I’m the leading engineer behind
fluent-rs - Rust implementation of the Fluent Localization System.
I’m working on final touches on the low level crate that operates on the FTL file format used by Fluent.
The AST is currently using
&str over a source String which makes the AST read-only. The most critical use of the parser/AST is runtime where read-only is totally enough.
But it’s tempting to use the same AST/parser for tooling where ability to modify the AST and serialize it after is important.
I see several possible approaches to that, and I’m not sure how to weight pros and cons and if maybe I missed some other option.
Here’s the issue about it - https://github.com/projectfluent/fluent-rs/issues/104
I’d appreciate help with making the decision and any feedback from other crates that implemented parser/AST/serializer and had to make similar decisions.
We also have a couple more decisions to make prior to 1.0 release - https://github.com/projectfluent/fluent-rs/labels/fluent-syntax
Appreciate feedback and help!