Declare 1.0 for libraries actively used by Chrome

[rsheeter]

We are LIVE from the sold out Chrome stable user base. Declare 1.0 for relevant libraries ($ ./resources/scripts/bump-version.sh font-test-data font-types read-fonts skrifa major)

BLOCKED ON:

• Moving variable font logic into read-fonts #1263 Moving variable font logic into read-fonts
• [ift] Put ift-related code behind a feature #1349 Putting IFT behind a feature gate so it doesn't cause read-fonts to race through major versions

[cmyr]

I do think this makes sense but I have some minor hesitancy about doing it super casually; at the very least I'd like to go through open issues and see if there's anything obvious that we'll want to break API for.

other current blockers:

• we know we're breaking the api in the ivs module of font_test_data, so it's worth at least doing that first
• ...nothing I can think of offhand 🤷

[rsheeter]

doing it super casually

I would argue we aren't doing it casually. We're acknowledging that we're shipping to a very large userbase and now need to be a bit more careful wrt compatibility. This makes the versioning reflect reality on the ground.

breaking the api in the ivs module of font_test_data, so it's worth at least doing that first

Agreed, imho that's the only thing we need to block on

[nicoburns]

Based on fontations' history, I'm assuming there will still be regular releases that are breaking from an API perspective (if not from a behaviour perspective)? So we should expect 2.0, 3.0 release soon and this just indicated maturity? Or is this indicating a change in policy with regard to API breaks?

[cmyr]

Based on fontations' history, I'm assuming there will still be regular releases that are breaking from an API perspective (if not from a behaviour perspective)? So we should expect 2.0, 3.0 release soon and this just indicated maturity? Or is this indicating a change in policy with regard to API breaks?

write-fonts will still stay on minor versions and should expect regular breaking changes, but we're going to try and avoid too much churn in read-fonts and skrifa; hopefully the majority of the work we need to do in these crates is just adding new API, not touching the old API.

That said, I expect we likely will need at least some breaking changes in read-fonts and skrifa in the future, in which case we would go to 2.0, 3.0 etc.

[nicoburns]

we're going to try and avoid too much churn in read-fonts and skrifa; hopefully the majority of the work we need to do in these crates is just adding new API, not touching the old API.
That said, I expect we likely will need at least some breaking changes in read-fonts and skrifa in the future, in which case we would go to 2.0, 3.0 etc.

This sounds good to me as an open source user of these crates (I'm not currently using write-fonts)! I wouldn't want y'all to hamstring yourself by forbidding breaking releases entirely. On the other hand, I had noticed that fontations crates were making breaking particularly frequently which has caused a lot of releases in downstream crates I use just to update those versions (especially font-types and read-fonts that often end up in the public API and so can't be updated in patch releases).

[cmyr]

On the other hand, I had noticed that fontations crates were making breaking particularly frequently which has caused a lot of releases in downstream crates I use just to update those versions (especially font-types and read-fonts that often end up in the public API and so can't be updated in patch releases).

Fair point, it is 100% true that we are not currently thinking very much about consumers other than chrome and fontc. That said, I do hope that our churn will start to slow down soon; in particular font-types should be almost done, and read-fonts should be getting close (although there will be a few more breaking releases before we finally get to 1.0, such as #1263)