Talk:Fonts

From ArchWiki
Latest comment: 15 April by Andrei Korshikov in topic List font containing a specific glyph?

‘Set the proper key map’?

FONT_MAP does not set the key map.

I put the following code into /etc/profile.d/locale.sh:

(case "${LANG}" in
pl_*) FONT=lat2-16;;
*) let 0;;
esac && setfont $FONT)

This allows the user to use a different locale than the system default. Of course, the list of supported languages must be extended.

--Yecril71pl (talk) 13:50, 11 May 2013 (UTC)Reply[reply]

Is xset fp rehash required?

When installing local fonts, this page suggests adding the following the .xinitrc file:

xset +fp /usr/share/fonts/local 
xset fp rehash

My understanding is that xset fp rehash is only required to reload the font path on an active X session, therefore not needed in the .xinitrc file.

Mirror web fonts

Hi! I mirrored two websites with fonts, made a PKGBUILD and added to AUR. Can it be any problems with it? ttf-myanmar-fonts, ttf-devanagarifonts. Fonts are free for downloading, should I ask website owners before publish them here or it might be ok?

Why have you had to mirror the websites if the fonts were "free for downloading"? Where have you put the links to the original sources? What does "free for download" mean exactly? What licence are those fonts released under? If you're asking it means you're not 100% sure you could re-distribute them, so you'd better double check that, possibly with the authors of the fonts. -- Kynikos (talk) 12:54, 23 September 2014 (UTC)Reply[reply]
Free for download: I made packages for installation, to install all for a language specialized fonts at once instead one by one. And for some of languages was no font packages in arch wiki. Original links are in a description of packages or should I add link to websites here in wiki? I will double check now. Here are my pkgs in aur. Andy Crowd 17:25, 23 September 2014 (UTC)
Yes please, you have to check whether you can redistribute them or not, and under what limitations, if any. If the urls you've put in the PKGBUILDs are the original sources, that's ok, there's no need to show them on the wiki. -- Kynikos (talk) 03:43, 25 September 2014 (UTC)Reply[reply]
I have send mails about licenses to the owners of those websites with fonts but got no answer yet. Andy Crowd 07:39, 25 September 2014 (UTC)

Diactric character input (lat2-16)

I've found it necessary to use FONT_MAP=8859-2 when using FONT=lat2-16 or ter-216n (Terminus) in /etc/vconsole.conf to be able to input characters like ą, ć, ę etc. (Polish diactric characters). Maybe worth adding here. Wujek.srujek (talk)

8859-2 is already specified by selecting FONT=lat2-16, setting FONT_MAP=8859-2 in this case does not have any effect. Terminus fonts are different in that they don't have built-in Unicode translation map (the fonts end with .psf.gz extension instead of .psfu.gz), so they deserve separate instructions. -- Lahwaacz (talk) 06:20, 12 October 2014 (UTC)Reply[reply]

My experiences have been very different than yours, then. The maps included in the fonts map Unicode code points to font glyphs, and there is no problem with this - all characters are printed correctly. The problem is with input from the keyboard. Please see the topic: https://bbs.archlinux.org/viewtopic.php?id=186736. Without FONT_MAP it just doesn't work completely - I can't write Polish characters with my keyboard. You are also wrong about the Terminus fonts not having the Unicode map - calling 'setfont -v ter-216n' results in loading the font and also logs 'Loading Unicode mapping table...'. I also read that fonts with embedded Unicode maps don't always use the psfu extension, which Terminus seems to be an example of. Please try it out, and if you find that I'm doing stuff wrong, maybe comment in the topic I linked? I also think you mixed up FONT_MAP with FONT_UNIMAP, as the latter is the map included in font files. Wujek.srujek (talk)

Improve rendering of fonts with bitmap version embedded

Hi, fonts like Calibri/Cambria doesn't always correctly render in every size (they have embedded bitmap version which activates on certain sizes). The following setting allows to fix this problem and get all sizes rendered correctly. I don't know if there is a better way to configure:

 <match target="font" >
   <edit name="embeddedbitmap" mode="assign">
     <bool>false</bool>
   </edit>
 </match>

If you need a screenshot I can upload some examples. Thanks.

FiNeX (talk) 22:26, 16 October 2017 (UTC)Reply[reply]

This configuration is covered in Font_configuration#Disable_bitmap_fonts. -- Lahwaacz (talk) 06:33, 17 October 2017 (UTC)Reply[reply]
First of all, disabling embed bitmaps absolutely does NOT make the render any more correct. Second of all, every single non-Microsoft TrueType renderer is incorrect (but not all Microsoft TrueType renderers are correct). Third of all, in Segoe UI/Calibri/Consolas/Cambria the hinting is deliberately broken by Microsoft so there isn’t even any correct way of rendering the outlines of these fonts in the first place. PiotrGrochowski (talk) 14:43, 6 March 2023 (UTC)Reply[reply]

Move CJK and Indic fonts to Localization subpages

The Localization subpages partially duplicate this page and I think it makes sense to group language-specific content together.

--Larivact (talk) 03:29, 31 July 2018 (UTC)Reply[reply]

I'm not sure about this, but if you think it is a good idea, feel free to go ahead. -- Lahwaacz (talk) 08:05, 13 August 2018 (UTC)Reply[reply]


I have not make any edit, but I am a recurrent user of localization tools. From my perspective is better to have language-specific content grouped together. The general language page could just have pointer to the specific language subpages, just like the Installation page. Calegria (talk) 07:35, 19 February 2020 (UTC)Reply[reply]

List font containing a specific glyph?

In the section "List installed fonts for a particular Unicode character", it uses `fc-match -s monospace:charset=1F4A9` to list the installed fonts supporting a particular Unicode. But I tried it, say I'm doing fc-match -s monospace:charset=2AED0 and also fc-match -s :charset=2AED0, and it will output so many fonts, even a large number of them don't contain the specified glyph. But use the method from https://unix.stackexchange.com/a/240563, it prints only one font which supports that Unicode, which is what I expect.

Maybe a better way is to use fc-list :charset=<codepoint>?

Bczhc (talk) 03:24, 6 March 2023 (UTC)Reply[reply]

In principle, I do agree. In fact, fc-list may say that glyph is supported, and you still have square in your terminal emulator. - Andrei Korshikov (talk) 09:05, 15 April 2024 (UTC)Reply[reply]