- From: John Daggett <jdaggett@mozilla.com>
- Date: Wed, 13 Feb 2013 21:39:49 -0800 (PST)
- To: Public CSS test suite mailing list <public-css-testsuite@w3.org>
Hi Gérard, Two quick notes. The fonts in question are both far larger than is needed for testing purposes (8MB?!?) and the terms of the license doesn't appear to allow their usage. If Japanese fonts are needed, I would suggest using a *single* font from the M+fonts project, these are far smaller. I didn't look closely at Taka's test to see what codepoints they use but I'm assuming that they can be structured to use the codepoints available in these fonts: http://mplus-fonts.sourceforge.jp/mplus-outline-fonts/index-en.html One of the main problems with using downloadable fonts is that it makes it difficult to automate reftests for some browsers, specifically Webkit. The reason for this is that the onload handler fires in Webkit fires before fonts needed for content have downloaded. Once support for font load events is implemented, this will be easier to do consistently across browsers. I would suggest being cautious about using subsetting tools if you don't know much about their quality or where their deficiencies lie. There are many tools that will create all sorts of funky problems (bad cmaps, incorrect or inconsistent tables, etc.). WEFT is a Microsoft tool for creating subsetted EOT fonts, not really appropriate in this case. Regards, John
Received on Thursday, 14 February 2013 05:40:16 UTC