[Rework] Update snowball stemmer to 2.0 and remove all crap aside of UTF8

4 files changed, 119 insertions, 95 deletions

diff --git a/contrib/snowball/runtime/api.c b/contrib/snowball/runtime/api.c
index 40039ef4a..21ea5f246 100644
--- a/contrib/snowball/runtime/api.c
+++ b/contrib/snowball/runtime/api.c
@@ -2,7 +2,7 @@
 #include <stdlib.h> /* for calloc, free */
 #include "header.h"
 
-extern struct SN_env * SN_create_env(int S_size, int I_size, int B_size)
+extern struct SN_env * SN_create_env(int S_size, int I_size)
 {
     struct SN_env * z = (struct SN_env *) calloc(1, sizeof(struct SN_env));
     if (z == NULL) return NULL;
@@ -27,12 +27,6 @@ extern struct SN_env * SN_create_env(int S_size, int I_size, int B_size)
         if (z->I == NULL) goto error;
     }
 
-    if (B_size)
-    {
-        z->B = (unsigned char *) calloc(B_size, sizeof(unsigned char));
-        if (z->B == NULL) goto error;
-    }
-
     return z;
 error:
     SN_close_env(z, S_size);
@@ -52,7 +46,6 @@ extern void SN_close_env(struct SN_env * z, int S_size)
         free(z->S);
     }
     free(z->I);
-    free(z->B);
     if (z->p) lose_s(z->p);
     free(z);
 }
@@ -63,4 +56,3 @@ extern int SN_set_current(struct SN_env * z, int size, const symbol * s)
     z->c = 0;
     return err;
 }
-
diff --git a/contrib/snowball/runtime/api.h b/contrib/snowball/runtime/api.h
index 8b997f0c2..ba9d1c14b 100644
--- a/contrib/snowball/runtime/api.h
+++ b/contrib/snowball/runtime/api.h
@@ -16,11 +16,17 @@ struct SN_env {
     int c; int l; int lb; int bra; int ket;
     symbol * * S;
     int * I;
-    unsigned char * B;
 };
 
-extern struct SN_env * SN_create_env(int S_size, int I_size, int B_size);
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+extern struct SN_env * SN_create_env(int S_size, int I_size);
 extern void SN_close_env(struct SN_env * z, int S_size);
 
 extern int SN_set_current(struct SN_env * z, int size, const symbol * s);
 
+#ifdef __cplusplus
+}
+#endif
diff --git a/contrib/snowball/runtime/header.h b/contrib/snowball/runtime/header.h
index 4d3078f50..85a42fdb8 100644
--- a/contrib/snowball/runtime/header.h
+++ b/contrib/snowball/runtime/header.h
@@ -54,5 +54,6 @@ extern int insert_v(struct SN_env * z, int bra, int ket, const symbol * p);
 extern symbol * slice_to(struct SN_env * z, symbol * p);
 extern symbol * assign_to(struct SN_env * z, symbol * p);
 
-extern void debug(struct SN_env * z, int number, int line_count);
+extern int len_utf8(const symbol * p);
 
+extern void debug(struct SN_env * z, int number, int line_count);
diff --git a/contrib/snowball/runtime/utilities.c b/contrib/snowball/runtime/utilities.c
index 1840f0280..1cfd1a17d 100644
--- a/contrib/snowball/runtime/utilities.c
+++ b/contrib/snowball/runtime/utilities.c
@@ -5,8 +5,6 @@
 
 #include "header.h"
 
-#define unless(C) if(!(C))
-
 #define CREATE_SIZE 1
 
 extern symbol * create_s(void) {
@@ -15,7 +13,7 @@ extern symbol * create_s(void) {
     if (mem == NULL) return NULL;
     p = (symbol *) (HEAD + (char *) mem);
     CAPACITY(p) = CREATE_SIZE;
-    SET_SIZE(p, CREATE_SIZE);
+    SET_SIZE(p, 0);
     return p;
 }
 
@@ -27,7 +25,7 @@ extern void lose_s(symbol * p) {
 /*
    new_p = skip_utf8(p, c, lb, l, n); skips n characters forwards from p + c
    if n +ve, or n characters backwards from p + c - 1 if n -ve. new_p is the new
-   position, or 0 on failure.
+   position, or -1 on failure.
 
