Oops, good catch on the decimal error. Your first fix was the correct one. The trailing null is added when the new string instance is created, so we don't need to pass that. The length form is the correct one.
Rick On Wed, Apr 1, 2009 at 1:43 PM, Mark Miesfeld <[email protected]> wrote: > Rick, > > This seems the proper patch to me. (It also seems we have a bug there > with removing a trailing decimal.) The test suite runs without > errors, but I've only run it on Windows: > > Index: interpreter/classes/StringClass.cpp > =================================================================== > --- interpreter/classes/StringClass.cpp (revision 4329) > +++ interpreter/classes/StringClass.cpp (working copy) > @@ -2009,14 +2009,14 @@ > { > char buffer[64]; > // format as a string > - gcvt(number, (int)precision, buffer); > + sprintf(buffer, "%.*f", (int)precision, number); > size_t len = strlen(buffer); > // if the last character is a decimal, we remove that > if (buffer[len - 1] == '.') > { > len--; > } > - return new_string(buffer); > + return new_string(buffer, len); > } > } > > Since we know the length, new_string(buffer, len) is better than > explicitly putting '\0' at len -1 isn't it? Or does that end up with > the string not having a trailing null when it should have? Should it > be: > > // if the last character is a decimal, we remove that > if (buffer[len - 1] == '.') > { > buffer[len - 1] = '\0'; > } > return new_string(buffer); > > instead? > > -- > Mark Miesfeld > > ------------------------------------------------------------------------------ > _______________________________________________ > Oorexx-devel mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/oorexx-devel > ------------------------------------------------------------------------------ _______________________________________________ Oorexx-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/oorexx-devel
