case insensitive find() in wxString

[Francesco Montorsi]

Robert Roebling ha scritto:
>>> FM> This patch would allow to append the "caseCompare cmp = exact" argument 
>>> FM> to various functions (e.g. Contains, Replace, After/Before First/Last, 
>>> FM> StartsWith / EndsWith). Specially for wxString::Contains and Replace() 
>>> FM> that would be very useful IMO.
>>> ....
>>> FM> What's your opinion ?
>>>
>>>  To be honest, I'm undecided. On one hand, I do see how could this be
>>> useful. OTOH we plan to switch to std::string at some moment in the future
>>> and adding yet more features which the standard class doesn't have is going
>>> to make switching even more difficult.
>> BTW, why should we switch to std::string in future ?
> 
> I'm not in favor of that, either, but it has been a request many
> time already.
sorry - you mean you're not in favour to switch to std::string or to 
implement more widely case-insensitive functions support ?

> But what I'm actually wondering is how you handle
> Unicode uppercase. It is easy for a-z, but even Westeuropean
> languages have more letters than these. Is there a native function
> for that or an algorithm etc?
good point. I thought to use wxTolower as all case insensitive wxString 
functions currently do... I see that GNU libc does not provide any 
stricmp() function unlike e.g. MSVC and probably because for some 
Unicode-supported characters the lowercase/uppercase has no sense.
In fact, I see most times wxTolower() is implemented with a call to 
tolower() using an explicit cast from wxChar to char.

Probably we should add a note in wxString docs saying that Case_Ignore 
flag has sense only for e.g. characters < 127. All other characters 
could be considered as "caseless" and thus always compared exactly even 
if Case_Ignore is given...

Francesco