diff options
| author | Roberto I <roberto@inf.puc-rio.br> | 2025-08-09 15:08:53 -0300 |
|---|---|---|
| committer | Roberto I <roberto@inf.puc-rio.br> | 2025-08-09 15:08:53 -0300 |
| commit | 5b179eaf6a78af5f000d76147af94669d04487b2 (patch) | |
| tree | a3466ed22def25294ff1647d1af91cd4ef7fa567 /ltable.c | |
| parent | 8fddca81e7d4137512e92f398ca638d61b935dbd (diff) | |
| download | lua-5b179eaf6a78af5f000d76147af94669d04487b2.tar.gz lua-5b179eaf6a78af5f000d76147af94669d04487b2.tar.bz2 lua-5b179eaf6a78af5f000d76147af94669d04487b2.zip | |
Details
Diffstat (limited to 'ltable.c')
| -rw-r--r-- | ltable.c | 2 |
1 files changed, 1 insertions, 1 deletions
| @@ -156,7 +156,7 @@ static Node *hashint (const Table *t, lua_Integer i) { | |||
| 156 | ** The main computation should be just | 156 | ** The main computation should be just |
| 157 | ** n = frexp(n, &i); return (n * INT_MAX) + i | 157 | ** n = frexp(n, &i); return (n * INT_MAX) + i |
| 158 | ** but there are some numerical subtleties. | 158 | ** but there are some numerical subtleties. |
| 159 | ** In a two-complement representation, INT_MAX does not has an exact | 159 | ** In a two-complement representation, INT_MAX may not have an exact |
| 160 | ** representation as a float, but INT_MIN does; because the absolute | 160 | ** representation as a float, but INT_MIN does; because the absolute |
| 161 | ** value of 'frexp' is smaller than 1 (unless 'n' is inf/NaN), the | 161 | ** value of 'frexp' is smaller than 1 (unless 'n' is inf/NaN), the |
| 162 | ** absolute value of the product 'frexp * -INT_MIN' is smaller or equal | 162 | ** absolute value of the product 'frexp * -INT_MIN' is smaller or equal |