freetype-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: ftview segfault on ArefRuqaaInk-Regular.ttf (one of google web fonts


From: Cosimo Lupo
Subject: Re: ftview segfault on ArefRuqaaInk-Regular.ttf (one of google web fonts)
Date: Wed, 24 May 2023 18:19:39 +0100

Apparently it's not the hosted font that is buggy (in fact the OT-SVG webfont works in Safari when loading https://fonts.google.com/specimen/Aref+Ruqaa+Ink; it's only the fonts included in the zip download. We're investigating the reason of that and will fix ASAP

On Wed, May 24, 2023 at 6:16 PM Hin-Tak Leung <htl10@users.sourceforge.net> wrote:
Argh thanks ! Glad Werner noticed it last week already... actually I had a draft of a blunt email  last night to Hosny and Dave about it being a font bug, but didn't send, for another thought - thinking that Google fonts cannot be hosting buggy fonts. I was wrong :-).

On Wednesday, 24 May 2023 at 21:16:54 GMT+8, Cosimo Lupo <cosimo@anthrotype.com> wrote:


if you downloaded that font from https://fonts.google.com/specimen/Aref+Ruqaa+Ink then it's a font bug, or rather the font has been incorrectly subsetted (the rest of the font lost a couple unreferenced glyphs but SVG table stayed the same because hb-subset can't subset it; it should have been subsetted -- if at all -- using the --retain-gids flag). See https://github.com/google/fonts/issues/6295

The original font file at https://github.com/google/fonts/tree/main/ofl/arefruqaaink has the correct glyph order and SVG table.

On Wed, May 24, 2023 at 1:39 PM Hin-Tak Leung <htl10@users.sourceforge.net> wrote:
On Wednesday, 24 May 2023, 05:49:00 BST, Werner LEMBERG <wl@gnu.org> wrote:

> > I finished adding the color toggle to ftgrid (trivial, just cut and
> > paste a few lines from ftview)

> Please submit a Merge Request!

That's okay when I finish the current investigation. I think I did have a gitlab account, then somehow it got suspended about half year ago from lack of use. I tried re-activating without success... btw, the mailing list also seems to change a few months ago - I used to get digests, now it comes as individual posts.

I think I found an anomaly with ftgrid/ftview with ArefRuqaaInk-Regular.ttf - using hb-shape/hp-view, I get at gid 468/469 which are a pair of blue and red braces ; hb-view can draw them as colours, so it is getting svg at 468/469. But on ftview/ftgrid, the blue and red braces are glyph 475 and 476 (below verbatim from FT2_DEBUG), and you get the black/white braces in my hacked ftgrid at 468/469:

========
<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" version="1.1"><defs><radialGradient id="g1" cx="277" cy="-196" r="475" gradientUnits="userSpaceOnUse"><stop offset="0" stop-color="#D22E1F" stop-opacity="0.252"/><stop offset="1" stop-color="#D22E1F"/></radialGradient><radialGradient id="g2" cx="255" cy="-195" r="260" gradientUnits="userSpaceOnUse"><stop offset="0" stop-color="#0074B6"/><stop offset="1" stop-color="#0074B6" stop-opacity="0.252"/></radialGradient><path d="M256,-199 Q284,-207 298,-222.5 Q312,-238 317,-255 Q322,-272 322,-284 Q322,-318 300,-338.5 Q278,-359 244,-359 Q221,-359 198,-345 Q175,-331 167,-331 Q154,-331 149,-336 L154,-344 Q156,-343 160,-343 Q174,-343 182,-357.5 Q190,-372 197,-380 Q224,-418 254,-418 Q288,-418 302,-383 L304,-385 Q294,-406 294,-431 Q294,-470 345,-521 Q350,-526 357.5,-533 Q365,-540 374,-548 L401,-574 Q419,-590 425,-603 Q407,-608 407,-622 Q407,-640 427,-640 Q446,-640 446,-621 Q446,-609 429,-583 Q416,-566 407.5,-552 Q399,-538 394,-528 Q348,-442 348,-327 Q348,-211 277,-196 Q348,-180 348,-65 Q348,50 394,136 Q399,146 407.5,160 Q416,174 429,191 Q446,217 446,229 Q446,248 427,248 Q407,248 407,230 Q407,215 425,211 Q419,198 401,182 L374,156 Q365,148 357.5,141 Q350,134 345,129 Q294,78 294,39 Q294,14 304,-7 L302,-9 Q288,26 254,26 Q224,26 197,-12 Q190,-20 182,-34.5 Q174,-49 160,-49 Q156,-49 154,-48 L149,-56 Q154,-61 167,-61 Q175,-61 198,-47 Q221,-33 244,-33 Q278,-33 300,-53.5 Q322,-74 322,-108 Q322,-120 317,-136.5 Q312,-153 298,-168 Q284,-183 256,-191 Z" id="uniFD3E.0" fill="url(#g1)"/><path d="M217,-51 Q172,-93 172,-144 Q162,-132 141,-132 Q115,-132 96,-152 L77,-173 Q68,-184 59,-184 Q48,-184 43,-181 L39,-191 Q52,-198 68,-215 Q84,-234 100.5,-248 Q117,-262 141,-262 Q158,-262 172,-248 Q172,-298 217,-340 L225,-336 Q220,-321 220,-311 Q220,-277 260,-266 Q292,-257 292,-233 Q292,-219 279.5,-210.5 Q267,-202 256,-195 Q270,-187 281,-180 Q292,-173 292,-159 Q292,-134 260,-125 Q220,-114 220,-81 Q220,-70 225,-55 Z M184,-165 Q193,-183 215,-195 L215,-197 Q193,-209 184,-227 Q148,-220 148,-196 Q148,-173 184,-165 Z" id="uniFD3E.1" fill="url(#g2)"/></defs><g id="glyph475"><use xlink:href="" xlink:href="" id="glyph476"><use xlink:href="" x="485" y="-392" transform="matrix(-1 0 0 -1 970 -784)"/><use xlink:href="" x="485" y="-392" transform="matrix(-1 0 0 -1 970 -784)"/></g></svg>
===

So I am not sure what's going on - ftgrid /ftview seems to get the wrong glyphs from 119 onwards, but the svg fragments themselves identify them as the wrong numbers. (I cut and paste the above into inkscape and got a pair of overlapping braces. see the "<g id="glyph475">" within.

Hin-Tak

reply via email to

[Prev in Thread] Current Thread [Next in Thread]