1715595 - Use char rather than uint8_t for utf-8 in unified components

Status: RESOLVED FIXED

(Core :: Internationalization, enhancement)

Description

[Dan Minor [:dminor]]

We should replace uint8_t with char in the unified components. I had originally used uint8_t for compatibility with Rust which is using u8, but the ICU4X C FFI has been developed around char instead.

This is pending the resolution of https://github.com/unicode-org/icu4x/issues/769, in case we end up choosing something different there.

Comment 1

[Dan Minor [:dminor]]

Attachment: Bug 1715595 - Use char rather than uint8_t for utf-8 in unified components

Comment 2

[Jonathan Kew [:jfkthame]]

My gut instinct would be to prefer an unsigned type here; UTF-8 uses (virtually) the full range of byte values from 0x00 - 0xFF, and having half of them be negative on the C side can be something of a potential footgun.

But I guess we need to see where the ICU4X issue goes....

Comment 3

[Dan Minor [:dminor]]

My take on the upstream issue is that the decision is to use char and the issue remains open to update the unit tests where needed. I'll double check.

Comment 4

[Dan Minor [:dminor]]

Confirmed that upstream is going to use char.

Comment 5

[Dan Minor [:dminor]]

https://treeherder.mozilla.org/jobs?repo=try&revision=a27e20fb96662d395601ffa8e481879fd3ea34fc

Comment 6

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/5e6a1afb2e39
Use char rather than uint8_t for utf-8 in unified components r=platform-i18n-reviewers,gregtatum

Comment 7

[Marian-Vasile Laza]

https://hg.mozilla.org/mozilla-central/rev/5e6a1afb2e39