stillwater-sc / universal / 30567691028

85%

master: 84%

push

github

fix(numeric): clear decimal.hpp narrowing + ignored-qualifier warnings (#1263) (#1269)

Three warnings in number/support/decimal.hpp (part of the warning-clean Epic
#1265), reachable widely because decimal.hpp is pulled in through the number-
system aggregators:

1. add() digit accumulation narrowed int -> uint8_t (-Wconversion): `*lit += *rit
   + carry` promotes to int and the compound assignment narrows back. Made the
   truncation explicit (`*lit = static_cast<uint8_t>(*lit + *rit + carry)`), and
   changed `char carry` to `uint8_t carry` -- it only ever holds 0 or 1, and char's
   implementation-defined signedness meant signed/unsigned mixing in digit
   arithmetic on signed-char platforms. This also makes add() consistent with
   sub()'s `uint8_t borrow` and mul()'s `uint8_t carry`.
2/3. mul() had `static_cast<const int64_t>(position)` at two sites
   (-Wignored-qualifiers): top-level const on a cast to a non-class type is
   meaningless. Dropped it.

Value-preserving (the casts state existing truncations; uint8_t carry holds the
same 0/1). gcc + clang: decimal.hpp is now clean under -Wconversion /
-Wignored-qualifiers; conversions_to_decimal, bt_decimal, fixpnt_mod_conversion
PASS on both.

Resolves #1263

4 of 4 new or added lines in 1 file covered. (100.0%)

41661 of 48877 relevant lines covered (85.24%)

7272861.96 hits per line