W3C

- DRAFT -

WebFonts Working Group Teleconference

22 Apr 2015

See also: IRC log

Attendees

Present
Vlad, jfkthame, ChrisL, +1.250.668.aaaa, j_hudson, RSheeter, kuettel, KhaledHosny
Regrets
Chair
Vlad
Scribe
ChrisL

Contents


<trackbot> Date: 22 April 2015

<RSheeter> if I scribe I fear you will hear my typing given how I setup hangouts

<scribe> scribenick: ChrisL

Vlad: we have a couple of comments following the publication
... one was typo, now fixed
... second was impl and how it handles a condition, not a spec issue
... resolved as no change
... who updates the Google implementation?

RSheeter: me, then I delegate depending

conformance test suite

Vlad: reviewed test plan, added all the new conformance items that were related to font collections and also the glyf and loca stuff

https://www.w3.org/Fonts/WG/wiki/TestPlan20-Format

https://www.w3.org/Fonts/WG/wiki/TestPlan20-AuthoringTool

https://www.w3.org/Fonts/WG/wiki/TestPlan20-UserAgent

ChrisL: prefer we adopt the ednote suggestion for http://dev.w3.org/webfonts/WOFF2/spec/#conform-mustSpecifyGlyfTableSize
... it should say "this is the size in the original font" and make not promises abot how big the reconstruction ends up

Vlad: and then no test needed

ChrisL: it would become an authoring tool test

Vlad: straw poll - remove test on size of reconstructed glyph table? not practical to test

RSheeter: remove it

<scribe> ACTION: vlad to remove the #conform-mustSpecifyGlyfTableSize requirement and ed note [recorded in http://www.w3.org/2015/04/22-webfonts-minutes.html#action01]

<trackbot> Created ACTION-173 - Remove the #conform-mustspecifyglyftablesize requirement and ed note [on Vladimir Levantovsky - due 2015-04-29].

kuettel: agree

<RSheeter> I think that was me and Khaled not Kuettel

ChrisL: easy as an authoring tool test. MUST be same

Vlad: but then why not for all the other tables

ok never mind then, no conformance on that

Vlad: wanted to prepare this for f2f meeting, go through the plans and see what we can do with the new ones

I changed a few things on https://www.w3.org/Fonts/WG/wiki/TestPlan20-UserAgent today

Vlad: extraneous-reject was similar in WOFF1, we had well defined padding requirements for tables, no overlap etc
... much more complex in woff2
... all variable length portions, cant see where they end without parsing. so needs a series of tests
... flagged as needing more work in the test plan

(no suggestions)

"For each place where extraneous data could occur, make a test."

Vlad: that is a lot of places. hard to find where those places are
... would need to be doctored by hand

RSheeter: make explicit tests for major and obvious ones?

Vlad: easy between main and meta, meta and private

j_hudson: is it that we have no clear prioritization of what to test?

Vlad: ambiguously defined. and also, variable length parts with no separation
... so hard to find where to insert space

ChrisL: need to reassess value of preventing ppl hiding data in odd places like between tables

Vlad: yes with per-table compression there were clear interstices that could be exploited. here it is all one compressed blob

ChrisL: suggest just testing in a few places

KhaledHosny: we already test those places

Vlad: in reality these three sections are contiguous. need to parse header to see where the joins between sections are
... and its harder to insert extra data, compared to woff1
... wonder if the test is too complicated for little gain

j_hudson: ask the w3c security folks?
... if there are security concerns they would know and we could flag it

<scribe> ACTION: ChrisL ask webappsec to review woff2 [recorded in http://www.w3.org/2015/04/22-webfonts-minutes.html#action02]

<trackbot> Created ACTION-174 - Ask webappsec to review woff2 [on Chris Lilley - due 2015-04-29].

j_hudson: maybe the person who commented on woff1 in last call would comment here

Vlad: even with some number of tests we can't test all possible insertion points. want to avoid a false sense of security too

ChrisL: fuzzing makes more sense on a per-implementation basis

RSheeter: restrict it to the ones we already test like in woff1

Vlad: yes

(we will limit the test plan to the basic file structure insertion points)

(agreement)

Vlad: we now have specific descriptions on bounding box calculations, loca tables, etc
... esp for font collection support
... description best done by whoever makes the tests
... KhaledHosny any ideas on those?

KhaledHosny: I can edit the descriptions for these

https://www.w3.org/Fonts/WG/wiki/TestPlan20-UserAgent#mustCalculateBBox

(and on from there)

KhaledHosny: I have started looking at these already

Vlad: authoring tool also changed because of the collections stuff

https://www.w3.org/Fonts/WG/wiki/TestPlan20-AuthoringTool

Vlad: often related to UA tests esp glyf and loca
... KhaledHosny if you could share your notes we can verify against test plan
... then ready for f2f meeting, when we are all in smae room things move fast

next meeting

Vlad: so who is coming to f2f meeting? chris responded
... suggested dates are june 9-11 but we probably don't need 3 days

j_hudson: could call in, will not be there

Vlad: probably hosted at monotype office, if not somewhere in boston

kuettel: we wanted to be sure khaled could join somehow, travel is problematic so remote access is important
... today was first succesful trial, joining via hangouts and today was the first time that worked
... otherwise we were reluctant to have the meeting

KhaledHosny: all the suggested days are okay
... can manage for attending most of each day

Vlad: certainly June 9, maybe spill over to June 10, don't see more than two days

<RSheeter> Rod & David K can make it :D

Vlad: jfkthame could you make it?

jfkthame: not in person
... maybe connect for part of the meeting, if needed.

Vlad: we can plan for an early start to ease remote participation

Resolved: Meeting 9 and 10 June, Boston, 3 remote participants and 5 on-site

kuettel: google has an office in cambridge if needed. hangouts already setup. lets discuss offline

Vlad: monotype has good connectivity in conference rooms too
... and hotel next door
... (marriott courtyard)

<RSheeter> https://github.com/w3c/woff2-tests

RSheeter: also look at the test repo on github

aob

Summary of Action Items

[NEW] ACTION: ChrisL ask webappsec to review woff2 [recorded in http://www.w3.org/2015/04/22-webfonts-minutes.html#action02]
[NEW] ACTION: vlad to remove the #conform-mustSpecifyGlyfTableSize requirement and ed note [recorded in http://www.w3.org/2015/04/22-webfonts-minutes.html#action01]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015/04/22 20:58:10 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.140  of Date: 2014-11-06 18:16:30  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Found ScribeNick: ChrisL
Inferring Scribes: ChrisL
Default Present: Vlad, jfkthame, ChrisL, +1.250.668.aaaa, j_hudson
Present: Vlad jfkthame ChrisL +1.250.668.aaaa j_hudson RSheeter kuettel KhaledHosny
Found Date: 22 Apr 2015
Guessing minutes URL: http://www.w3.org/2015/04/22-webfonts-minutes.html
People with action items: ask chrisl vlad webappsec

[End of scribe.perl diagnostic output]