On 11/21/2016 2:17 PM, Philippe Verdy wrote:
2016-11-21 23:02 GMT+01:00 Philippe Verdy <[email protected] <mailto:[email protected]>>:You don't need them, I just used lowercase letters for strong LTR characters and uppercase for RTL, just like in the existing Bidi test page. Use some random Arabic or Japanese words if you prefer. 2016-11-21 22:40 GMT+01:00 Asmus Freytag (c) <[email protected] <mailto:[email protected]>>: On 11/21/2016 1:17 PM, Philippe Verdy wrote:Examples were in the initial post I sent in this thread, or in other replies. In encoded order, it should be testing this: ARABIC-ONE "【japanese1】japanese2: “english1, « french1 », or「 japanese3」”。" ARABIC-TWOReplacing "japanese" by its translation in Japanese, and translating ARABIC-ONE and TWO into Arabic (Note: japanese3 is been also translated in Arabic):العربية واحدة "【日本の1】日本の2: “english1, « french1 », or「اليابانية واحد」。" العربية-اثنينThe CJK square quote are not mirrored, they are just swapped, but still do not embed their content as pairs... This is an example of where the simple assignement of direction for quotes from the paragraph direction only does not work, and where detecting pairs or quotes would be necessary to fix their enclosure as isolates at inner levels.n
I get where is the problem? A./
