[Cuis-dev] [Ann] Unicode Text Editors

[Luciano Notarfrancesco]

On Fri, 13 May 2022 at 2:33 AM Juan Vuletich <JuanVuletich at zoho.com> wrote:

> On 5/11/2022 3:41 PM, Luciano Notarfrancesco via Cuis-dev wrote:
>
> > ...
> > If you mean “together” in the Browser, maybe we can move Utf8String,
> > Utf32String and UnicodeCodePoint to the category ‘Kernel-Text’
> > together with String and Character.
> >
>
> Someday I'd like to bootstrap a small kernel image. It won't contain
> Unicode at all. That's why I didn't put them in Kernel.


Awesome. As I mentioned in other thread, I wish we started getting some
stuff out of the base image and into packages, and maybe make a “standard
distribution image” that is like the current base image but built from
packages. I think this approach has a lot of benefits, and will make
bootstrapping easier.

> We are not storing redundant
> > contour data for all the code points that don’t have a glyph like
> > 16r1D53A, right? I took a quick look at the instance of
> > TTFontDescription and there seems to be data for this code point, but
> > probably I made a mistake because I’m not familiar with this…
>
> Oh! Good catch! You are right! That didn't matter when I originally
> wrote that code, for 8 bit Character and String. Now it is a huge waste.
> Earlier today I fixed it and pushed the updates. Memory usage should
> drop by about a half.
>

Cool!
I was also pondering if we should have a Float16Array. It could reduce
memory usage by another half…
https://en.wikipedia.org/wiki/Half-precision_floating-point_format
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cuis.st/mailman/archives/cuis-dev/attachments/20220513/5f4f0981/attachment.htm>