applications:

rustls v0.23:

  • asciinema
  • atuin
  • aw-server-rust
  • fips-provider
  • kryoptic
  • ntpd-rs
  • rust-cargo-rpmstatus
  • rust-cargo-semver-checks
  • rust-gst-plugin-reqwest
  • rust-rbw
  • rust-routinator
  • rust-sccache
  • rustup
  • sandogasa
  • selenium-manager
  • uv

rustls v0.22:

  • rust-redlib (via redis + hyper-rustls v0.25)

rustls v0.21:

  • awatcher (via reqwest v0.11)
  • rust-resctl-bench (via rust-minreq)
  • rust-app-store-connect (via reqwest v0.11)

Bundled rustls:

  • arapuca
  • bpfman
  • cosmic-initial-setup
  • cosmic-store
  • flare-im
  • fractal
  • goose
  • pnpm12
  • pop-launcher
  • task
  • trunk
  • vaultwarden

Yet unclear affectedness:

These packages pull in rust-rustls-devel but don't contain a LICENSE.dependencies file so it cannot be checked directly whether it is actually linked into the built executables.

  • matrix-synapse
  • pdns-recursor

See also this mailing list post about dropping old and unsupported rustls versions.

The latest release of rustls, version 0.23.46, addresses four low-to-medium security vulnerabilities: - <https://rustsec.org/advisories/RUSTSEC-2026-0346.html> - <https://rustsec.org/advisories/RUSTSEC-2026-0347.html> - <https://rustsec.org/advisories/RUSTSEC-2026-0348.html> - <https://rustsec.org/advisories/RUSTSEC-2026-0349.html> This is in addition to many other security fixes that have only landed in the v0.23 branch of rustls. Applications that still depend on rustls v0.22 or even v0.21 will need to be ported to newer versions of their dependencies. Dependent applications: **rustls v0.23**: - [x] asciinema - [x] atuin - [ ] aw-server-rust - [x] fips-provider - [x] kryoptic - [x] ntpd-rs - [x] rust-cargo-rpmstatus - [x] rust-cargo-semver-checks - [x] rust-gst-plugin-reqwest - [x] rust-rbw - [x] rust-routinator - [x] rust-sccache - [x] rustup - [x] sandogasa - [x] selenium-manager - [x] uv **rustls v0.22**: - [ ] rust-redlib (via redis + hyper-rustls v0.25) **rustls v0.21**: - [ ] awatcher (via reqwest v0.11) - [ ] rust-resctl-bench (via rust-minreq) - [ ] rust-app-store-connect (via reqwest v0.11) **Bundled rustls**: - [ ] arapuca - [ ] bpfman - [ ] cosmic-initial-setup - [ ] cosmic-store - [ ] flare-im - [ ] fractal - [ ] goose - [x] pnpm12 - [ ] pop-launcher - [ ] task - [ ] trunk - [ ] vaultwarden **Yet unclear affectedness**: These packages pull in rust-rustls-devel but don't contain a LICENSE.dependencies file so it cannot be checked directly whether it is actually linked into the built executables. - [ ] matrix-synapse - [ ] pdns-recursor --- See also this [mailing list post](https://lists.fedoraproject.org/archives/list/rust@lists.fedoraproject.org/thread/G3OXXZRNGCZGJ6XVN5INOQ2GVNMNIKRN/) about dropping old and unsupported rustls versions.
decathorpe changed title from update packages to and rebuild packages with rustls v0.23.46 to address many security advisories to update packages to rustls v0.23.46 to address many security advisories 2026-10-14 20:44:16 +00:00
Author
Owner

Builds of rustls v0.23.46 are now available:

Dependent packages can either be rebuilt into these side-tags and attached to the existing updates, or built directly for the corresponding release(s) after these updates are stable.

