Fix logging error font-patcher

[llc0930]

Description

Python complains about error on line 1881 in font-patcher.
Maybe you guys should take the time to check if the function global logger is doing exactly the right thing.

Requirements / Checklist

• Read the Contributing Guidelines
• Verified the license of any newly added font, glyph, or glyph set

What does this Pull Request (PR) do?

Fix typo in font-patcher.

How should this be manually tested?

I'm not sure, maybe this feature isn't finished yet?

Any background context you can provide?

Python complains about error when I manually patch fonts.
The shell script I use is:

for p in ../iosevka-custom-mono/ttf/*.ttf; do
	echo -e "Patching for font \e[0;36m${p:27}\e[0m ..."
	fontforge -script font-patcher "$p" -cq --careful -out iosevka-custom-mono-nerd --makegroups 0
done

When I do not set the parameter --makegroups 0, it reports an error.

[Finii]

Thanks for finding and fixing this!

There are two more affected calls in the code. To prevent that I added an additional commit that simplifies the whole logger business and allows the programmer to use logger throughout and not decide on a case-to-case basis if logger or logging is the correct thing to use.

Maybe you guys should take the time to check if the function

Yes. That is often skipped due to laziness.
On the other hand that logging message you stumbled over is just issued if some user did not follow the instructions and is kind of a saveguard for 'shall never happen'. Misuse cases are tested less rigorously, I have to admit.

P.S. I guess you should use --makegroups 4, but of course I do not know your font files. At least the 'original' Iosevka needes that.

[Finii]

Add skip-ci, force push

[llc0930]

In fact in older versions I used to just use the command:
fontforge -script font-patcher "$p" -q -c --careful -out iosevka-custom-mono-nerd

But when I did the long-lost update yesterday, I found that there would be a warning, and it is still the same now.

So why doesn't the default --makegroups 1 work?

Thanks for the suggestion, but --makegroups 4 doesn't work either.

[llc0930]

I use Debian sid, here is the script file I use if it helps.
iosevka_build.sh.txt
func.sh.txt

[Finii]

How did you obtain the font-patcher?

https://github.com/ryanoasis/nerd-fonts#font-patcher

Edit:

--makegroups 0 is a last resort when the FontnameParser can not be loaded. It handles all the font name related stuff.
The resulting fonts might work, but most times not if there are multiple weights in a font set.

[llc0930]

Do I need python3-argparse-addons?
I think I have all other dependencies.

[Finii]

I believe you dont. Just try to run the patcher, it should tell you missing dependencies.

[llc0930]

After installing python3-argparse-addons, the same error is reported without the parameter --makegroups, and --makegroups 4 also does not work.
Ok, I'll try running the patcher directly.

[Finii]

Same error, meaning FontnameParser missing?

How did you obtain the font-patcher?

[llc0930]

I figured out what was wrong, I didn't need python3-argparse-addons, my git sparse-checkout command in shell script was wrong.
Instead of git sparse-checkout set "src" I should use git sparse-checkout set --no-cone bin/scripts/name_parser src/glyphs /font-patcher.
There should be no problem now, I will look at the difference between --makegroups 1 and --makegroups 4, thank you.

[Finii]

As written in the readme, reccommended is using the FontPatcher.zip or the docker image, and not a (sparse) checked out repo.

But now it does not matter anymore ;) 👍

[Finii]

difference between --makegroups 1 and --makegroups 4

It's not very well documented.

1 has all name parts in full (no abbreviations). Often the resulting name are longer than some old systems can work with. This is the best for modern systems but can have problems on some.

2 abbreviates the the style and weight parts a bit, like "ExtraBold" becomes "ExtBld".

3 abbreviates them even more aggressive, so "ExtraBold" becomes "XBd" or something.

The options 4 to 6 are similar, but tackle additionally the problem when not the styles and weights are the problem but the plain family name. Instead of adding "Nerd Font Mono" or similar to the family name it adds the abbreviated form "NFM". This is needed where the original name is already very long.

5 and 6 behave like 2 and 3, but with NFM etc.

Option 0 produces names with a much simpler scheme, which breaks often.

Option -1 tells the patcher to leave the names alone and not touch them. This can be used for example when patching over an already Nerd Fonts patched font (yes, people do that) or similar.

[llc0930]

Thanks for the detailed explanation, I ended up going with option -1 yesterday because the fonts I patched are only used on my own system.