Re: test font with both SVG and COLR table

2022-01-26 Thread Adam Twardoch (Lists) 
Apologies for a bit off-topic:

BTW, given that some folks at Apple don’t seem to be great fans of COLRv1,
while Google isn’t a great fan of SVG, chances are that we’ll see the
phasing out of "sbix" and "CBDT", but both COLRv1 and SVG will stick around
— also because SVG allows bitmaps. Check out:

- https://creativemarket.com/search?q=opentype+svg
- https://creativemarket.com/romanjusdado/2950202-Wooden-Tiles-Font
-
https://creativemarket.com/SamParrett/5150302-Glory-Culture-SVG-Font-Extras
-
https://creativemarket.com/helloimgreg/2330313-Stranger-Times-OpenType-SVG-Font
of su
This stuff is all bitmap-based, it’s all released fonts available on the
market, and for now, they’re OT+SVG only. COLRv1 won’t replace them.

This means that we’ll keep seeing OT+SVG fonts, OT+COLRv1 fonts and quite
possibly _hybrid_ OT+SVG+COLRv1 fonts, which kind of make sense if the font
vendor wants to make the users’ life simple.

I’m just saying — the hybrid OT+SVG+COLRv1 font is not an eccentricity, I
think it’ll be reality — especially if font editor vendors like FontLab
make creation of such hybrids easy.

With 4 separate color font flavors, the market acceptance was low (but
still, people have made many such fonts). But as soon as those flavors are
down to just 2, I think both will stick around. (Though I do see a reason
for sbix to also exist, especially in Apple’s own HEIC flavor, not PNG,
which they use on iOS).

Best,
Adam


On Wed, Jan 26, 2022 at 5:07 PM Cosimo Lupo via FreeType development <
freetype-devel@nongnu.org> wrote:

> Hi Werner and all,
> please find attached a test COLRv1 + SVG font, containing only one color
> glyph ✍ "WRITING HAND" (U+270D) emoji.
> The font was built using nanoemoji
>  using the following command,
> from the root of the noto-emoji
>  repository:
>
> $ nanoemoji --color_format=glyf_colr_1_and_picosvg --keep_glyph_names
> --pretty_print --family "Noto Color Emoji COLRv1 And SVG"
> svg/emoji_u270d.svg
>
> Let me know if that is what you are looking for.
> Cheers
>
> Cosimo
>
> On Tue, Jan 25, 2022 at 7:51 PM Werner LEMBERG  wrote:
>
>>
>> > Do you still need such a test font with SVG and COLR in it? I guess,
>> > we can make one if it's needed.
>>
>> This would be still very helpful, yes.  I think it would be helpful
>> for for fuzzing, too.  A single glyph (besides '.notdef) would be
>> enough.
>>
>> >> The question doesn't arise for serious `COLR` handling, as
>> >> described above.  In case of the convenience `COLR` rendering,
>> >> `SVG` takes precendence.
>> >
>> > I think the table preference decision should be made by the
>> > application, or in the FT_LoadGlyph then with flags that allow
>> > separate selection?
>>
>> Since the convenience stuff is tagged as experimental, and Alexei has
>> some serious concerns I probably won't change anything right now.
>>
>> > FWIW, overall, I think if a font has COLRv1 and SVG, COLRv1 should
>> > have preference as it enables variable capabilities.
>>
>> This is up to the application; there is no issue w.r.t. precendence at
>> all.
>>
>>
>> Werner
>>
>

Re: Odp: Re: The TrueType rendering accuracy nightmare (Re: Freetype-devel Digest, Vol 184, Issue 24)

2020-05-16 Thread Adam Twardoch (Lists) 
On Sat, 16 May 2020 at 05:32, piotrunio-2...@wp.pl 
wrote:

> It is very important to take research into TrueType rendering to properly
simulate the Microsoft bilevel renderer because it would help render fonts
correctly.

That is excellent piece of advice. So it seems that you know what you need
to do: you should take research into TrueType rendering to properly
simulate the Microsoft bilevel renderer.

As you've noticed, FreeType is complex. It has many different modes.
Screenshots from FontForge don't have any decisive value, because they're
just screenshots from one client app.

As for the rest: Hin-Tak provided you with some useful pointers so you can
start that research.

And yes, the arrogant-aggressive stance isn’t exactly encouraging or
helpful.