Builds of rustls v0.23.46 are now available: - rawhide (<https://bodhi.fedoraproject.org/updates/FEDORA-2026-dcf642f4c3>) - f45-build-side-155802 (<https://bodhi.fedoraproject.org/updates/FEDORA-2026-4685ebea4f>) - f44-build-side-155804 (<https://bodhi.fedoraproject.org/updates/FEDORA-2026-fa08e0d381>) - f43-build-side-155806 (<https://bodhi.fedoraproject.org/updates/FEDORA-2026-c2e6d24cb0>) - epel10.4-build-side-155808 (<https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2026-2c0cdb5ae9>) - epel10.3-build-side-155810 (<https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2026-4a72b956bb>) - epel10.2-build-side-155816 (<https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2026-54f44a79bb>) - epel9-build-side-155818 (<https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2026-baf169e0a7>) Dependent packages can either be rebuilt into these side-tags and attached to the existing updates, or built directly for the corresponding release(s) after these updates are stable.
Member

I got impatient to ship a set of trustfall-rustdoc-adapter updates, so I’m taking care of rust-cargo-semver-checks after all. The needed faster-hex and rustls are stable everywhere except F45, where I’ll re-use and update https://bodhi.fedoraproject.org/updates/FEDORA-2026-ca950aef8f, tagging the needed dependencies into the side tag for long enough to do the rebuild.

I’m also taking care of rust-sccache (with the same approach in F45) as I patch it to accommodate a new typed-path release.

I got impatient to ship a set of `trustfall-rustdoc-adapter` updates, so I’m taking care of `rust-cargo-semver-checks` after all. The needed `faster-hex` and `rustls` are stable everywhere except F45, where I’ll re-use and update https://bodhi.fedoraproject.org/updates/FEDORA-2026-ca950aef8f, tagging the needed dependencies into the side tag for long enough to do the rebuild. I’m also taking care of `rust-sccache` (with the same approach in F45) as I patch it to accommodate a new `typed-path` release.
Member

@decathorpe wrote in #43 (comment):

  • kryptico

This seems to… not exist?

@decathorpe wrote in https://forge.fedoraproject.org/rust/backlog/issues/43#issue-340762: > * [ ] kryptico This seems to… not exist?
Member

What do you think is a reasonable strategy in situations like this for the applications that don’t have rust-sig as a co-maintainer? Repeatedly going through the ceremony of trying to get consent to rebuild these as provenpackager is not something I’m interested in. And of course, Pagure doesn’t have good support for PR’s that only contain empty commits, so opening PR’s to rebuild isn’t a good approach. I guess just filing bugs and hoping the maintainers do the rebuilds would be the way to go, but this is pretty tedious without automation, too.

What do you think is a reasonable strategy in situations like this for the applications that don’t have `rust-sig` as a co-maintainer? Repeatedly going through the ceremony of trying to get consent to rebuild these as provenpackager is not something I’m interested in. And of course, Pagure doesn’t have good support for PR’s that only contain empty commits, so opening PR’s to rebuild isn’t a good approach. I guess just filing bugs and hoping the maintainers do the rebuilds would be the way to go, but this is pretty tedious without automation, too.
Member

@music wrote in #43 (comment):

What do you think is a reasonable strategy in situations like this for the applications that don’t have rust-sig as a co-maintainer?

(In the above category are aw-server-rust and fips-provider from the packages that need to be rebuilt, plus awatcher from those that need to be ported. I’m currently rebuilding selenium-manager since it has rust-sig as co-maintainer.)

@music wrote in https://forge.fedoraproject.org/rust/backlog/issues/43#issuecomment-1368687: > What do you think is a reasonable strategy in situations like this for the applications that don’t have `rust-sig` as a co-maintainer? (In the above category are `aw-server-rust` and `fips-provider` from the packages that need to be rebuilt, plus `awatcher` from those that need to be ported. I’m currently rebuilding `selenium-manager` since it has `rust-sig` as co-maintainer.)
Author
Owner

@music wrote in #43 (comment):

@decathorpe wrote in #43 (comment):

  • kryptico

This seems to… not exist?

