Bug 51 – String cast overrides the char type of decorated string literals.

Observe the following code: char[6] cstr = "123456"c; auto str1 = cast(wchar[3])(cstr); auto str2 = cast(wchar[3])("123456"c); writefln("str1: ", (cast(char[])str1).length , " : ", (cast(char[])str1)); // prints: str1: 6 : 123456 writefln("str2: ", (cast(char[])str2).length , " : ", (cast(char[])str2)); // prints: str2: 6 : ?1?2?3 One would expect the same behaviour from str1 and str2, however the cast(wchar[3]) in str2 overrides the string decorator 'c', and makes the string literal "123456" (i.e. the string "instance" data) be of type wchar[6] instead of char[6].

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 [email protected] schrieb am 2006-03-14: > Observe the following code: > > char[6] cstr = "123456"c; > auto str1 = cast(wchar[3])(cstr); > auto str2 = cast(wchar[3])("123456"c); > > writefln("str1: ", (cast(char[])str1).length , " : ", (cast(char[])str1)); > // prints: str1: 6 : 123456 > writefln("str2: ", (cast(char[])str2).length , " : ", (cast(char[])str2)); > // prints: str2: 6 : ?1?2?3 > > One would expect the same behaviour from str1 and str2, however the > cast(wchar[3]) in str2 overrides the string decorator 'c', and makes the string > literal "123456" (i.e. the string "instance" data) be of type wchar[6] instead > of char[6]. This might seem confusing, but is the correct behaviour. http://www.digitalmars.com/d/arrays.html # The type of a string is determined by the semantic phase of # compilation. The type is one of: char[], wchar[], dchar[], and is # determined by implicit conversion rules. If there are two equally # applicable implicit conversions, the result is an error. To # disambiguate these cases, a cast is appropriate: # # cast(wchar [])"abc" // this is an array of wchar characters Thomas -----BEGIN PGP SIGNATURE----- iD8DBQFEHRi13w+/yD4P9tIRAmgpAJ9F/KNd1JSBTBOp1QME7RA6Lwja9wCfT7Gx SlkT9jiEQ1rxtIl/cc7wT7s= =aqz7 -----END PGP SIGNATURE-----

(In reply to comment #1) > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > This might seem confusing, but is the correct behaviour. > http://www.digitalmars.com/d/arrays.html > # The type of a string is determined by the semantic phase of > # compilation. The type is one of: char[], wchar[], dchar[], and is > # determined by implicit conversion rules. If there are two equally > # applicable implicit conversions, the result is an error. To > # disambiguate these cases, a cast is appropriate: > # > # cast(wchar [])"abc" // this is an array of wchar characters > Thomas > -----BEGIN PGP SIGNATURE----- > iD8DBQFEHRi13w+/yD4P9tIRAmgpAJ9F/KNd1JSBTBOp1QME7RA6Lwja9wCfT7Gx > SlkT9jiEQ1rxtIl/cc7wT7s= > =aqz7 > -----END PGP SIGNATURE----- That behaviour is true for undecorated string literals, however for decorated string literals (i.e. those with postfix 'c', 'd', or 'w') it is not so. In http://www.digitalmars.com/d/lex.html#stringliteral it is said: "The optional Postfix character gives a specific type to the string, rather than it being inferred from the context. ..." Thus postfix-decorated string literals should not be vulnerable to type change due to casts or other context influences, no?

Fixed DMD 0.161 according to Bruno's analysis.

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 [email protected] schrieb am 2006-03-14: > http://d.puremagic.com/bugzilla/show_bug.cgi?id=51 > Observe the following code: > > char[6] cstr = "123456"c; > auto str1 = cast(wchar[3])(cstr); > auto str2 = cast(wchar[3])("123456"c); > > writefln("str1: ", (cast(char[])str1).length , " : ", (cast(char[])str1)); > // prints: str1: 6 : 123456 > writefln("str2: ", (cast(char[])str2).length , " : ", (cast(char[])str2)); > // prints: str2: 6 : ?1?2?3 > > One would expect the same behaviour from str1 and str2, however the > cast(wchar[3]) in str2 overrides the string decorator 'c', and makes the string > literal "123456" (i.e. the string "instance" data) be of type wchar[6] instead > of char[6]. Added to DStress as http://dstress.kuehne.cn/run/s/string_postfix_07_A.d http://dstress.kuehne.cn/run/s/string_postfix_07_B.d Thomas -----BEGIN PGP SIGNATURE----- iD8DBQFEtOHcLK5blCcjpWoRAgakAKCOhY7VG1AkTF8IFOz1Ur067Dk+dQCeIlbj h9EvDdFmTjEL3U2/pdpgnqk= =rrCR -----END PGP SIGNATURE-----