Open Bug 1951981 Opened 1 year ago Updated 1 year ago

Can't find monospace font for thai on Linux (was: Somehow a specific plaintext file or its view-source equivalent aren't monospace)

Categories

(Core :: Layout: Text and Fonts, defect)

Product:
Core
Component:
Layout: Text and Fonts
Type:
defect

Tracking

()

People

(Reporter: emilio, Unassigned)

References

(
URL
)

Details

Going to https://www.nat.go.th/jquery.min.map shows non-monospace text for me and James Graham at least (on Linux).

This repros on a clean profile and I can't make any sense of it (this is without any font fingerprinting protection stuff or what not, plus other plaintext files work fine).

Changing font-family on the inspector doesn't change anything (other than the size due to the monospace quirk). Even sans-serif or something doesn't change the font... But other properties do work (??)

Flags: needinfo?(jfkthame)

Err, of course forgot to paste the last bit:

Jonathan do you have any idea off the top of your head about what might be going on?

I was thinking that maybe content-language or something could affect it but it doesn't seem to be specified according to devtools.

So the plaintext file has ~never been monospace (tried far back to FF 40 or so).

But the view-source used to be monospace until some change around here: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=0f6958f498427fad830cdf94e279d8acd16402c9&tochange=f09f24f2b545919ab0db4bd352c276779c58bcc6

Bug 1551276 somehow is the only loosely related thing, I believe?

See Also: → chardetng

So document.characterSet is windows-874 for us and windows-1252 for Chrome. I don't think that's necessarily the culprit but is that problematic somehow Henri?

Flags: needinfo?(hsivonen)

Ok, so that does explain... when the charset changes, RecomputeLanguageFromCharset() says th, and I guess we don't have any monospace fonts on Linux for th somehow? Seems we should still fallback to something monospace tho :)

So, reduced test-case:

data:text/html,<html lang="th"><pre>Am I monospace</pre>
Summary: Somehow a specific plaintext file (or its view-source equivalent) aren't monospace → Can't find monospace font for thai on Linux (was: Somehow a specific plaintext file or its view-source equivalent aren't monospace)

Yeah, if we decide the content language is th, then we'll use the generic font pref settings for Thai, and probably fontconfig doesn't find a monospaced Thai font.

Given that the content of the file is (largely, if not entirely?) Latin-script text, it would be preferable for us to use the Latin font prefs, regardless of what we think the content language is. The changes planned for bug 1748636 should make that work.

Flags: needinfo?(jfkthame)

(In reply to Emilio Cobos Álvarez (:emilio) from comment #5)

So document.characterSet is windows-874 for us and windows-1252 for Chrome. I don't think that's necessarily the culprit but is that problematic somehow Henri?

I didn't actually debug this, but I'm pretty sure that what's happening here is:

  • Autodetection happens, because this is text/plain without an encoding label.
  • The content doesn't have enough non-ASCII (superficially looks 100% ASCII) for detection from content to work, so the encoding comes from the .th TLD, which maps to windows-874.
Flags: needinfo?(hsivonen)

The severity field is not set for this bug.
:jfkthame, could you have a look please?

For more information, please visit BugBot documentation.

Flags: needinfo?(jfkthame)

Ok, so that does explain... when the charset changes, RecomputeLanguageFromCharset() says th, and I guess we don't have any monospace fonts on Linux for th somehow? Seems we should still fallback to something monospace tho :)

So, reduced test-case:

data:text/html,<html lang="th"><pre>Am I monospace</pre>

For me (on Ubuntu), this (and the original example) does render with monospace (Noto Mono, to be exact). Presumably your fontconfig setup is somewhat different, and that leads to the differing behavior.

Out of curiosity, what does LANG=th fc-match -a monospace | head list for you? And what font is actually being used for the content? (I get Noto Sans Thai as the top match, followed by Noto Mono and other monospaced fonts.)

Severity: -- → S4
Flags: needinfo?(jfkthame) → needinfo?(emilio)

(Marking this as S4, given that it seems fairly edge-casey, depending on local fontconfig setup. Hopefully future font-prefs redesign will help here, anyhow.)

(In reply to Jonathan Kew [:jfkthame] from comment #10)

Out of curiosity, what does LANG=th fc-match -a monospace | head list for you? And what font is actually being used for the content? (I get Noto Sans Thai as the top match, followed by Noto Mono and other monospaced fonts.)

