Note that Chrome's behavior is obviously incorrect and is not intentional
in any way. See crbug.com/6122 (yes, it's a four digit bug in Chromium).
On Fri, May 23, 2014 at 1:54 PM, Najib Tounsi <ntounsi@gmail.com> wrote:
> I think one should take into account
> - font size that may differ largely from one font to another ('Arabic
> Typesetting' letters are small compared to others),
> - ligatures, different shapes of same Arabic letters,
> - isolated (non joining) letters should (or not) have diacritic
>
> I share two screen shots (hope self-explained) from Mozilla and Chrome
> In the second line of Arabic, I have added U+200D, the zwj character, to
> force the joining.
>
> Mozilla
> [image: Mozilla]
>
> Chrome
>
> [image: Chrome]
>
> (source here http://196.200.140.8/Tests/testJoinOrNot.html)
>
> I prefer the Gecko approach: if joining is not to be broken this should
> apply ONLY to change in color. Moreover, even with the zwj control
> character, Mozilla BREAKS the joining when font or size change.
>
> Best regards,
>
> Najib
>
>
>
> On 5/22/14 8:09 PM, Behdad Esfahbod wrote:
>
> On 14-05-22 03:28 AM, Matitiahu Allouche wrote:
>
> Fantasai wrote: " I don't think change of font or color should cause a joining break, however."
>
> I agree that change of color should not mandate a joining break. On the other hand, a change of font, or a change of size in the same font, seems to make joining across the change very difficult technically. It would be quite extraordinary that letters from two fonts crafted by different designers join properly.
> If the technical hurdle is such that there is little chance that implementations will overcome it, then it might be better to pronounce that there will be no joining across font or size changes.
>
> I disagree. Joining should NOT be broken even if font changes. Using a bold
> font to highlight a letter in print is a valid use. It's hard. But the
> alternative is wrong. Whether user uses a different font weight or a
> different color to highlight shouldn't affect the correctness of the script
> behavior.
>
>
>
>
