Require digit size to support values 0-99 #1013
Conversation
eemeli
added
functions
| Implementations MAY define an upper limit on the _resolved value_ | ||
| of a digit size option option consistent with that implementation's practical limits. | ||
| of a digit size option option consistent with that implementation's practical limits, | ||
| as long as all integer values from 0 to 99 (inclusive) are supported. |
There was a problem hiding this comment.
99 is way beyond the significant digits supported by an IEEE754 double. If we are going to require minimal values, it shouldn't be more than needed with double, which means it may vary according to the option.
There was a problem hiding this comment.
I agree that this isn't the right fix. I think what you're trying to say is maybe:
| as long as all integer values from 0 to 99 (inclusive) are supported. | |
| of a digit size option consistent with that implementation's practical limits. | |
| Implementations that do so MUST emit an _Unsupported Operation_ error. | |
| The literal values for integers from 0 to 99 (inclusive) MUST NOT produce | |
| a _Bad Option_ error. |
(Note that your text has "option option")
There was a problem hiding this comment.
My intent here was to clarify the general case, but arguably that does complicate matters; the current language allows more clearly for an implementation to define a different upper bound for each digit size option.
The "option option" text is currently in the spec; we should fix that separately if discarding this change.
|
closing per 2025-02-17 call discussion. typo to be fixed in 1007 |
Digit size option values MAY have an implementation-defined upper limit, but let's make sure that does not exclude anything allowed by the
digit-size-optionrule for literal values.