Wed, 31 Dec 2014 13:27:57 +0100
Ignore runtime configuration files generated during quality assurance.
michael@0 | 1 | <!DOCTYPE HTML> |
michael@0 | 2 | <html> |
michael@0 | 3 | <head> |
michael@0 | 4 | <meta charset="utf-8"> |
michael@0 | 5 | <title>ALM character</title> |
michael@0 | 6 | <style type="text/css"> |
michael@0 | 7 | p { font-size: 2em; } |
michael@0 | 8 | </style> |
michael@0 | 9 | <!-- |
michael@0 | 10 | Test that U+061C functions as an invisible Arabic character. |
michael@0 | 11 | |
michael@0 | 12 | In the test text it is followed by two European numbers with a |
michael@0 | 13 | European separator between them. By rules W2, W3 and W6 of the |
michael@0 | 14 | Bidi algorithm (http://unicode.org/reports/tr9/#W2) the |
michael@0 | 15 | European numbers change to Arabic numbers, the ARABIC LETTER |
michael@0 | 16 | MARK changes to R and the separator changes to Other Neutral, |
michael@0 | 17 | and then by rule N1 the Arabic numbers make the separator |
michael@0 | 18 | change to R, so the final visual order is "a2+1؜", but |
michael@0 | 19 | the ؜ should not appear |
michael@0 | 20 | --> |
michael@0 | 21 | </head> |
michael@0 | 22 | <body> |
michael@0 | 23 | <p>a؜1+2</p> |
michael@0 | 24 | </body> |
michael@0 | 25 | </html> |