Re: GSUnicodeString and NSString disagree on rangeOfComposedCharacterSequenceAtIndex:

[David Chisnall]

On 7 Apr 2018, at 19:51, David Chisnall <address@hidden> wrote:
> 
> I’m now hitting a failure in the unichar_const tests, because for some reason 
> a GSMutableString and a (UTF-16) NSConstantString are not comparing equal, in 
> spite of having the same hash, the same length, and the same values for both 
> characters...

Ignore that bit - I’d made the same fix in GSMutableString, but hadn’t 
recompiled before testing.  I’m now passing all of the NSString tests, with 
both rangeOfSequence_c and rangeOfSequence_u disabled and the subclasses that 
use it falling back to the superclass implementation.  I don’t know if the 
NSString or the GSString implementation is the correct one, but at least one of 
them is wrong, and the NSString one at least works with other NSString 
subclasses that implement the required primitive methods of NSString.

David