186273 - Separator control characters are displayed literally

Status: RESOLVED WONTFIX

(Core :: Layout: Text and Fonts, defect, P5)

Description

[Ilya Konstantinov]

Unicode control characters 001C, 001D, 001E and 001F are displayed literally
(that is, the relevant character from that position in the font is printed),
while they should be interpretted only logically (just as a newline character is
not displayed as-is).

Most handy use of those characters is for BiDi language support (to prevent
unrelated parts of text from being blended by the reordering), but the BiDi
engine in Mozilla seem to classify those characters properly already, so this is
a Layout issue (shouldn't be put under the "BiDi: Hebrew and Arabic" component).

Comment 1

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

The characters you cite aren't bidi control characters, but rather "file
separator", "group separator", "record separator", and "unit separator".  Bidi
control codes are:

LRM   200E
RLM   200F
LRE   202A
RLE   202B
PDF   202C
LRO   202D
RLO   202E

Comment 2

[Ilya Konstantinov]

That is true. I never claimed them to be such either. I just don't know of any
use for them in HTML except for hinting the BiDi ordering engine what pieces of
text are to be considered separate records.

Comment 3

[Simon Montagu :smontagu]

I'm not sure that these characters are even legal in HTML -- they certainly
aren't in XML [1]. Is there no alternative way to do what you want to do, either
by markup or by using other characters? Can you give a concrete example?

[1] http://www.w3.org/TR/REC-xml#NT-Char

Comment 4

[Ilya Konstantinov]

Since it's invalid in XML, I guess I'll just find a way to do what I want with 
BiDi control characters.