Indeed. But I consistently fail to spell kryptoci ... kcryptoc ... kryotpcoc ... kryoptic correctly. It's like a tongue twister for the fingers that one (or maybe that's just me).

@music wrote in https://forge.fedoraproject.org/rust/backlog/issues/43#issuecomment-1368683: > @decathorpe wrote in #43 (comment): > > > * [ ] kryptico > > This seems to… not exist? Indeed. But I consistently fail to spell kryptoci ... kcryptoc ... kryotpcoc ... kryoptic correctly. It's like a tongue twister for the fingers that one (or maybe that's just me).
Author
Owner

@music wrote in #43 (comment):

What do you think is a reasonable strategy in situations like this for the applications that don’t have rust-sig as a co-maintainer? Repeatedly going through the ceremony of trying to get consent to rebuild these as provenpackager is not something I’m interested in. And of course, Pagure doesn’t have good support for PR’s that only contain empty commits, so opening PR’s to rebuild isn’t a good approach. I guess just filing bugs and hoping the maintainers do the rebuilds would be the way to go, but this is pretty tedious without automation, too.

I see two (well, three) options here:

  1. Ask maintainers to add @rust-sig as co-maintainer on those packages.
  2. You can "hack" pagure to allow empty PRs. Open the browser inspector on the disabled "Merge" button and remove the "disabled" attribute, and it will become a working button (not really a serious option, but ... the more you know).
  3. I'm working on a script to file automated bugs for packages that need to be rebuilt and / or have their vendored dependencies updated.

I think the end result will be some combination of options 1 and 3.

@music wrote in https://forge.fedoraproject.org/rust/backlog/issues/43#issuecomment-1368687: > What do you think is a reasonable strategy in situations like this for the applications that don’t have `rust-sig` as a co-maintainer? Repeatedly going through the ceremony of trying to get consent to rebuild these as provenpackager is not something I’m interested in. And of course, Pagure doesn’t have good support for PR’s that only contain empty commits, so opening PR’s to rebuild isn’t a good approach. I guess just filing bugs and hoping the maintainers do the rebuilds would be the way to go, but this is pretty tedious without automation, too. I see two (well, three) options here: 1. Ask maintainers to add @rust-sig as co-maintainer on those packages. 2. You can "hack" pagure to allow empty PRs. Open the browser inspector on the disabled "Merge" button and remove the "disabled" attribute, and it will become a working button (not really a serious option, but ... the more you know). 3. I'm working on a script to file automated bugs for packages that need to be rebuilt and / or have their vendored dependencies updated. I think the end result will be some combination of options 1 and 3.
Member

@decathorpe wrote in #43 (comment):

Indeed. But I consistently fail to spell kryptoci ... kcryptoc ... kryotpcoc ... kryoptic correctly. It's like a tongue twister for the fingers that one (or maybe that's just me).

Aha! This one has rust-sig, so I’ll go ahead and rebuild it.

@decathorpe wrote in https://forge.fedoraproject.org/rust/backlog/issues/43#issuecomment-1368772: > Indeed. But I consistently fail to spell kryptoci ... kcryptoc ... kryotpcoc ... kryoptic correctly. It's like a tongue twister for the fingers that one (or maybe that's just me). Aha! This one has `rust-sig`, so I’ll go ahead and rebuild it.
Member

@music wrote in #43 (comment):

Aha! This one has rust-sig, so I’ll go ahead and rebuild it.

Actually, the primary maintainer is shipping an update right now, so I’ll wait a bit to see which branches get left at an old version and therefore still need rebuilding.

@music wrote in https://forge.fedoraproject.org/rust/backlog/issues/43#issuecomment-1368797: > Aha! This one has `rust-sig`, so I’ll go ahead and rebuild it. Actually, the primary maintainer is shipping an update *right now*, so I’ll wait a bit to see which branches get left at an old version and therefore still need rebuilding.
Sign in to join this conversation.
No description provided.