On 14.07.2012 23:05, Nikita Egorov wrote: >> Reopened on request of OP. >> >> It seems clear to me that the attempted fix in r 9363 doesn't solve the >> OP's problem, since it would preserve *negative* char (int) values. >> >> Using both casts looks appropriate to make sure char codes > 127 are >> positive int's. >> >> Nikita, could you please confirm that svn r9635 solves the problem? >> Leaving the STR open for confirmation... > > Hi, Albrecht. At the moment fluid works very well.
Thanks for confirmation. > But I don't understand why you had not removed the (int). isspace((unsigned > char)*n)) is enough! (1) to be absolutely sure that NO compiler would issue a warning. (2) because I don't know why EXACTLY Fabien changed the existing (unsigned char) cast to (int). (3) to document that we REALLY want both casts in sequence. > Compilers (I tested VC++, mingw-gcc, ubuntu-gcc) cast to type int in any case > and they print no warning because it's absolutely safe. We only say to > compiler: "Don't extend the sign bit when you will convert our character to > int". Yes, I also think that it would have been enough, but see above. BTW, since this is only a compiler information, this additional cast doesn't add any runtime (performance) penalty, so I think that it is okay anyway. I'm going to close the STR now... -- Albrecht _______________________________________________ fltk-bugs mailing list fltk-bugs@easysw.com http://lists.easysw.com/mailman/listinfo/fltk-bugs