Ticket #880 (closed defect: fixed)
~F format directive incorrectly uses exponential form when "w" and "d" are omitted.
Reported by: | aidenn0 | Owned by: | rme |
---|---|---|---|
Priority: | normal | Milestone: | |
Component: | ANSI CL Compliance | Version: | 1.7 |
Keywords: | Cc: |
Description
From Section 22.3.3.1
If both w and d are omitted, then the effect is to print the value using ordinary free-format output; prin1 uses this format for any number whose magnitude is either zero or between 10^{-3 (inclusive) and 10}7 (exclusive).
free-format is defined in this section to be the same format used by prin1 for the range 10^{-3} through 10^{7}. This is covered in section 22.1.3.1.3:
If the magnitude of the float is either zero or between 10^{-3} (inclusive) and 10^{7} (exclusive), it is printed as the integer part of the number, then a decimal point, followed by the fractional part of the number; there is always at least one digit on each side of the decimal point. If the sign of the number (as determined by float-sign) is negative, then a minus sign is printed before the number. If the format of the number does not match that specified by *read-default-float-format*, then the exponent marker for that format and the digit 0 are also printed. For example, the base of the natural logarithms as a short float might be printed as 2.71828S0.
22.3.3.1 does go on to say that if the width would be greater than 100, exponential notation may be used. However, it follows from the specification that (for example)
(format t "~F" 1e7)
should print:
10000000.0
Not:
1.0E+7
Change History
comment:2 Changed 18 months ago by rme
- Status changed from new to closed
- Resolution set to fixed
(In [16338]) Pull out the code that prints a free-form floating-point number from print-a-float, and put it into a new helper function print-float-free-form-1.
New function print-float-free-form, which uses the above helper function.
Make format-fixed-aux (which handles "~f") call print-float-free-form instead of prin1 when the w and d parameters are null.
Closes ticket:880.
For a long time, I didn't understand this. I was interpreting 22.3.3.1 to mean that then when both the w and d parameters are omitted, "~f" means to print the value like prin1 does. Since prin1 uses computerized scientific notation for numbers 1e7 and greater, I didn't see a problem.
But I think now that the spec is using the output of prin1 on numbers in [1e-3, 1e7) as an example of "ordinary free-form output". So, I now (finally) agree that (format t "~f" 1e7) should print 10000000.0.