unix/coverage: Expand mp_printf coverage tests.

Test 'l' and 'll' sized objects.  When the platform's `mp_int_t` is not 64
bits, dummy values are printed instead so the test result can match across
all platforms.

Ensure hex test values have a letter so 'x' vs 'X' is tested.

And test 'p' and 'P' pointer printing.

Signed-off-by: Jeff Epler <jepler@gmail.com>

ports/unix/coverage.c

@@ -204,8 +204,20 @@ static mp_obj_t extra_coverage(void) {
         mp_printf(&mp_plat_print, "%d %+d % d\n", -123, 123, 123); // sign
         mp_printf(&mp_plat_print, "%05d\n", -123); // negative number with zero padding
         mp_printf(&mp_plat_print, "%ld\n", 123); // long
-        mp_printf(&mp_plat_print, "%lx\n", 0x123); // long hex
-        mp_printf(&mp_plat_print, "%X\n", 0x1abcdef); // capital hex
+        mp_printf(&mp_plat_print, "%lx\n", 0x123fl); // long hex
+        mp_printf(&mp_plat_print, "%lX\n", 0x123fl); // capital long hex
+        if (sizeof(mp_int_t) == 8) {
+            mp_printf(&mp_plat_print, "%llx\n", LLONG_MAX); // long long hex
+            mp_printf(&mp_plat_print, "%llX\n", LLONG_MAX); // capital long long hex
+            mp_printf(&mp_plat_print, "%llu\n", ULLONG_MAX); // unsigned long long
+        } else {
+            // fake for platforms without narrower mp_int_t
+            mp_printf(&mp_plat_print, "7fffffffffffffff\n", LLONG_MAX);
+            mp_printf(&mp_plat_print, "7FFFFFFFFFFFFFFF\n", LLONG_MAX);
+            mp_printf(&mp_plat_print, "18446744073709551615\n", ULLONG_MAX);
+        }
+        mp_printf(&mp_plat_print, "%p\n", (void *)0x789f); // pointer
+        mp_printf(&mp_plat_print, "%P\n", (void *)0x789f); // pointer uppercase
         mp_printf(&mp_plat_print, "%.2s %.3s '%4.4s' '%5.5q' '%.3q'\n", "abc", "abc", "abc", MP_QSTR_True, MP_QSTR_True); // fixed string precision
         mp_printf(&mp_plat_print, "%.*s\n", -1, "abc"); // negative string precision
         mp_printf(&mp_plat_print, "%b %b\n", 0, 1); // bools
@@ -216,11 +228,31 @@ static mp_obj_t extra_coverage(void) {
         #endif
         mp_printf(&mp_plat_print, "%d\n", 0x80000000); // should print signed
         mp_printf(&mp_plat_print, "%u\n", 0x80000000); // should print unsigned
-        mp_printf(&mp_plat_print, "%x\n", 0x80000000); // should print unsigned
-        mp_printf(&mp_plat_print, "%X\n", 0x80000000); // should print unsigned
+        mp_printf(&mp_plat_print, "%x\n", 0x8000000f); // should print unsigned
+        mp_printf(&mp_plat_print, "%X\n", 0x8000000f); // should print unsigned
         mp_printf(&mp_plat_print, "abc\n%"); // string ends in middle of format specifier
         mp_printf(&mp_plat_print, "%%\n"); // literal % character
         mp_printf(&mp_plat_print, ".%-3s.\n", "a"); // left adjust
+
+        // Check that all kinds of mp_printf arguments are parsed out
+        // correctly, by having a char argument before and after each main type
+        // of value that can be formatted.
+        mp_printf(&mp_plat_print, "%c%%%c\n", '<', '>');
+        mp_printf(&mp_plat_print, "%c%p%c\n", '<', (void *)0xaaaa, '>');
+        mp_printf(&mp_plat_print, "%c%b%c\n", '<', true, '>');
+        mp_printf(&mp_plat_print, "%c%d%c\n", '<', 0xaaaa, '>');
+        mp_printf(&mp_plat_print, "%c%ld%c\n", '<', 0xaaaal, '>');
+        mp_printf(&mp_plat_print, "%c" INT_FMT "%c\n", '<', (mp_int_t)0xaaaa, '>');
+        mp_printf(&mp_plat_print, "%c%s%c\n", '<', "test", '>');
+        mp_printf(&mp_plat_print, "%c%f%c\n", '<', MICROPY_FLOAT_CONST(1000.), '>');
+        mp_printf(&mp_plat_print, "%c%q%c\n", '<', (qstr)MP_QSTR_True, '>');
+        if (sizeof(mp_int_t) == 8) {
+            mp_printf(&mp_plat_print, "%c%lld%c\n", '<', LLONG_MAX, '>');
+        } else {
+            mp_printf(&mp_plat_print, "<9223372036854775807>\n");
+        }
+
+
     }
 
     // GC

py/mpprint.c

@@ -446,16 +446,36 @@ int mp_vprintf(const mp_print_t *print, const char *fmt, va_list args) {
             }
         }
 
