- From: Geoffrey Sneddon <me@gsnedders.com>
- Date: Fri, 6 Nov 2015 12:54:11 +0000
- To: Gérard Talbot <css21testsuite@gtalbot.org>
- Cc: John Daggett <jdaggett@mozilla.com>, public-css-testsuite@w3.org
On 06/11/15 00:49, Gérard Talbot wrote: > Le 2015-11-05 19:16, Geoffrey Sneddon a écrit : >> CSSTest/csstest-familyname-funkyA.ttf has names "small-caps 1in CSSTest >> FamilyName Funky" and "ＣＳＳテスト フォント名". Unless I'm missing >> anything, this has nothing else special about it (and I could well have >> missed something!). > > In that font, glyphs "1" "2" "3" "4" have been respectively replaced > with "F" "A" "I" "L" and "5" "6" "7" "8" have been respectively > replaced with "P" "A" "S" "S" glyphs. I meant about the font's metadata, insofar as it effects font-matching. i.e., given <p style="font-family: ＣＳＳテスト フォント名">5678</p> what font is initially selected (by step one of the font matching algorithm)? As far as I can tell, this is UA dependent (per "If there are two fonts with exactly the same properties, the user agent selects one of them"), because it's equally valid to choose csstest-familyname.ttf, CSSTest/csstest-familyname-funkyA.ttf, CSSTest/csstest-familyname-funkyB.ttf and CSSTest/csstest-familyname-funkyC.ttf. This seems not great for a test font! This affects <css-fonts-3/font-family-name-010.xht> and <css-fonts-3/font-family-name-011.xht> if I'm not mistaken. /Geoffrey
Received on Friday, 6 November 2015 12:54:41 UTC