Best,
Adam

Re: [ft-devel] An analysis of Librsvg

2019-05-21 Thread Adam Twardoch (Lists) 
SVG Native Viewer of course offers another kind of link:

The
http://www.w3.org/TR/svg-native spec is authored by Adobe’s Sairus Patel
and Apple’s Myles Maxfield. Before SVG was put into the OpenType spec, the
SVG Glyphs spec was co-developed by Sairus
https://www.w3.org/2013/10/SVG_in_OpenType/ and Sairus was the first person
who jumped at my idea of putting SVG into OpenType, and steered the working
group all the way through.

Given that Adobe has contributed the PS rendering code into FreeType, and
are co-developing the SVG Native Viewer, I think it's likely that they'll
be more than happy to work with FreeType devs and, is needed, answer
questions and "make things happen" on their end. And Sairus is a great,
very open-minded fellow, who, well, co-authored both OT-SVG and SVG Native,
and also has once written portions of the Adobe FDK for OpenType. He is
both a font guy and an SVG guy :)

A.

On Tue, 21 May 2019 at 07:13, Vincent Torri  wrote:

> On Tue, May 21, 2019 at 6:35 AM Werner LEMBERG  wrote:
> >
> >
> > > I've mentioned this discussion to the resvg dev, and he seems pretty
> > > interested:
> > >
> > > https://github.com/RazrFalcon/resvg/issues/135
> >
> > Thanks.
> >
> > > I can imagine that in case of resvg, the dev could be willing to
> > > implement some profiles (such as OT-SVG and SVG Native) which, when
> > > used, could filter unwanted content completely.
> >
> > Sounds good.  Moazin plans to publish similar overviews for other SVG
> > engines, too.  Right now it seems that some engines miss an important
> > feature necessary for SVG OT, namely the possibility to select an
> > object by its ID...
>
> what annoys me with resvg is that another compiler is needed. I guess
> that for unix, there is no problem with package managers, as resvg is
> is already compiled. For (native) Windows, i will be a bit more
> complicated
>
> Vincent Torri
>
___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] An analysis of Librsvg

2019-05-20 Thread Adam Twardoch (Lists) 
I've mentioned this discussion to the resvg dev, and he seems pretty
interested:

https://github.com/RazrFalcon/resvg/issues/135

I can imagine that in case of resvg, the dev could be willing to implement
some profiles (such as OT-SVG and SVG Native) which, when used, could
filter unwanted content completely.

A.
___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] The criterion for comparing SVG Rendering libraries

2019-05-14 Thread Adam Twardoch (Lists) 
For completeness, I’d like to add that the license should also be a
criterion. I think librsvg is GPL+LGPL, resvg is MPL, and the Adobe SVG
Native viewer is Apache 2.

It’s worth realizing that SVG actually originated in the late 1990s with
strong participation from Adobe, and the Adobe SVG Viewer (
https://www.adobe.com/devnet/svg/adobe-svg-viewer-download-area.html) was
one of the earliest implementations.

But after the Macromedia acquisition in 2006, Adobe dropped their interest
in SVG and concentrated on Flash. Apple’s resistance to put Flash into iOS
gradually devalued the Adobe Flash platform, and now we’re seeing Adobe
(along with others) pick up their work on SVG after a 10+-year hiatus.

I wouldn’t be surprised if the newly-opensourced Adobe SVG Native viewer
was a spin-off of the Adobe SVG Viewer from 20 years ago.

It seems to me that the recent Adobe opensourcing of the Adobe SVG Native
viewer (which is  C++ 11) and their work on the SVG Native spec shows
serious commitment on Adobe’s side. It's hard to predict, but I suspect
that this project will be well-maintained and will become the institutional
support from possibly also other players.

It wouldn’t be a great surprise to me if, say, someone like Qt would at
some point integrate the Adobe code into the Qt framework and replace their
old SVG Tiny implementation.

I also know that resvg (which is written in Rust) is rather active, but it
seems that it's an individual development.

