Custom Postfix for font

[mietzen]

Hey,

I looking for a way to attach a post fix to a font name, e.g. I want:

docker run --rm -v `pwd`/in:/in -v `pwd`/out:/out nerdfonts/patcher --complete --careful 

To create the usual font name, which it does. And:

docker run --rm -v `pwd`/in:/in -v `pwd`/out:/out nerdfonts/patcher --complete

To create the usual font name with the post fix forced.

Is this possible? I tried --name But I don't fully understand this option. Do I then need to create bold, italic, regular, etc. all separated?

If I run with --name 'MyFontForced', I get:

Done with Patch Sets, generating font...
WARNING: Possible problem with the weight metadata detected, check with --debug
DEBUG: Weight approximations: OS2/PS/Name: 600/600/400 (from 600/'Demi'/'')
DEBUG: =====> Family (ID 1)      ok       (14 <=31): MyFont Ced For
DEBUG: =====> SubFamily (ID 2)   ok       ( 7 <=31): Regular
DEBUG: =====> Fullname (ID 4)    ok       (14 <=63): MyFont Ced For
DEBUG: =====> PSN (ID 6)         ok       (12 <=63): MyFontCedFor
DEBUG: =====> Filename '/out/MyFontCedFor-Regular.ttf'

All fonts are called MyFontCedFor-Regular.ttf, why? Also where does Ced For come from? Is this Forced splitted and swapped?

[Finii]

Is this possible? I tried --name But I don't fully understand this option. Do I then need to create bold, italic, regular, etc. all separated?

Fonts are a very organically grown file format, and lots of information is strewn around and contradicting in typical fonts.
The name of a font is stored multiple times in the file, with different twists; but usually that is the best information source.
So the patcher analyses the name (for some definition of 'the') and tries to find out if the concrete font is bold or condensed etc.

The --name option lets you choose what 'the' means:

  --name FORCE_NAME     Specify naming source ('full', 'postscript', 'filename', or concrete free name-string)

The option is kind of mixed, because it takes keywords ('full', 'postscript', 'filename') or a verbatim string.

So if you give --name filename the patcher will use the filename itself as name source, giving --name full is what is usually used automatically (the 'full' name inside the font file is used).

Note that all information - also Bold, Condensed, ExtraLight, etc - are in a font's name, and need to be there to analyze.

Do I then need to create [...] all separated?

Yes and no. When you use --name with a concrete name string that string has to contain the weights and styles, and thus is different for each font in that family. For example --name MyFontForced-CondensedBold:

Note that the weight and styles are separated from the 'real' name by a dash. That is just a convention, but it is usually followed, and the patcher script detects that. And as you see above, when you specify the weights-and-styles group separated with a dash it succeeds in correct determination of the parts.

For your use case, where you have multiple fonts in some family (I guess), you need to call the patcher individually on each font file with the correct --name Whatever-WeightStyle (where Weight and Style must be correct for that font).

Easier probably is to rename the font files prior to patching to something you like, and then use --name filename. But that will append the "Nerd Font" stuff.

mv SomeFont.ttf MyFontForced-CondensedBold.ttf
fontforge font-patcher --name filename --makegroups 4 MyFontForced-CondensedBold.ttf

(--makegroups 4 selects "NF" instead of "NerdFont" in the name, if you prefer, see addendum below)

Hmm, that --name with a concrete name string does not attach the "Nerd Font" suffix to the name is undocumented in the Wiki (https://github.com/ryanoasis/nerd-fonts/wiki/ScriptOptions)
But that behavior is explicitly mentioned when introduced here

• Feature: Select name source #1319

Fixing the Wiki.

---

Addendum: