RepoMirrors
/
mpv
mirror of https://github.com/mpv-player/mpv
1
0
Fork 0
Code Issues Releases Wiki Activity

m_option: attempt to fix two rounding issues 

Since double has a mantissa too small to hold INT64_MAX in full
precision, converting INT64_MAX to double rounds up. Insert some casts
to silence corresponding warnings (as shown by clang 11).

Also, the comparison in multiply_int64() was incorrect (I think...),
because if v==(double)INT64_MAX, then v==(1<<64), which cannot be
represented as int64_t.

There are probably better ways to solve this.
This commit is contained in:
wm4 2020-03-18 20:19:13 +01:00
parent 26f4f18c06
commit 5a81de59a8
1 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
options/m_option.c
View File

@ -56,10 +56,10 @@ const char m_option_path_separator = OPTION_PATH_SEPARATOR;
// For integer types: since min/max are floats and may not be able to represent
// the real min/max, and since opt.min/.max may use +/-INFINITY, some care has
// to be taken. (Also tricky rounding.)
#define OPT_INT_MIN(opt, T, Tm) \
((opt)->min < (opt)->max ? ((opt)->min <= (Tm) ? (Tm) : (T)((opt)->min)) : (Tm))
#define OPT_INT_MAX(opt, T, Tm) \
((opt)->min < (opt)->max ? ((opt)->max >= (Tm) ? (Tm) : (T)((opt)->max)) : (Tm))
#define OPT_INT_MIN(opt, T, Tm) ((opt)->min < (opt)->max \
? ((opt)->min <= (double)(Tm) ? (Tm) : (T)((opt)->min)) : (Tm))
#define OPT_INT_MAX(opt, T, Tm) ((opt)->min < (opt)->max \
? ((opt)->max >= (double)(Tm) ? (Tm) : (T)((opt)->max)) : (Tm))
char *m_option_strerror(int code)
{
@ -390,7 +390,7 @@ static void multiply_int64(const m_option_t *opt, void *val, double f)
int64_t iv = v;
if (v < INT64_MIN)
iv = INT64_MIN;
if (v > INT64_MAX)
if (v >= (double)INT64_MAX)
iv = INT64_MAX;
*(int64_t *)val = iv;
clamp_int64(opt, val);