test: Handle 32-bit double in test-efcvt

Skip tests which require more than 32-bit float precision in this
configuration.

Signed-off-by: Keith Packard <[email protected]>

Author: keith-packard

test/test-efcvt.c

@@ -155,6 +155,16 @@ const struct test fcvtf_tests[] = {
 #define SKIP_LONG_FLOAT 0
 #endif
 
+#if (!defined(TINY_STDIO) || !defined(_IO_FLOAT_EXACT)) && __SIZEOF_DOUBLE__ < 8
+#undef SKIP_LONG_FLOAT
+#define SKIP_LONG_FLOAT 1
+#define SKIP_LONGISH_FLOAT 1
+#endif
+
+#ifndef SKIP_LONGISH_FLOAT
+#define SKIP_LONGISH_FLOAT 0
+#endif
+
 #define many_tests(tests, func, n, skip_long) do {                      \
     for (i = 0; i < n; i++) {                                           \
         int decpt;                                                      \
@@ -188,13 +198,13 @@ main(void)
     unsigned i;
     static char buf[2048];
 
-    many_tests(fcvt_tests, fcvt_r, N_FCVT_TESTS, 0);
+    many_tests(fcvt_tests, fcvt_r, N_FCVT_TESTS, SKIP_LONGISH_FLOAT);
     many_tests(fcvt_extra_tests, fcvt_r, N_FCVT_EXTRA_TESTS, SKIP_LONG_FLOAT);
-    many_tests(ecvt_tests, ecvt_r, N_ECVT_TESTS, 0);
+    many_tests(ecvt_tests, ecvt_r, N_ECVT_TESTS, SKIP_LONGISH_FLOAT);
 #ifdef __PICOLIBC__
     many_tests(fcvt_tests, fcvtf_r, N_FCVT_TESTS, SKIP_LONG_FLOAT);
     many_tests(fcvtf_tests, fcvtf_r, N_FCVTF_TESTS, SKIP_LONG_FLOAT);
-    many_tests(ecvt_tests, ecvtf_r, N_ECVT_TESTS, 0);
+    many_tests(ecvt_tests, ecvtf_r, N_ECVT_TESTS, SKIP_LONGISH_FLOAT);
 #endif
     return error;
 }