Best,
Adam
___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] TTFAutohint Without TT Point Numbers

2016-04-11 Thread Adam Twardoch (Lists) 
Things like "point #gre356hlue3752hlye" or "point that is located at
x=36.87% y=78.56% of the glyph bbox"
On Di., 12. Apr. 2016 at 07:41, Werner LEMBERG  wrote:

>
> > I've had situations where I relied on glyph names but I had those
> > where I relied on glyph ids (numbers) and it was a superior choice.
> > If I have 1239 glyphs in one font and 1239 glyphs in another,
> > chances that the glyph IDs are consistent are actually high. I only
> > need to check what's going on if one font has 1239 and the other
> > 1242 glyphs.
> >
> > But with names, I have to check everything.
> >
> > I think the same is true for point numbers vs some other
> > identification.  Relying on point numbers is actually not as bad as
> > some might think. :)
>
> Well, what could `some other identification' be?
>
>
> Werner
>
___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] ft2demos with X11 on OS X (XQuartz) (Re: Freetype-devel Digest, Vol 134, Issue 22)

2016-03-26 Thread Adam Twardoch (Lists) 
Oh, I see now that
/usr/X11 and /usr/X11R6 are symlinked (rightly) to /opt/X11

So there's a bunch of things like libfreetype, libcairo etc. in /opt/X11
(aka /usr/X11) and, separately, in /usr/local

But this is not different from what I had a month ago.



On Sat, Mar 26, 2016 at 5:52 PM, Adam Twardoch (Lists) <
list.a...@twardoch.com> wrote:

> Ah, yes. I do indeed have cairo, mupdf installed via brew in /usr/local,
> freetype is also there (built myself). mupdf comes from the "x11" group of
> things installed by brew. When I try to launch mupdf-x11, I get
>
> error: cannot open display
> uncaught exception: cannot open display
>
> Then I have XQuartz installed in /opt/x11
>
> I guess these two things are in conflict somehow.
>
> So, I do have freetype both in
> /usr/local/include/freetype2/ and /usr/local/lib
> and in
> /opt/X11/include/freetype2/ and /opt/X11/lib
>
> Do you know how I could build the freetype2-demos tools to link to
> freetype that is in /opt/X11 ?
>
> A.
>
>
> On Sat, Mar 26, 2016 at 4:52 PM, Hin-Tak Leung <
> ht...@users.sourceforge.net> wrote:
>
>> do you have multiple versions of some libraries (possibly freetype) under
>> both /opt/X11/lib and /usr/local/lib ? Possibly also cairo, etc?
>>
>> The X server should be launched on demand by launchd, but it does depend
>> on you trying to access those versions of libraries under /opt/X11, and
>> not other
>> compatible ones under /usr/local/lib, AFAIK.
>>
>> 
>> On Sat, 26/3/16, freetype-devel-requ...@nongnu.org <
>> freetype-devel-requ...@nongnu.org> wrote:
>>
>>  Date: Sat, 26 Mar 2016 08:29:07 +0100
>>  From: "Adam Twardoch (Lists)" <list.a...@twardoch.com>
>>  To: "freetype-devel@nongnu.org"
>>  <freetype-devel@nongnu.org>
>>  Subject: [ft-devel] ft2demos with X11 on OS X (XQuartz)
>>  Message-ID:
>>
>>  <cag32g7nqw428x6jaexegroku2_wth6n+maxzsz_k+1dfdqd...@mail.gmail.com>
>>  Content-Type: text/plain; charset="utf-8"
>>
>>  I once somehow had managed to build a version of ft2demos
>>  tools which
>>  worked with the OS X X11 (XQuartz) server.
>>
>>  But now, with the standard "autogen.sh then make" of current
>>  git FreeType
>>  plus the "make" of current git ft2demos, the tools build but
>>  when trying to
>>  use them, I get:
>>
>>  ftmulti 72 CinzelGX.ttf
>>  cannot open X11 display
>>  could not allocate display surface
>>
>>error = 0x
>>
>>  Perhaps this is because in the recent versions of Mac OS X,
>>  X11 is not
>>  included by default but the optional XQuartz installer (
>>  http://www.xquartz.org/ ) installs X11 into:
>>
>>  /opt/X11
>>
>>  Maybe this could be automatically recognized by FT?
>>
>>
>> ___
>> Freetype-devel mailing list
>> Freetype-devel@nongnu.org
>> https://lists.nongnu.org/mailman/listinfo/freetype-devel
>>
>
>
___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] ft2demos with X11 on OS X (XQuartz)

2016-03-26 Thread Adam Twardoch (Lists) 
In xterm, there is extra

X11_PREFS_DOMAIN=org.macosforge.xquartz.X11
XTERM_LOCALE=en_US.UTF-8
XTERM_SHELL=/bin/bash
XTERM_VERSION=XTerm(322)

but otherwise the env vars are the same.
A.


