New std.uni: ready for more beating

[Dmitry Olshansky]

26-Feb-2013 12:27, dennis luehring пишет:
> Am 25.02.2013 19:31, schrieb Dmitry Olshansky:
>> 24-Feb-2013 12:32, dennis luehring пОшет:
>>> would it make sense to incoporate test from the ICU testsuite - there
>>> are api tests and many data tests around
>>
>> For key algorithms I'm using consortium's test data files + plus
>> running random generated stress-tests against ICU.
>>
>> It might make sense to incorporate some of their tests but I'm wondering
>> if it'll end up only as a difference in the API.
>
> are doing any benchmark tests again other implementations like ICU or
> Qt's QString, etc.?

Not all that much but this is on todo list.
If anything it takes a fair bit of time to install the lib,
decipher the API and see how to make it do things like normalization
of UTF32 buffer in memory.

The 2 major moments I was concerned with are:
a) all old things should not get slower then std.uni (in fact they are 
many times faster)
b) correct (as in spec) and in line with others - currently only ICU but 
that's the defacto standard

> are there any operations that are slower then other implementations?

Till the data is available. It would be of tremendous help if you can 
test it against some library that you are comfortable with.

-- 
Dmitry Olshansky