Skip to content

Commit

Permalink
pythongh-112301: Use literal format strings in unicode_fromformat_arg (
Browse files Browse the repository at this point in the history
  • Loading branch information
encukou committed Sep 25, 2024
1 parent 162d152 commit da5855e
Showing 1 changed file with 35 additions and 43 deletions.
78 changes: 35 additions & 43 deletions Objects/unicodeobject.c
Original file line number Diff line number Diff line change
Expand Up @@ -2694,11 +2694,6 @@ unicode_fromformat_write_wcstr(_PyUnicodeWriter *writer, const wchar_t *str,
#define F_SIZE 3
#define F_PTRDIFF 4
#define F_INTMAX 5
static const char * const formats[] = {"%d", "%ld", "%lld", "%zd", "%td", "%jd"};
static const char * const formats_o[] = {"%o", "%lo", "%llo", "%zo", "%to", "%jo"};
static const char * const formats_u[] = {"%u", "%lu", "%llu", "%zu", "%tu", "%ju"};
static const char * const formats_x[] = {"%x", "%lx", "%llx", "%zx", "%tx", "%jx"};
static const char * const formats_X[] = {"%X", "%lX", "%llX", "%zX", "%tX", "%jX"};

static const char*
unicode_fromformat_arg(_PyUnicodeWriter *writer,
Expand Down Expand Up @@ -2840,47 +2835,44 @@ unicode_fromformat_arg(_PyUnicodeWriter *writer,
case 'd': case 'i':
case 'o': case 'u': case 'x': case 'X':
{
/* used by sprintf */
char buffer[MAX_INTMAX_CHARS];
const char *fmt = NULL;
switch (*f) {
case 'o': fmt = formats_o[sizemod]; break;
case 'u': fmt = formats_u[sizemod]; break;
case 'x': fmt = formats_x[sizemod]; break;
case 'X': fmt = formats_X[sizemod]; break;
default: fmt = formats[sizemod]; break;
}
int issigned = (*f == 'd' || *f == 'i');

// Fill buffer using sprinf, with one of many possible format
// strings, like "%llX" for `long long` in hexadecimal.
// The type/size is in `sizemod`; the format is in `*f`.

// Use macros with nested switches to keep the sprintf format strings
// as compile-time literals, avoiding warnings and maybe allowing
// optimizations.

// `SPRINT` macro does one sprintf
// Example usage: SPRINT("l", "X", unsigned long) expands to
// sprintf(buffer, "%" "l" "X", va_arg(*vargs, unsigned long))
#define SPRINT(SIZE_SPEC, FMT_CHAR, TYPE) \
sprintf(buffer, "%" SIZE_SPEC FMT_CHAR, va_arg(*vargs, TYPE))

// One inner switch to handle all format variants
#define DO_SPRINTS(SIZE_SPEC, SIGNED_TYPE, UNSIGNED_TYPE) \
switch (*f) { \
case 'o': len = SPRINT(SIZE_SPEC, "o", UNSIGNED_TYPE); break; \
case 'u': len = SPRINT(SIZE_SPEC, "u", UNSIGNED_TYPE); break; \
case 'x': len = SPRINT(SIZE_SPEC, "x", UNSIGNED_TYPE); break; \
case 'X': len = SPRINT(SIZE_SPEC, "X", UNSIGNED_TYPE); break; \
default: len = SPRINT(SIZE_SPEC, "d", SIGNED_TYPE); break; \
}

// Outer switch to handle all the sizes/types
switch (sizemod) {
case F_LONG:
len = issigned ?
sprintf(buffer, fmt, va_arg(*vargs, long)) :
sprintf(buffer, fmt, va_arg(*vargs, unsigned long));
break;
case F_LONGLONG:
len = issigned ?
sprintf(buffer, fmt, va_arg(*vargs, long long)) :
sprintf(buffer, fmt, va_arg(*vargs, unsigned long long));
break;
case F_SIZE:
len = issigned ?
sprintf(buffer, fmt, va_arg(*vargs, Py_ssize_t)) :
sprintf(buffer, fmt, va_arg(*vargs, size_t));
break;
case F_PTRDIFF:
len = sprintf(buffer, fmt, va_arg(*vargs, ptrdiff_t));
break;
case F_INTMAX:
len = issigned ?
sprintf(buffer, fmt, va_arg(*vargs, intmax_t)) :
sprintf(buffer, fmt, va_arg(*vargs, uintmax_t));
break;
default:
len = issigned ?
sprintf(buffer, fmt, va_arg(*vargs, int)) :
sprintf(buffer, fmt, va_arg(*vargs, unsigned int));
break;
case F_LONG: DO_SPRINTS("l", long, unsigned long); break;
case F_LONGLONG: DO_SPRINTS("ll", long long, unsigned long long); break;
case F_SIZE: DO_SPRINTS("z", Py_ssize_t, size_t); break;
case F_PTRDIFF: DO_SPRINTS("t", ptrdiff_t, ptrdiff_t); break;
case F_INTMAX: DO_SPRINTS("j", intmax_t, uintmax_t); break;
default: DO_SPRINTS("", int, unsigned int); break;
}
#undef SPRINT
#undef DO_SPRINTS

assert(len >= 0);

int sign = (buffer[0] == '-');
Expand Down

0 comments on commit da5855e

Please sign in to comment.