$ LANG=th fc-match -a monospace
FreeSerif.otf: "FreeSerif" "ปกติ"
NotoSansMono-Regular.ttf: "Noto Sans Mono" "Regular"
NotoSansMono-Medium.ttf: "Noto Sans Mono" "Medium"
NotoSansMono-Light.ttf: "Noto Sans Mono" "Light"
NotoSansMono-Thin.ttf: "Noto Sans Mono" "Thin"
NotoSansMono-Bold.ttf: "Noto Sans Mono" "Bold"
NotoSansMono-Black.ttf: "Noto Sans Mono" "Black"
NimbusMonoPS-Regular.otf: "Nimbus Mono PS" "Regular"
NimbusMonoPS-Bold.otf: "Nimbus Mono PS" "Bold"
NimbusMonoPS-Italic.otf: "Nimbus Mono PS" "Italic"
NimbusMonoPS-BoldItalic.otf: "Nimbus Mono PS" "Bold Italic"
FreeMono.otf: "FreeMono" "Regular"
FreeMonoBold.otf: "FreeMono" "Bold"
...

And what font is actually being used for the content?

Free Serif.

I guess Free Serif gets chosen because it has th explicitly on its styles and so on:

Pattern has 44 elts (size 48)
	family: "FreeSerif"(w)
	familylang: "en"(s)
	style: "ปกติ"(w) "Regular"(w) "нормален"(w) "normal"(w) "obyčejné"(w) "Mittel"(w) "µεσαία"(w) "Normaali"(w) "Normál"(w) "Normale"(w) "Gemiddeld"(w) "odmiana zwykła"(w) "Обычный"(w) "Normálne"(w) "menengah"(w) "прямій"(w) "Navadno"(w) "vidējs"(w) "normalusis"(w) "عادی"(w) "vừa"(w) "Arrunta"(w) "सामान्य"(w)
	stylelang: "th"(s) "en"(w) "bg"(w) "ca"(w) "cs"(w) "de"(w) "el"(w) "fi"(w) "hu"(w) "it"(w) "nl"(w) "pl"(w) "ru"(w) "sk"(w) "id"(w) "uk"(w) "sl"(w) "lv"(w) "lt"(w) "fa"(w) "vi"(w) "eu"(w) "hi"(w)
	fullname: "FreeSerif"(w)
	fullnamelang: "en"(s)
	slant: 0(i)(s)
	weight: 80(f)(s)
	width: 100(f)(s)
	size: 12(f)(s)
	pixelsize: 12.5(f)(s)
	foundry: "GNU "(w)
	antialias: True(w)
	hintstyle: 1(i)(w)
	hinting: True(w)
	verticallayout: False(s)
	autohint: False(s)
	globaladvance: True(s)
	file: "/usr/share/fonts/gnu-free/FreeSerif.otf"(w)
	index: 0(i)(w)
	outline: True(w)
	scalable: True(w)
	dpi: 75(f)(s)
	rgba: 1(i)(w)
	scale: 1(f)(s)
	charset: [ommitted]
(w)
	lang: aa|ab|af|am|ar|as|ast|av|ay|az-az|az-ir|ba|bm|be|bg|bh|bho|bi|bin|bn|br|bs|bua|ca|ce|ch|chm|chr|co|cs|cu|cv|cy|da|de|el|en|eo|es|et|eu|fa|fi|fj|fo|fr|ff|fur|fy|ga|gd|gez|gl|gn|gu|gv|ha|haw|he|hi|ho|hr|hu|hy|ia|ig|id|ie|ik|io|is|it|ka|kaa|ki|kk|kl|kok|ks|ku-am|ku-ir|kum|kv|kw|ky|la|lb|lez|ln|lt|lv|mg|mh|mi|mk|ml|mo|mr|mt|nb|nds|ne|nl|nn|no|nr|nso|ny|oc|om|or|os|pa|pl|ps-af|ps-pk|pt|rm|ro|ru|sa|sah|sco|se|sel|sh|shs|si|sk|sl|sm|sma|smj|smn|sms|so|sq|sr|ss|st|sv|sw|ta|tg|th|ti-er|ti-et|tig|tk|tl|tn|to|tr|ts|tt|tw|tyv|ug|uk|ur|uz|ve|vi|vo|vot|wa|wen|wo|xh|yap|yi|yo|zu|ak|an|ber-dz|byn|crh|csb|dv|ee|fat|fil|hne|hsb|ht|hz|jv|kab|kj|kr|ku-iq|ku-tr|kwm|lg|li|mai|mn-mn|ms|na|ng|nv|ota|pa-pk|pap-an|pap-aw|qu|quz|rn|rw|sc|sd|sg|sid|sn|su|ty|wal|za|lah|brx|sat|doi|mni|und-zmth|anp|bhb|hif|mag|raj|the|agr|ayc|bem|ckb|dsb|lij|mfe|mhr|miq|mjw|nhn|niu|rif|sgs|szl|tpi|unm|wae|yuw|got(s)
	fontversion: 27015662(i)(s)
	capability: "otlayout:DFLT otlayout:arab otlayout:armn otlayout:beng otlayout:bng2 otlayout:bugi otlayout:copt otlayout:cyrl otlayout:dev2 otlayout:deva otlayout:gjr2 otlayout:glag otlayout:goth otlayout:grek otlayout:gujr otlayout:gur2 otlayout:guru otlayout:hano otlayout:hebr otlayout:latn otlayout:mlm2 otlayout:mlym otlayout:musc otlayout:ory2 otlayout:orya otlayout:sinh otlayout:taml otlayout:thai otlayout:tml2"(w)
	fontformat: "CFF"(w)
	embeddedbitmap: True(s)
	decorative: False(s)
	lcdfilter: 1(i)(w)
	namelang: "th"(s)
	prgname: "fc-match"(s)
	postscriptname: "FreeSerif"(w)
	color: False(w)
	symbol: False(s)
	variable: False(s)
	fonthashint: False(w)
	order: 0(i)(s)
	desktop: "KDE"(s)
	namedinstance: False(w)
	fontwrapper: "SFNT"(w)
Flags: needinfo?(emilio)

