RFC: 6178739 - Formatter - Zero padding flag with zero width (original) (raw)

Brian Burkhalter brian.burkhalter at oracle.com
Thu Jun 27 18:27:20 UTC 2013

On Jun 26, 2013, at 10:55 PM, David DeHaven wrote:

Specifically, I was referred to how C handles "%0.4f\n". No width, decimal truncated (rounded? floored? I forgot which it is) to four digits. -DrD-

printf("%0.4f\n", 56789.456789F); ... 56789.4570 ^ ^ ^ ^ ^ ^ ^ ^ ... "A leading zero in the width value is interpreted as the zero-padding flag mentioned above […]." Only if there's a valid width following, which there isn't in the case above. Try "%016.4" with the above test. Note that the width is the full width of the entire field, including decimal point and following digits. printf("%016.4f\n", 56789.456789F); printf("%0.4f\n", 56789.456789F); Produces: 00000056789.4570 56789.4570

printf("%016.4f\n", 56789.456789F);
printf("%1.4f\n", 56789.456789F);
printf("%0.4f\n", 56789.456789F);

produces (on Mac OS X)

00000056789.4570 56789.4570 56789.4570

whereas

    System.out.printf("%016.4f\n", 56789.456789);
    System.out.printf("%1.4f\n", 56789.456789);
    System.out.printf("%0.4f\n", 56789.456789);

produces

00000056789.4568 56789.4568 Exception in thread "main" java.util.MissingFormatWidthException: %0.4f

Two possible solutions are:

  1. Change the specification of the width field to

"The optional width is a positive decimal integer indicating the minimum number of characters to be written to the output. A leading zero in the width value is interpreted to be the zero-padding flag discussed below."

  1. Handle a 0.x as indicating zero width, not zero padding.

Brian

More information about the core-libs-dev mailing list