Hey all,

Stepping in again to provide some context. For those unfamiliar I’m an engineer 
at Font Awesome since the version 4 days.

1. SVG files that you find in any of our packages (even back to version 4) are 
artifacts and NOT source files. We’ve authored icons in a variety of tools over 
the years but no one has ever directly edited SVG files. They are always an 
output format that are intentionally optimized and lossy to some degree.

2. Any FontForge metadata you see is facilitated by a build system. A human 
never ran the desktop FontForge app to then generate files. It was always 
scripted.

3. The fontawesome-webfont.svg file from version 4 was also never a source file 
(regardless of whether the ForkAwesome folks considered it as such) It was also 
a generated file from FontLab _I think_.

Happy to answer questions that pop up so speculation doesn’t have to win the 
day.

> On May 18, 2026, at 10:42 AM, Julian Gilbey <[email protected]> wrote:
> 
> Dear all,
> 
> On Fri, May 15, 2026 at 06:39:10PM +0100, Julian Gilbey wrote:
>> On Fri, May 15, 2026 at 05:52:49PM +0100, Nathan Willis wrote:
>> [...]
>>> 
>>> I may have missed something in here, but how do you know that the SVGs in 
>>> the
>>> font are not the source SVGs?
>> 
>> Hi Nathan!
>> 
>> Good question; I'm not sure we do.  The SVG font file
>> (fontawesome-webfont.svg) begins:
>> [...]
>> 
>> So perhaps this is the original source for the font.  FontForge is
>> quite happy to open it and allow editing, and this would not be a
>> reason to consider FontAwesome 4.7.0 to be non-DFSG free.
>> [...]
>> Looking further at FontAwesome 4.7.0:
>> 
>> * fontawesome-webfont.{eot,ttf} and FontAwesome.otf all contain the
>>  string "FONTLAB:OTFEXPORT", so presumably these were generated by
>>  FontLab (non-free software)
>> 
>> * I don't know how to uncompress the .woff/.woff2 files to see if
>>  there is any identifying information there.  But my guess is that
>>  these were also generated using FontLab.
>> [...]
> 
> Thinking more about it, how about the following transition process?
> 
> * Strip all of the fontawesome-webfont.* and FontAwesome.otf files
>  from the source package except for the SVG one.
> 
> * Regenerate all of these fonts using FontForge; here's a simple
>  script that will do fine with a little tweaking:
> 
> -----
> #!/usr/bin/python3
> 
> import fontforge
> 
> SVG = "/tmp/fontawesome-webfont.svg"
> TTF = "/tmp/fontawesome-webfont.ttf"
> WOFF = "/tmp/fontawesome-webfont.woff"
> WOFF2 = "/tmp/fontawesome-webfont.woff2"
> OTF = "/tmp/FontAwesome.otf"
> 
> fa = fontforge.open(SVG)
> fa.copyright = "Copyright Dave Gandy 2016. All rights reserved."
> fa.version = "4.7.0"
> fa.sfntRevision = 4.007
> fa.ascent = 1536
> fa.descent = 256
> 
> for font in [TTF, WOFF, WOFF2, OTF]:
>    fa.generate(font)
> -----
> 
> Note that this does not generate the EOT font; that is fine, as it is
> specific to Internet Explorer and is not needed for Microsoft Edge, so
> is not needed on Debian systems (unless they are serving to old MS
> Windows machines).  If really needed, we could generate that too using
> eot-utils.
> 
> Then my suggestion is that fonts-font-awesome-legacy stores its fonts
> in /usr/share/fonts-font-awesome (not
> /usr/share/fonts-font-awesome-legacy), and includes FontAwesom 4.7.0
> (with this minor change).  The control fields will be roughly, for the
> time being:
> 
> Package: fonts-font-awesome-legacy
> Conflicts and Replaces: fonts-font-awesome (<< 7)
> contains FontAwesome versions 4.7.0, 5.x, 6.x
> 
> Package: fonts-font-awesome
> Version: 7.x
> Depends: fonts-font-awesome-legacy
> contains FontAwesome version 7.x
> 
> Once these are both in testing, we file normal severity bugs against
> all packages that build-depend or depend on fonts-font-awesome asking
> them to either switch to fonts-font-awesome-legacy or to ensure that
> their package will continue to work with fonts-font-awesome 7.x, and
> close the bug when they have done so.  Then in forky+1 (or perhaps
> sooner), we can drop the fonts-font-awesome ->
> fonts-font-awesome-legacy dependency.
> 
> Thoughts before I proceed?
> 
> Best wishes,
> 
>   Julian
> 
> -- 
> To unsubscribe, send mail to [email protected].

Reply via email to