(In reply to Emilio Cobos Álvarez (:emilio) from comment #12)

I guess Free Serif gets chosen because it has th explicitly on its styles and so on:

Yeah, given that it explicitly claims Thai support:

lang: aa|ab|af|am|ar|as|ast|av|ay|az-az|az-ir|ba|bm|be|bg|bh|bho|bi|bin|bn|br|bs|bua|ca|ce|ch|chm|chr|co|cs|cu|cv|cy|da|de|el|en|eo|es|et|eu|fa|fi|fj|fo|fr|ff|fur|fy|ga|gd|gez|gl|gn|gu|gv|ha|haw|he|hi|ho|hr|hu|hy|ia|ig|id|ie|ik|io|is|it|ka|kaa|ki|kk|kl|kok|ks|ku-am|ku-ir|kum|kv|kw|ky|la|lb|lez|ln|lt|lv|mg|mh|mi|mk|ml|mo|mr|mt|nb|nds|ne|nl|nn|no|nr|nso|ny|oc|om|or|os|pa|pl|ps-af|ps-pk|pt|rm|ro|ru|sa|sah|sco|se|sel|sh|shs|si|sk|sl|sm|sma|smj|smn|sms|so|sq|sr|ss|st|sv|sw|ta|tg|th|ti-er|ti-et|tig|tk|tl|tn|to|tr|ts|tt|tw|tyv|ug|uk|ur|uz|ve|vi|vo|vot|wa|wen|wo|xh|yap|yi|yo|zu|ak|an|ber-dz|byn|crh|csb|dv|ee|fat|fil|hne|hsb|ht|hz|jv|kab|kj|kr|ku-iq|ku-tr|kwm|lg|li|mai|mn-mn|ms|na|ng|nv|ota|pa-pk|pap-an|pap-aw|qu|quz|rn|rw|sc|sd|sg|sid|sn|su|ty|wal|za|lah|brx|sat|doi|mni|und-zmth|anp|bhb|hif|mag|raj|the|agr|ayc|bem|ckb|dsb|lij|mfe|mhr|miq|mjw|nhn|niu|rif|sgs|szl|tpi|unm|wae|yuw|got(s)

(and yes, it does include Thai chars), it makes sense that fontconfig will return it as the best match when given lang=th in the request. And sadly, that font includes (non-monospaced) Latin chars, so that's what you see.

For me, I get Noto Thai as the first match, but it does not include Latin letters, it's Thai-only. So it doesn't get in the way of the Latin chars falling back to the next match in the list, which is a suitably-monospaced font.

You need to log in before you can comment on or make changes to this bug.