bstr VS mold

Libraries for a lot of this stuff exist (albeit in many cases not very mature yet):

- https://github.com/pop-os/cosmic-text does text layout (which Taffy explicitly considers out of scope)

- https://github.com/AccessKit/accesskit does accessibility

- https://github.com/servo/rust-cssparser does value-agnostic CSS parsing (it will parse the general syntax but leaves value parsing up to the user, meaning you can easily add support for whatever properties you what). Libraries like https://github.com/parcel-bundler/lightningcss implement parsing for the standard css properties.

- There are crates like https://github.com/BurntSushi/bstr and https://docs.rs/wtf8/latest/wtf8/ for working with non-unicode text

We are planning to add a C API to Taffy, but tbh I feel like C is not very good for this kind of modularised approach. You really want to be able to expose complex APIs with enforced type safety and this isn't possible with C.

As the author of bstr and also the regex implementation that bstr uses to implement word breaking, it is linear time.

NSFL: https://github.com/BurntSushi/bstr/blob/86947727666d7b21c97e...

OsStr uses WTF-8 on Windows, and just represents the raw underlying bytes on Unix.

Byte strings can be WTF-8. They can be anything. The problem is that there is no real way to (easily) get the underlying WTF-8 bytes of an OsStr on Windows. So there's no free conversion to and from byte strings.

I wrote more about this in the bstr docs (and don't miss the link to os_str_bytes): https://docs.rs/bstr/latest/bstr/#file-paths-and-os-strings

I'd be happy to answer more questions if you have them. :-) https://github.com/BurntSushi/bstr/discussions

Check out bstr, which does this exact thing for its BString and BStr types.

Currently slated for next Monday: https://github.com/BurntSushi/bstr/issues/40

This is just an FYI. I don't mean to say much to your overall point, although, as someone else who has spent a lot of time doing Unicode-y things, I do tend to agree with you. I had a very similar discussion a bit ago.[1]

Putting that aside, at least with respect to grapheme segmentation, it might be a little simpler than you think. But maybe only a little. The unicode-segmentation crate also does word segmentation, which is quite a bit more complicated than grapheme segmentation. For example, you can write a regex to parse graphemes without too much fuss[2]. (Compare that with the word segmentation regex, much to my chagrin.[3]) Once you build the regex, actually using it is basically as simple as running the regex.[4]

Sadly, not all regex engines will be able to parse that regex due to its use of somewhat obscure Unicode properties. But the Rust regex crate can. :-)

And of course, this somewhat shifts code size to heap size. So there's that too. But bottom line is, if you have a nice regex engine available to you, you can whip up a grapheme segmenter pretty quickly. And some regex engines even have grapheme segmentation built in via \X.

[1]: https://github.com/BurntSushi/aho-corasick/issues/72

[2]: https://github.com/BurntSushi/bstr/blob/e38e7a7ca986f9499b30...

[3]: https://github.com/BurntSushi/bstr/blob/e38e7a7ca986f9499b30...

[4]: https://github.com/BurntSushi/bstr/blob/e38e7a7ca986f9499b30...

This is a great idea. I realize the find implementation is not ideal and have considered bringing in an optional dependency to improve performance. I remembered bstr using two-way search, so I was wondering if depending on the full crate for searching would be worthwhile, but I see that changed. Thanks for the tip!

Fun little nit-pick that does not detract from your overall point: you can actually count graphemes with a regex and that's exactly what bstr does. :-)

I think this is unlikely to gain traction. I say that no to discourage you, just to explain.

- The community has an instinctive distrust of closed source or a compiler from an untrusted source. If you’re familiar with the Trusting Trust attack you’ll understand why.

- Dev tools in every language ecosystem are almost always free, unless they involve some kind of hosting. People aren’t used to opening their wallets. Look the experience of the guy who built the mold linker(https://github.com/rui314/mold). Far superior to the state of art, improves incremental compiles a lot, widely applicable across ecosystems (C, C++, Rust), CPU architectures and Operating Systems. You don’t even have to modify your compiler, just need to point to his linker. He’s even giving it away for free for personal use. But still, almost no one uses it. The inertia of the established options is really high.

- It’s not complex enough. Think about the complexity involved in the cranelift backend. No one can seriously recreate the efforts of bjorn3. If we could have, we would have. But the idea idea here can be recreated, especially by the experts who already built incremental compilation into rustc.

- But if your solution is truly complex, like the parallel frontend, the burden of maintaining a fork would be too high. You’d have to spend all your time rebasing.

Again I’m not trying to discourage you, just stating the difficulties of making a business in the dev tools space. You would be better off contributing this excellent work to the community and trying a different tack.

I initially thought this would be about the mold linker (https://github.com/rui314/mold)

I assume this submission is trying to highlight the specific message (2023-01-24) : https://github.com/rui314/mold/issues/190#issuecomment-14028...

Fyi... the author wrote a more expansive blog post about selling dev tools a few months later (2023-06-06) and there was a related HN thread about it: https://news.ycombinator.com/item?id=36225016

Loongson's LoongArch CPU has been supported. (03b1a1c)

I'm amazed at how quickly the author responds to requests: https://github.com/rui314/mold/issues/1057

From the report to the fix in less than two days.

I'm not sure how competitive it will be with lld, especially if we consider ThinLTO (which takes multiple minutes on 64-core machine) - it can make the advantages of mold insignificant.

What other solutions have you tried for the problem of slow linking? You haven't even said which linker and what flags you're using. I haven't actually tried it, but the author of gold has an even faster linker called mold: https://github.com/rui314/mold

> Mold did it first, though: https://github.com/rui314/mold

Before LLD?