On Sat, Mar 26, 2016 at 4:15 PM, suzuki toshiya <mpsuz...@hiroshima-u.ac.jp>
wrote:

> There is no difference in environment variables
> between the shell in xterm and that in Terminal.app?
>
> Adam Twardoch (Lists) wrote:
>
>> Well, I guess the problem may be on a different level. When I launch
>> `ftmulti` from the XQuartz-provided terminal (xterm), then it launches
>> correctly (with the GUI). But if I launch it from within the regular "text"
>> Terminal or iTerm apps, then I get the error. Same with `ftview` etc.
>> cannot open X11 display
>> could not allocate display surface
>>
>>   error = 0x0002
>>
>> (The error is sometimes 0x and sometimes 0x0002).
>> But the weird thing is that this used to work, 100%, just a few weeks
>> ago. I did pull the new FreeType and built it, and then I rebuilt the
>> ft2demos tools.
>> So this has to have something to do with the way FreeType has been built.
>> A.
>>
>>
>> On Sat, Mar 26, 2016 at 12:44 PM, suzuki toshiya <
>> mpsuz...@hiroshima-u.ac.jp <mailto:mpsuz...@hiroshima-u.ac.jp>> wrote:
>>
>> Dear Adam,
>>
>> Do you mean that other X11 programs (e.g. xterm) could
>> be started correctly, but ft2demos could not be?
>> I will try to find some access to the recent Mac OS X...
>>
>> Regards,
>> mpsuzuki
>>
>> Adam Twardoch (Lists) wrote:
>>  > I once somehow had managed to build a version of ft2demos tools
>> which worked with the OS X X11 (XQuartz) server.
>>  >
>>  > But now, with the standard "autogen.sh then make" of current git
>> FreeType plus the "make" of current git ft2demos, the tools build
>> but when trying to use them, I get:
>>  >
>>  > ftmulti 72 CinzelGX.ttf
>>  > cannot open X11 display
>>  > could not allocate display surface
>>  >
>>  >   error = 0x
>>  >
>>  > Perhaps this is because in the recent versions of Mac OS X, X11
>> is not included by default but the optional XQuartz installer (
>> http://www.xquartz.org/ ) installs X11 into:
>>  >
>>  > /opt/X11
>>  >
>>  > Maybe this could be automatically recognized by FT?
>>  >
>>  > A.
>>  >
>>  >
>>
>>
>>
>
___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] ft2demos with X11 on OS X (XQuartz)

2016-03-26 Thread Adam Twardoch (Lists) 
Nikolaus,

I tried both

On Sat, Mar 26, 2016 at 4:20 PM, Nikolaus Waxweiler 
wrote:

> Out of curiosity, are you starting the tools from the build directory or
> after you `make install`?
>
> ___
> Freetype-devel mailing list
> Freetype-devel@nongnu.org
> https://lists.nongnu.org/mailman/listinfo/freetype-devel
>
___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] ft2demos with X11 on OS X (XQuartz) (Re: Freetype-devel Digest, Vol 134, Issue 22)

2016-03-26 Thread Adam Twardoch (Lists) 
Ah, yes. I do indeed have cairo, mupdf installed via brew in /usr/local,
freetype is also there (built myself). mupdf comes from the "x11" group of
things installed by brew. When I try to launch mupdf-x11, I get

error: cannot open display
uncaught exception: cannot open display

Then I have XQuartz installed in /opt/x11

I guess these two things are in conflict somehow.

So, I do have freetype both in
/usr/local/include/freetype2/ and /usr/local/lib
and in
/opt/X11/include/freetype2/ and /opt/X11/lib

Do you know how I could build the freetype2-demos tools to link to freetype
that is in /opt/X11 ?

A.


