Re: Question about OUString::compareToAscii

On Friday 13 of July 2012, Christophe JAILLET wrote:
> Hi,
>
> according to my understanding,
>
> OUString::compareToAscii(RTL_CONSTASCII_STRINGPARAM("A string")) gives

 compareAsciiL - the L means it gets the string length from the macro

> exactly the same result as
> OUString::compareToAscii("A string")
>
> However, the first one:
>      - is harder to read
>      - uses the deprecated RTL_CONSTASCII_STRINGPARAM macro
>      - is a bit slower

 It is actually a bit faster (the macro computes string length at 
compile-time), which is presumably how this whole RTL_CONSTASCII_STRINGPARAM 
madness came to be.

> If you agree with that, I will provide a patch to clean it.

 String literals can now mostly be used just like OUString, because many 
string functions now have specialized overloads for them. IIRC compare() does 
not, for a reason I don't quite recall at the moment, but it's probably 
better to leave those as they are for the time being.

> Moreover, and as a 2nd step, I plan to propose a patch that turns:
>      s.compareToAscii("A string") == 0
> into
>      s.equalsAscii("A string")

 s.equals( "A string" )

 or even simpler

 s == "A string"

> and
>
>      ! s.compareToAscii("A string")
> into
>      s.equalsAscii("A string")
>
> I think that equalsAscii is better because:
>      - shorter
>      - more meaningful
>      - doing some ! on sal_Int32 is not as clean as using sal_Bool directly

 Yes, except again, just == will do.

> Do you think that such a clean-up is interesting ?

 Yes, if it's interesting to you to do it :).

-- 
 Lubos Lunak
 l.lunak@suse.cz