-        // parse long specifiers (only for LP64 model where they make a difference)
-        #ifndef __LP64__
-        const
+        // parse long and long long specifiers (only where they make a difference)
+        #if defined(MICROPY_UNIX_COVERAGE) || (LONG_MAX > INT_MAX)
+        #define SUPPORT_L_FORMAT (1)
+        #else
+        #define SUPPORT_L_FORMAT (0)
         #endif
+        #if SUPPORT_L_FORMAT
         bool long_arg = false;
+        #endif
+
+        #if (MICROPY_OBJ_REPR == MICROPY_OBJ_REPR_D) || defined(_WIN64) || defined(MICROPY_UNIX_COVERAGE)
+        #define SUPPORT_LL_FORMAT (1)
+        #else
+        #define SUPPORT_LL_FORMAT (0)
+        #endif
+        #if SUPPORT_LL_FORMAT
+        bool long_long_arg = false;
+        #endif
+
         if (*fmt == 'l') {
             ++fmt;
-            #ifdef __LP64__
+            #if SUPPORT_L_FORMAT
             long_arg = true;
             #endif
+            #if SUPPORT_LL_FORMAT
+            if (*fmt == 'l') {
+                ++fmt;
+                long_long_arg = true;
+            }
+            #endif
         }
 
         if (*fmt == '\0') {
@@ -501,35 +521,50 @@ int mp_vprintf(const mp_print_t *print, const char *fmt, va_list args) {
                 chrs += mp_print_strn(print, str, len, flags, fill, width);
                 break;
             }
-            case 'd': {
-                mp_int_t val;
-                if (long_arg) {
-                    val = va_arg(args, long int);
-                } else {
-                    val = va_arg(args, int);
-                }
-                chrs += mp_print_int(print, val, 1, 10, 'a', flags, fill, width);
-                break;
-            }
+            case 'd':
+            case 'p':
+            case 'P':
             case 'u':
             case 'x':
             case 'X': {
-                int base = 16 - ((*fmt + 1) & 6); // maps char u/x/X to base 10/16/16
-                char fmt_c = (*fmt & 0xf0) - 'P' + 'A'; // maps char u/x/X to char a/a/A
+                char fmt_chr = *fmt;
                 mp_uint_t val;
-                if (long_arg) {
-                    val = va_arg(args, unsigned long int);
-                } else {
-                    val = va_arg(args, unsigned int);
+                if (fmt_chr == 'p' || fmt_chr == 'P') {
+                    val = va_arg(args, intptr_t);
                 }
-                chrs += mp_print_int(print, val, 0, base, fmt_c, flags, fill, width);
+                #if SUPPORT_LL_FORMAT
+                else if (long_long_arg) {
+                    val = va_arg(args, unsigned long long);
+                }
+                #endif
+                #if SUPPORT_L_FORMAT
+                else if (long_arg) {
+                    if (sizeof(long) != sizeof(mp_uint_t) && fmt_chr == 'd') {
+                        val = va_arg(args, long);
+                    } else {
+                        val = va_arg(args, unsigned long);
+                    }
+                }
+                #endif
+                else {
+                    if (sizeof(int) != sizeof(mp_uint_t) && fmt_chr == 'd') {
+                        val = va_arg(args, int);
+                    } else {
+                        val = va_arg(args, unsigned);
+                    }
+                }
+                int base;
+                // Map format char x/p/X/P to a/a/A/A for hex letters.
+                // It doesn't matter what d/u map to.
+                char fmt_c = (fmt_chr & 0xf0) - 'P' + 'A';
+                if (fmt_chr == 'd' || fmt_chr == 'u') {
+                    base = 10;
+                } else {
+                    base = 16;
+                }
+                chrs += mp_print_int(print, val, fmt_chr == 'd', base, fmt_c, flags, fill, width);
                 break;
             }
-            case 'p':
-            case 'P': // don't bother to handle upcase for 'P'
-                // Use unsigned long int to work on both ILP32 and LP64 systems
-                chrs += mp_print_int(print, va_arg(args, unsigned long int), 0, 16, 'a', flags, fill, width);
-                break;
             #if MICROPY_PY_BUILTINS_FLOAT
             case 'e':
             case 'E':
@@ -545,18 +580,6 @@ int mp_vprintf(const mp_print_t *print, const char *fmt, va_list args) {
                 #endif
                 break;
             }
-            #endif
-                // Because 'l' is eaten above, another 'l' means %ll.  We need to support
-                // this length specifier for OBJ_REPR_D (64-bit NaN boxing).
-                // TODO Either enable this unconditionally, or provide a specific config var.
-            #if (MICROPY_OBJ_REPR == MICROPY_OBJ_REPR_D) || defined(_WIN64)
-            case 'l': {
-                unsigned long long int arg_value = va_arg(args, unsigned long long int);
-                ++fmt;
-                assert(*fmt == 'u' || *fmt == 'd' || !"unsupported fmt char");
-                chrs += mp_print_int(print, arg_value, *fmt == 'd', 10, 'a', flags, fill, width);
-                break;
-            }
             #endif
             default:
                 // if it's not %% then it's an unsupported format character

tests/ports/unix/extra_coverage.py.exp

@@ -2,19 +2,34 @@
 -123 +123  123
 -0123
 123
-123
-1ABCDEF
+123f
+123F
+7fffffffffffffff
+7FFFFFFFFFFFFFFF
+18446744073709551615
+789f
+789F
 ab abc ' abc' ' True' 'Tru'
 
 false true
 (null)
 -2147483648
 2147483648
-80000000
-80000000
+8000000f
+8000000F
 abc
 %
 .a  .
+<%>
+<aaaa>
+<true>
+<43690>
+<43690>
+<43690>
+<test>
+<1000.000000>
+<True>
+<9223372036854775807>
 # GC
 0
 0