Please release a new version of fluent-bundle

[sylvestre]

The last update has been a year ago and it depends on some old libs like self_cell (which is now 4 years ago)
https://crates.io/crates/fluent-bundle/0.15.3/dependencies
https://github.com/Voultapher/self_cell/releases

Same story for rust-hash

It makes cargo-deny sad for some packages:

error[duplicate]: found 2 duplicate entries for crate 'self_cell'
    ┌─ /github/workspace/Cargo.lock:225:1
    │  
225 │ ╭ self_cell 0.10.3 registry+https://github.com/rust-lang/crates.io-index
226 │ │ self_cell 1.2.0 registry+https://github.com/rust-lang/crates.io-index
    │ ╰─────────────────────────────────────────────────────────────────────┘ lock entries

error[duplicate]: found 2 duplicate entries for crate 'rustc-hash'
    ┌─ /github/workspace/Cargo.lock:217:1
    │  
217 │ ╭ rustc-hash 1.1.0 registry+https://github.com/rust-lang/crates.io-index
218 │ │ rustc-hash 2.1.1 registry+https://github.com/rust-lang/crates.io-index
    │ ╰──────────────────────────────────────────────────────────────────────┘ lock entries

Thanks!

[zbraniecki]

sgtm. Any of the active maintainers want to do it? If not, I can take it next week. Just ping me

[alerque]

I can probably do it.

I recall their being a problem with bumping rustc-hash though. Working from memory there was going to be some breaking change involved in that.

[sylvestre]

@alerque seems that it was done already : d7f0d48

[alerque]

Unfortunately this isn't going to be quite as easy a bump as I expected. I forgot we added some features (e.g. implemented NUMBER()) and hence the next bump will technically be a breaking release (0.16.0 instead of 0.15.4). I don't think it will be a huge deal, but I'm working through the changes to all the workspace crates making sure they get appropriate bumps and hopefully avoid any regressions.

[alerque]

Sigh.

Even with everything freshened up internally, we're still triggering duplicate versions of rustc-hash because our own intl-memoizer depends on type-map which hasn't had a release in 4 years and uses the 1.x hash interface.

[alerque]

The type-map maintainer responded super fast and accepted my PR to update it, so that hurdle is past! (Thanks @kardeiz.)

[zbraniecki]

thank you @alerque !