Re: [css-inline] Baseline Alignment Ahem Variant Needed

Hi Myles,

On Fri, Sep 23, 2016 at 6:57 PM, Myles C. Maxfield <mmaxfield@apple.com>
wrote:

> +John & Rossen & Dominik for the question at the end. (Sorry if I added
> the wrong person, but I'm hoping you could find the answer out 😊)
>
> In the font I attached to the previous email, I modified the control
> points of all the glyphs. I subtracted 200 from all the glyphs'
> Y-coordinates, as well as from the font's ascent and the font's descent.
> Then I fixed up the paths for the É and p glyphs so their edges are
> correctly still at Y=0. (And then deleted the C glyph because the Ç glyph
> is already in the font and has the same shape.)
>
> Regarding the other two fonts I made: there are many tables which can be
> in a font, and most browsers ignore most of those tables (which isn't a
> bug; it just means they don't get extra fancy stuff). It appears that none
> of the browsers use the "bsln" table nor the "BASE" table (WebKit doesn't
> even claim to use them, but I can't speak for the other browser vendors).
> Testing browsers using a mechanism which isn't claimed to be supported
> anywhere would be a mistake.
>

If this is the question you meant to have answered: blsn seems to AAT only
so we dont't support that. I briefly checked info on BASE
<https://www.microsoft.com/typography/otspec/base.htm> but I don't think we
currently incorporate information from the BASE table into line layout in
Chrome. Would probably be interesting to look into it more, though.

Dominik




> I would be thrilled to know if other browsers claim to use these baseline
> tables. If they do, we can test with these two fonts with the additional
> baseline tables. Otherwise, testing with them would be meaningless.
>
> Thanks,
> Myles
>
> On Sep 23, 2016, at 12:16 PM, fantasai <fantasai.lists@inkedblade.net>
> wrote:
>
> On 09/22/2016 07:37 PM, Myles C. Maxfield wrote:
>
> Here it is. It seems to work on Chrome, Firefox, and Safari (and I haven’t
> tried on Edge).
>
>
> This font hardcodes the baseline shift by simply adjusting the control
>
> points of all the glyph contours. Alternatively, I also made two extra
>
> fonts, one which uses the “bsln” table, and one which uses the “BASE”
>
> table, to move the baseline without modifying the glyph contours.
>
> However, I’ve found that no browsers seem to honor these tables (but
>
> I know the tables are correct because a native app directly using
>
> CoreText reacts to the tables appropriately). If you want these two
>
> additional fonts, feel free to contact me, but I figured I wouldn’t
>
> include them here since they are likely not what you are looking for.
>
>
> Okay, I don't think I actually understood what you're saying here. :)
> In the first case, did you just shift the glyph above the y=0 line?
> Or did you move the y=0 line with respect to the ascent/descent?
> Or something else?
>
> (If those tables are supposed to work, then, yes, please give me a copy
> of those fonts as well; all three should have different font names though.)
>
> ~fantasai
>
>