On Sat, Mar 26, 2016 at 4:52 PM, Hin-Tak Leung <ht...@users.sourceforge.net>
wrote:

> do you have multiple versions of some libraries (possibly freetype) under
> both /opt/X11/lib and /usr/local/lib ? Possibly also cairo, etc?
>
> The X server should be launched on demand by launchd, but it does depend
> on you trying to access those versions of libraries under /opt/X11, and
> not other
> compatible ones under /usr/local/lib, AFAIK.
>
> 
> On Sat, 26/3/16, freetype-devel-requ...@nongnu.org <
> freetype-devel-requ...@nongnu.org> wrote:
>
>  Date: Sat, 26 Mar 2016 08:29:07 +0100
>  From: "Adam Twardoch (Lists)" <list.a...@twardoch.com>
>  To: "freetype-devel@nongnu.org"
>  <freetype-devel@nongnu.org>
>  Subject: [ft-devel] ft2demos with X11 on OS X (XQuartz)
>  Message-ID:
>
>  <cag32g7nqw428x6jaexegroku2_wth6n+maxzsz_k+1dfdqd...@mail.gmail.com>
>  Content-Type: text/plain; charset="utf-8"
>
>  I once somehow had managed to build a version of ft2demos
>  tools which
>  worked with the OS X X11 (XQuartz) server.
>
>  But now, with the standard "autogen.sh then make" of current
>  git FreeType
>  plus the "make" of current git ft2demos, the tools build but
>  when trying to
>  use them, I get:
>
>  ftmulti 72 CinzelGX.ttf
>  cannot open X11 display
>  could not allocate display surface
>
>error = 0x
>
>  Perhaps this is because in the recent versions of Mac OS X,
>  X11 is not
>  included by default but the optional XQuartz installer (
>  http://www.xquartz.org/ ) installs X11 into:
>
>  /opt/X11
>
>  Maybe this could be automatically recognized by FT?
>
>
> ___
> Freetype-devel mailing list
> Freetype-devel@nongnu.org
> https://lists.nongnu.org/mailman/listinfo/freetype-devel
>
___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] ft2demos with X11 on OS X (XQuartz)

2016-03-26 Thread Adam Twardoch (Lists) 
Well, I guess the problem may be on a different level. When I launch
`ftmulti` from the XQuartz-provided terminal (xterm), then it launches
correctly (with the GUI). But if I launch it from within the regular "text"
Terminal or iTerm apps, then I get the error. Same with `ftview` etc.

cannot open X11 display
could not allocate display surface

  error = 0x0002

(The error is sometimes 0x and sometimes 0x0002).

But the weird thing is that this used to work, 100%, just a few weeks ago.
I did pull the new FreeType and built it, and then I rebuilt the ft2demos
tools.

So this has to have something to do with the way FreeType has been built.

A.


On Sat, Mar 26, 2016 at 12:44 PM, suzuki toshiya <mpsuz...@hiroshima-u.ac.jp
> wrote:

> Dear Adam,
>
> Do you mean that other X11 programs (e.g. xterm) could
> be started correctly, but ft2demos could not be?
> I will try to find some access to the recent Mac OS X...
>
> Regards,
> mpsuzuki
>
> Adam Twardoch (Lists) wrote:
> > I once somehow had managed to build a version of ft2demos tools which
> worked with the OS X X11 (XQuartz) server.
> >
> > But now, with the standard "autogen.sh then make" of current git
> FreeType plus the "make" of current git ft2demos, the tools build but when
> trying to use them, I get:
> >
> > ftmulti 72 CinzelGX.ttf
> > cannot open X11 display
> > could not allocate display surface
> >
> >   error = 0x
> >
> > Perhaps this is because in the recent versions of Mac OS X, X11 is not
> included by default but the optional XQuartz installer (
> http://www.xquartz.org/ ) installs X11 into:
> >
> > /opt/X11
> >
> > Maybe this could be automatically recognized by FT?
> >
> > A.
> >
> >
>
___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

[ft-devel] ft2demos with X11 on OS X (XQuartz)

2016-03-26 Thread Adam Twardoch (Lists) 
I once somehow had managed to build a version of ft2demos tools which
worked with the OS X X11 (XQuartz) server.

But now, with the standard "autogen.sh then make" of current git FreeType
plus the "make" of current git ft2demos, the tools build but when trying to
use them, I get:

ftmulti 72 CinzelGX.ttf
cannot open X11 display
could not allocate display surface

  error = 0x

Perhaps this is because in the recent versions of Mac OS X, X11 is not
included by default but the optional XQuartz installer (
http://www.xquartz.org/ ) installs X11 into:

/opt/X11

Maybe this could be automatically recognized by FT?

A.
___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel