- From: Roel Nieskens <roel@pixelambacht.nl>
- Date: Mon, 22 Aug 2016 20:29:43 +0200
- To: Cameron McCormack <cam@mcc.id.au>
- Cc: Sairus Patel <sppatel@adobe.com>, Chris Lilley <chris@w3.org>, public-svgopentype@w3.org, Jonathan Kew <jfkthame@gmail.com>
Hi folks, The tool used to create LapisLegit and EmojiOne has been fixed: https://github.com/eosrei/scfbuild/issues/7 LapisLegit shows correctly in Edge now: https://pixelambacht.nl/lapislegit/ -Roel On Wed, Aug 17, 2016 at 3:54 AM, Cameron McCormack <cam@mcc.id.au> wrote: > Hi Sairus, > > On Wed, Aug 17, 2016, at 05:43 AM, Sairus Patel wrote: >> I’d also heard of a Firefox bug in which unitsPerEm is assumed to be >> 1000. >> >> Could Jonathan, Cameron, or someone else at Mozilla please confirm that >> this is still a known limitation of Firefox’s OT-SVG impl? > > Firefox should honour the unitsPerEm in the head table to determine how > to scale glyphs when rendering: > > http://searchfox.org/mozilla-central/rev/03b3c20a6ec60435feb995a2a23c5926188c85a1/gfx/thebes/gfxFont.cpp#2186-2193 > > But I notice that we do use 1000x1000 as the size of the "window" when > setting up the SVG document: > > http://searchfox.org/mozilla-central/rev/03b3c20a6ec60435feb995a2a23c5926188c85a1/gfx/thebes/gfxSVGGlyphs.cpp#134-153 > > This can affect how percentage values in the SVG content are resolved if > the root <svg> element does not specify a width="" and height="". For > example, assume you have a font with unitsPerEm = 2048: > > <svg xmlns="http://www.w3.org/2000/svg"> > <g id="glyph0"> > <rect width="100%" height="128"/> > </g> > </svg> > > will define glyph 0 with a rectangle that is 1000 units wide, regardless > of what the unitsPerEm value is, but the height of the rectangle will > still take up 2048 / 128 = 16th of the height of the em square. > > (This is just from code inspection; I didn't test.) > > This is something that should be clarified in the specification. Do you > know what other implementations do for percentage values like this? Are > they resolved against a window size derived from the unitsPerEm? > > Thanks, > > Cameron >
Received on Monday, 22 August 2016 18:30:43 UTC