Re: [icu-support] Collation rule giving unexpected results

[Keith S. <de...@th...>]

Mark Davis wrote:
> The problem is that {vU}{MA}{sASAT}{sDOT BELOW} is not in normalized 
> order.
>
> Whenever you are sorting languages that use more than one combining 
> mark, you need to turn normalization on. So try putting this statement 
> at the top of your rules:
>
> [normalization on]
>
Yes, using [normalization on] does seem to fix it.

I am using the encoding order defined in:
http://www.unicode.org/versions/Unicode5.1.0/#Myanmar Table 11-3.
It is unfortunate that normalization is broken for Burmese.
I guess it is OK for collation, so long as the normalized strings are 
not used for anything else.

thanks for you help,

Keith