Lines Matching defs:digits

32 //  not based on the Ryu algorithm, instead generating the digits by
36 //  of the digits above the decimal point, making it inefficient for %e calls
51 //  results for the first ~50 digits of the output. In practice this shouldn't
52 //  be a problem since long doubles are only accurate for ~35 digits, but the
145   // us the floor, whereas counting the digits of the power of 2 gives us the
159 // digits a given e-bit number could have.
227   // for the first 50 digits, but it's much faster. Since even 128 bit long
228   // doubles are only accurate to ~35 digits, the 50 digits of accuracy are
337   // for the first 50 digits, but it's much faster. Since even 128 bit long
338   // doubles are only accurate to ~35 digits, the 50 digits of accuracy are
391 // request blocks of digits and convert them from integers to strings themself.
392 // Blocks contain the most digits that can be stored in an BlockInt. This is 9
393 // digits for a 32 bit int and 18 digits for a 64 bit int.
398 // zero. This will give all digits before the decimal point. Then the user can
399 // start calling get_negative_block in a loop from 0 until the number of digits
402 // be zero after the decimal point and before the non-zero digits. Additionally,
429   // get_block returns an integer that represents the digits in the requested
484       const BlockInt digits =
486       return digits;
545       BlockInt digits =
547       return digits;
572   // and return if it's sure that all of the digits after it are zero.
575     // The decimal representation of 2**(-i) will have exactly i digits after
585     // If the remaining digits are all 0, then this is the lowest block.
679       // generates the digits from right to left, but the digits will be written
706       // If there are still digits above the decimal point, handle those.
720         // Zero all digits above the decimal point.
770     // The decimal representation of 2**(-i) will have exactly i digits after