On 02.05.2014 07:45, LacaK wrote: > Hi, > I am trying understand how FPC determines which version of overloaded > function to call. > Take as example these 2 functions (from SysUtils): > 1. function StrPLCopy(Dest: PChar; Source: string; MaxLen: SizeUInt): > PChar; overload; > 2. function StrPLCopy(Dest: PWideChar; const Source: UnicodeString; > MaxLen: SizeInt): PWideChar; overload; > > And in program I use: > var p: pointer; > s: string; > begin > StrPLCopy(p, s, 10); > end; > > It seems, that 2nd "unicode" version of StrPLCopy is called. But why? > I am thinking as: > 1st parameter is untyped pointer, so no one of function can be prefered > 2nd parameter is "string" which exactly match 1st function signature > 3rd parameter is "integer constant" which mach both signatures > > So it seems to me, that 1st version of function signature matches > better, because of exact match of 2nd parameter ... why is then called > 2nd version? I don't have a explanation for now, but it *might* be important to note whether you are using 2.6.x or 2.7.1. Regards, Sven _______________________________________________ fpc-pascal maillist - http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal

Hi again, I reply also to fpc-devel list (excuse me, if it is off-topic), because I would like understand wat is going on. I attach also simple test program. I use FPC 2.6.4 but I think, that same happens also with fresh 2.7.1 Thanks -Laco. > On 02.05.2014 07:45, LacaK wrote: >> Hi, >> I am trying understand how FPC determines which version of overloaded >> function to call. >> Take as example these 2 functions (from SysUtils): >> 1. function StrPLCopy(Dest: PChar; Source: string; MaxLen: SizeUInt): >> PChar; overload; >> 2. function StrPLCopy(Dest: PWideChar; const Source: UnicodeString; >> MaxLen: SizeInt): PWideChar; overload; >> >> And in program I use: >> var p: pointer; >> s: string; >> begin >> StrPLCopy(p, s, 10); >> end; >> >> It seems, that 2nd "unicode" version of StrPLCopy is called. But why? >> I am thinking as: >> 1st parameter is untyped pointer, so no one of function can be prefered >> 2nd parameter is "string" which exactly match 1st function signature >> 3rd parameter is "integer constant" which mach both signatures >> >> So it seems to me, that 1st version of function signature matches >> better, because of exact match of 2nd parameter ... why is then called >> 2nd version? > > I don't have a explanation for now, but it *might* be important to > note whether you are using 2.6.x or 2.7.1. > > Regards, > Sven > > _______________________________________________ > fpc-pascal maillist - > http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal >