I'm having a bit of trouble wrapping my head around what the underline is expected to look like for upright vertical text.  Do you happen to have an example png lying around? :)

Yes, unfortunately that rotation is not playing nicely with text blobs and the bloberizer is hacking around it.  I recall at the time we considered adding per-run rotation info to text blobs, but gave up on it for some reason (either decided that upright vertical was rare enough not to worry about it, or there were other complications, I don't recall).

If we resurrected that idea, and instead of rotating the canvas we would add rotation info to each text blob run, do you think that would provide sufficient info for Skia to compute the correct intercepts?

> I'm having a bit of trouble wrapping my head around what the underline is expected to look like for upright vertical text.  Do you happen to have an example png lying around? :)

I don't have one, but I could construct one if needed. The Japanese glyph sequence in vertical should be rotated counterclockwise (it is not in this image, because I disable the rotation), so that it follows the aaag... sequence downwards. The intersects between the CJK glyphs and the vertical underline should then be clipped. The CJK glyphs are a little bit more "left" then the Latin glyphs, due to the baseline shift alphabetic-ideographic.

> If we resurrected that idea, and instead of rotating the canvas we would add rotation info to each text blob run, do you think that would provide sufficient info for Skia to compute the correct intercepts?

This sounds to me like it would be enough to follow a different baseline direction for calculating the intersects, but I am not sure whether we should push the rotation down in the stack in this way. Either way, we need to handle upright in vertical somehow, ideally in a clean way. Maybe Cary can take a look and comment?

I imagine it should look something like this. (photoshopped)

Deleted: vertical_skip_ink.png 44.5 KB

	vertical_skip_ink.png 44.5 KB View Download

Coincidentally WebKit has the bug http://wkb.ug/128518

This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available. If you change it back, also remove the "Hotlist-Recharge-Cold" label.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Still applies, but remains low prio.

This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

@drott - please retriage

The NextAction date has arrived: 2018-11-28