maint: build fix for ucptest and related code tyding up

Since 9a868b0 (Tidy up config.h management (task from README) (#658),
2025-01-11), it fails to build with an "#error" because "config.h" is
no longer included directly.

Move the check and change some types in the utf8 tables it uses through
pcre2_ord2utf() so it better matches its current use.

Author: carenas

maint/ucptest.c

@@ -82,17 +82,17 @@ characters, the list ends with ...
 The command "list" must be followed by one of property names script, bool,
 type, gbreak or bidi. The defined values for that property are listed. */
 
-#ifndef SUPPORT_UNICODE
-#error "Unicode support not enabled"
-#endif
-
 #include <ctype.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include "../src/pcre2_internal.h"
 #include "../src/pcre2_ucp.h"
 
+#ifndef SUPPORT_UNICODE
+#error "Unicode support not enabled"
+#endif
+
 #ifdef HAVE_UNISTD_H
 #include <unistd.h>
 #endif

src/pcre2_internal.h

@@ -2233,7 +2233,7 @@ tables are needed only when compiling the 8-bit library. */
 
 #if PCRE2_CODE_UNIT_WIDTH == 8
 extern const int              PRIV(utf8_table1)[];
-extern const int              PRIV(utf8_table1_size);
+extern const unsigned         PRIV(utf8_table1_size);
 extern const int              PRIV(utf8_table2)[];
 extern const int              PRIV(utf8_table3)[];
 extern const uint8_t          PRIV(utf8_table4)[];

src/pcre2_ord2utf.c

@@ -80,16 +80,17 @@ PRIV(ord2utf)(uint32_t cvalue, PCRE2_UCHAR *buffer)
 /* Convert to UTF-8 */
 
 #if PCRE2_CODE_UNIT_WIDTH == 8
-int i, j;
+unsigned int i;
+
 for (i = 0; i < PRIV(utf8_table1_size); i++)
   if ((int)cvalue <= PRIV(utf8_table1)[i]) break;
 buffer += i;
-for (j = i; j > 0; j--)
+for (unsigned int j = i; j != 0; j--)
  {
  *buffer-- = 0x80 | (cvalue & 0x3f);
  cvalue >>= 6;
  }
-*buffer = PRIV(utf8_table2)[i] | cvalue;
+*buffer = (PCRE2_UCHAR)(PRIV(utf8_table2)[i] | (int)cvalue);
 return i + 1;
 
 /* Convert to UTF-16 */

src/pcre2_tables.c

@@ -51,6 +51,8 @@ defined. */
 #endif
 
 
+/* Utility macros */
+#define ARR_SIZE(x) sizeof(x)/sizeof(x[0])
 
 /* Table of sizes for the fixed-length opcodes. It's defined in a macro so that
 the definition is next to the definition of the opcodes in pcre2_internal.h.
@@ -100,15 +102,15 @@ handling wide characters. */
 character. */
 
 const int PRIV(utf8_table1)[] =
-  { 0x7f, 0x7ff, 0xffff, 0x1fffff, 0x3ffffff, 0x7fffffff};
+  { 0x7f, 0x7ff, 0xffff, 0x1fffff, 0x3ffffff, 0x7fffffff };
 
-const int PRIV(utf8_table1_size) = sizeof(PRIV(utf8_table1)) / sizeof(int);
+const unsigned PRIV(utf8_table1_size) = ARR_SIZE(PRIV(utf8_table1));
 
 /* These are the indicator bits and the mask for the data bits to set in the
 first byte of a character, indexed by the number of additional bytes. */
 
-const int PRIV(utf8_table2)[] = { 0,    0xc0, 0xe0, 0xf0, 0xf8, 0xfc};
-const int PRIV(utf8_table3)[] = { 0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01};
+const int PRIV(utf8_table2)[] = { 0,    0xc0, 0xe0, 0xf0, 0xf8, 0xfc };
+const int PRIV(utf8_table3)[] = { 0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01 };
 
 /* Table of the number of extra bytes, indexed by the first byte masked with
 0x3f. The highest number for a valid UTF-8 first byte is in fact 0x3d. */

src/pcre2test.c

@@ -3236,7 +3236,8 @@ utf8_to_ord(PCRE2_SPTR8 utf8bytes, PCRE2_SPTR8 end, uint32_t *vptr)
 {
 uint32_t c = *utf8bytes++;
 uint32_t d = c;
-int i, j, s;
+unsigned j;
+int i, s;
 
 for (i = -1; i < 6; i++)               /* i is number of additional bytes */
   {
@@ -3252,7 +3253,7 @@ if (i == 0 || i == 6) return 0;        /* invalid UTF-8 */
 s = 6*i;
 d = (c & utf8_table3[i]) << s;
 
-for (j = 0; j < i; j++)
+for (j = 0; (int)j < i; j++)
   {
   if (utf8bytes >= end) return 0;
 
@@ -3266,7 +3267,7 @@ for (j = 0; j < i; j++)
 
 for (j = 0; j < utf8_table1_size; j++)
   if (d <= (uint32_t)utf8_table1[j]) break;
-if (j != i) return -(i+1);
+if ((int)j != i) return -(i+1);
 
 /* Valid value */
 
@@ -3340,7 +3341,7 @@ ord_to_utf8(uint32_t cvalue, uint8_t *utf8bytes)
 int i, j;
 if (cvalue > 0x7fffffffu)
   return -1;
-for (i = 0; i < utf8_table1_size; i++)
+for (i = 0; i < (int)utf8_table1_size; i++)
   if (cvalue <= (uint32_t)utf8_table1[i]) break;
 utf8bytes += i;
 for (j = i; j > 0; j--)
@@ -3746,6 +3747,7 @@ if (!utf && (pat_patctl.control & CTL_UTF8_INPUT) == 0)
   {
   for (; len > 0; len--) *pp++ = *p++;
   }
+
 else while (len > 0)
   {
   uint32_t c;
@@ -3830,7 +3832,6 @@ if (pbuffer32_size < 4*len + 4)
   }
 
 pp = pbuffer32;
-
 if (!utf && (pat_patctl.control & CTL_UTF8_INPUT) == 0)
   {
   for (; len > 0; len--) *pp++ = *p++;
@@ -8189,15 +8190,22 @@ while ((c = *p++) != 0)
 
   if (c != '\\' || subject_literal)
     {
+#ifdef SUPPORT_PCRE2_32
     uint32_t topbit = 0;
     if (test_mode == PCRE32_MODE && c == 0xff && *p != 0)
       {
       topbit = 0x80000000;
       c = *p++;
       }
+#endif
     if ((utf || (pat_patctl.control & CTL_UTF8_INPUT) != 0) &&
-      HASUTF8EXTRALEN(c)) { GETUTF8INC(c, p); }
+        HASUTF8EXTRALEN(c))
+      {
+      GETUTF8INC(c, p);
+      }
+#ifdef SUPPORT_PCRE2_32
     c |= topbit;
+#endif
     }
 
   /* Handle backslash escapes */