We break runs of CJK ideographs between certain ideograph pairs using
some standard algorithm. We suppress this breaking in runs that are
white-space:pre. one hand, that means authors using white-space to
prevent breaking in regular text also have it working by analogy in CJK
text. the other hand, it makes no sense; there is no whitespace here
for white-space to act upon. What's correct?
Rob
dev-tech-layout mailing list
dev-tech-layout (AT) lists (DOT) mozilla.org

Robert 'Callahan wrote:
We break runs of CJK ideographs between certain ideograph pairs using
some standard algorithm. We suppress this breaking in runs that are
white-space:pre. one hand, that means authors using white-space to
prevent breaking in regular text also have it working by analogy in CJK
text. the other hand, it makes no sense; there is no whitespace here
for white-space to act upon. What's correct?

Without me having much (if anything) to add to this conversation, do you
have any testcases to run on other UA's, such as IE, , Safari,
(others?) to see how interoperable this functionality is.

I'd personally suggest <pre>; and <div style="whitespace:pre"tests
seperate.

~Justin Wood (Callek)

dev-tech-layout mailing list
dev-tech-layout (AT) lists (DOT) mozilla.org

Justin Wood (Callek) wrote:
Robert 'Callahan wrote:
>We break runs of CJK ideographs between certain ideograph pairs using
>some standard algorithm. We suppress this breaking in runs that are
>white-space:pre. one hand, that means authors using white-space to
>prevent breaking in regular text also have it working by analogy in CJK
>text. the other hand, it makes no sense; there is no whitespace here
>for white-space to act upon. What's correct?

Well I'm not expert about that but I just had a look in the CSS2 norm,
and you can find this about using 'white-space: pre' :
#white-space-prop
[] Lines are only broken at newlines in the source, or at
occurrences of "\A" in generated content.

So it seems the current implementation is correct.

Without me having much (if anything) to add to this conversation, do you
have any testcases to run on other UA's, such as IE, , Safari,
(others?) to see how interoperable this functionality is.

I'd personally suggest <pre>; and <div style="whitespace:pre"tests
seperate.

Shifting the topic a bit, currently using <pre>, and therefore also
text/plain, breaks a lot of i18n features. It's not nice when combining
character stop to combine and I believe strange things also happen for
bidi. So I'm all for roc changing things to that regard, and make it
work better, current behavior is not satisfying at all generally, even
if that particular point seems correct.

dev-tech-layout mailing list
dev-tech-layout (AT) lists (DOT) mozilla.org

Jean-Marc Desperrier wrote:
Well I'm not expert about that but I just had a look in the CSS2 norm,
and you can find this about using 'white-space: pre' :
#white-space-prop
[] Lines are only broken at newlines in the source, or at
occurrences of "\A" in generated content.

So it seems the current implementation is correct.

That's a good point, although I don't know if the person who wrote that
was thinking about CJK at the time.

Rob

dev-tech-layout mailing list
dev-tech-layout (AT) lists (DOT) mozilla.org

Jean-Marc Desperrier wrote:
Shifting the topic a bit, currently using <pre>, and therefore also
text/plain, breaks a lot of i18n features. It's not nice when combining
character stop to combine []

when you can not use ZWJ (U+200D) in gmail composition window :

(should look like ):

Images coming from

dev-tech-layout mailing list
dev-tech-layout (AT) lists (DOT) mozilla.org

Robert 'Callahan wrote:
Jean-Marc Desperrier wrote:
>Well I'm not expert about that but I just had a look in the CSS2 norm,
>and you can find this about using 'white-space: pre' :
>#white-space-prop
>[] Lines are only broken at newlines in the source, or at
>occurrences of "\A" in generated content.
>>
>So it seems the current implementation is correct.

That's a good point, although I don't know if the person who wrote that
was thinking about CJK at the time.

I don't think it matters. If you turn text wrapping off, it's off.

~fantasai

dev-tech-layout mailing list
dev-tech-layout (AT) lists (DOT) mozilla.org