Dear Eike,

In message "Re: [porting-dev] MinGW port and ICU patches",
Eike Rathke wrote...

 >Do you know off-hand or could give me a pointer where assigning UChar*
 >to sal_Unicode* or vice versa was done? I consider relying on such
 >implementation detail a misuse ... if it wasn't for very good reason.
 >

In both line #185 and 186 in 
i18npool/source/breakiterator/breakiterator_unicode.cxx,
constructor of UnicodeString, which requires const UChar* as a 1st argument, is 
called with const sal_Unicode* string.

I may be possible to implement a switch in icu configure script to force using 
unsigned short instead of wchar_t and propose it upstream.

 >And I want to have less patches against ICU ;-)
 >
 >Having conditionals or macros in libs.mk and packaging would probably be
 >a one-time effort with little maintenance needed, while platform patches
 >against ICU would had to be reapplied and maintained for every ICU
 >upgrade. Given the awful unstructured all-in-one patch bulk I consider
 >having to handle that much worse than any "I don't like conditionals"
 >comment.

I think your coment makes sense and will not implement more modifications in 
patches 
for 3rd party modules to have less conditionals elsewhere. Probably I can make 
the 
patch for icu a bit smaller in the future with the strategy.

Takashi Ono ([EMAIL PROTECTED])

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to