Why ttf_parser::Face need a lifetime? about ttf-parser HOT 6 CLOSED

You should use owned-ttf-parser instead.

Is font parse not one-time action?

No. ttf-parser parses the font parts on demand. Parsing the whole font into the memory is absurd and no one is doing this.

The lifetime is required, since Face doesn't allocate anything (at least on heap). At all.

from ttf-parser.

You should use owned-ttf-parser instead.

Is font parse not one-time action?

No. ttf-parser parses the font parts on demand. Parsing the whole font into the memory is absurd and no one is doing this.

The lifetime is required, since Face doesn't allocate anything (at least on heap). At all.

Hey @RazrFalcon ,
I'm currently utilizing ttf-parser and exploring the possibility of caching the Face object for repeated use. Below is a snippet of my current implementation (which didn't work out due to the lifetime constraints and the fact that I'm no experienced Rust developer yet):

#[derive(Resource, Default)]
pub struct FontCacheRes {
    font_content: HashMap<u64, CachedFont>,
}

#[derive(Default)]
pub struct CachedFont {
    pub content: Vec<u8>,
    // https://github.com/RazrFalcon/ttf-parser/issues/37
    // pub face: Option<Face>, // TODO: Cache Face?
    pub font: Font,
}

impl CachedFont {
    // pub fn get_or_create_face(&'a mut self) -> &'a Face<'a> {
    //     if self.face.is_none() {
    //         self.face = Face::from_slice(&self.content, 0);
    //     }
    //     self.face.as_ref().unwrap()
    // }
}

Given the lifetime requirements of Face, I'm debating whether caching the Face makes any sense or if there's a recommended practice for handling such scenarios. Is it advisable to cache Face within a struct, or should I reparse the font data each time it's required? Insights or best practices on this matter would be very helpful.

Thanks :)

from ttf-parser.

As the comment above mentioned - just use owned-ttf-parser.

And yes, depending on your performance requirements you might be fine with re-parsing the Face on each access.
Face::parse doesn't actually parse the whole font, only some basic metadata.

from ttf-parser.

As the comment above mentioned - just use owned-ttf-parser.

And yes, depending on your performance requirements you might be fine with re-parsing the Face on each access. Face::parse doesn't actually parse the whole font, only some basic metadata.

thanks for the quick reply 🙏 Can I use owned-ttf-parser with rustybuzz though?

from ttf-parser.

Yes. owned-ttf-parser is just a wrapper for ttf-parser. See docs.

from ttf-parser.

Yes. owned-ttf-parser is just a wrapper for ttf-parser. See docs.

Thanks for the quick information. I'll explore both caching and not caching the Face to see what works best for my SVG text rendering project. Also, a big thanks for developing rustybuzz & ttf-parser – it looks like a great fit for my needs 👍🙏

from ttf-parser.