    -- used to implement hop and next in the utf8 case.
 */
@@ -66,77 +64,95 @@ extern int skip_utf8(const symbol * p, int c, int lb, int l, int n) {
 /* Code for character groupings: utf8 cases */
 
 static int get_utf8(const symbol * p, int c, int l, int * slot) {
-    int b0, b1;
+    int b0, b1, b2;
     if (c >= l) return 0;
     b0 = p[c++];
     if (b0 < 0xC0 || c == l) {   /* 1100 0000 */
-        * slot = b0; return 1;
+        *slot = b0;
+        return 1;
     }
-    b1 = p[c++];
+    b1 = p[c++] & 0x3F;
     if (b0 < 0xE0 || c == l) {   /* 1110 0000 */
-        * slot = (b0 & 0x1F) << 6 | (b1 & 0x3F); return 2;
+        *slot = (b0 & 0x1F) << 6 | b1;
+        return 2;
+    }
+    b2 = p[c++] & 0x3F;
+    if (b0 < 0xF0 || c == l) {   /* 1111 0000 */
+        *slot = (b0 & 0xF) << 12 | b1 << 6 | b2;
+        return 3;
     }
-    * slot = (b0 & 0xF) << 12 | (b1 & 0x3F) << 6 | (p[c] & 0x3F); return 3;
+    *slot = (b0 & 0xE) << 18 | b1 << 12 | b2 << 6 | (p[c] & 0x3F);
+    return 4;
 }
 
 static int get_b_utf8(const symbol * p, int c, int lb, int * slot) {
-    int b0, b1;
+    int a, b;
     if (c <= lb) return 0;
-    b0 = p[--c];
-    if (b0 < 0x80 || c == lb) {   /* 1000 0000 */
-        * slot = b0; return 1;
+    b = p[--c];
+    if (b < 0x80 || c == lb) {   /* 1000 0000 */
+        *slot = b;
+        return 1;
     }
-    b1 = p[--c];
-    if (b1 >= 0xC0 || c == lb) {   /* 1100 0000 */
-        * slot = (b1 & 0x1F) << 6 | (b0 & 0x3F); return 2;
+    a = b & 0x3F;
+    b = p[--c];
+    if (b >= 0xC0 || c == lb) {   /* 1100 0000 */
+        *slot = (b & 0x1F) << 6 | a;
+        return 2;
     }
-    * slot = (p[c] & 0xF) << 12 | (b1 & 0x3F) << 6 | (b0 & 0x3F); return 3;
+    a |= (b & 0x3F) << 6;
+    b = p[--c];
+    if (b >= 0xE0 || c == lb) {   /* 1110 0000 */
+        *slot = (b & 0xF) << 12 | a;
+        return 3;
+    }
+    *slot = (p[--c] & 0xE) << 18 | (b & 0x3F) << 12 | a;
+    return 4;
 }
 
 extern int in_grouping_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) {
     do {
-	int ch;
-	int w = get_utf8(z->p, z->c, z->l, & ch);
-	unless (w) return -1;
-	if (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)
-	    return w;
-	z->c += w;
+        int ch;
+        int w = get_utf8(z->p, z->c, z->l, & ch);
+        if (!w) return -1;
+        if (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)
+            return w;
+        z->c += w;
     } while (repeat);
     return 0;
 }
 
 extern int in_grouping_b_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) {
     do {
-	int ch;
-	int w = get_b_utf8(z->p, z->c, z->lb, & ch);
-	unless (w) return -1;
-	if (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)
-	    return w;
-	z->c -= w;
+        int ch;
+        int w = get_b_utf8(z->p, z->c, z->lb, & ch);
+        if (!w) return -1;
+        if (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)
+            return w;
+        z->c -= w;
     } while (repeat);
     return 0;
 }
 
 extern int out_grouping_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) {
     do {
-	int ch;
-	int w = get_utf8(z->p, z->c, z->l, & ch);
-	unless (w) return -1;
-	unless (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)
-	    return w;
-	z->c += w;
+        int ch;
+        int w = get_utf8(z->p, z->c, z->l, & ch);
+        if (!w) return -1;
+        if (!(ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0))
+            return w;
+        z->c += w;
     } while (repeat);
     return 0;
 }
 
 extern int out_grouping_b_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) {
     do {
-	int ch;
-	int w = get_b_utf8(z->p, z->c, z->lb, & ch);
-	unless (w) return -1;
-	unless (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)
-	    return w;
-	z->c -= w;
+        int ch;
+        int w = get_b_utf8(z->p, z->c, z->lb, & ch);
+        if (!w) return -1;
+        if (!(ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0))
+            return w;
+        z->c -= w;
     } while (repeat);
     return 0;
 }
@@ -145,48 +161,48 @@ extern int out_grouping_b_U(struct SN_env * z, const unsigned char * s, int min,
 
 extern int in_grouping(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) {
     do {
-	int ch;
-	if (z->c >= z->l) return -1;
-	ch = z->p[z->c];
-	if (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)
-	    return 1;
-	z->c++;
+        int ch;
+        if (z->c >= z->l) return -1;
+        ch = z->p[z->c];
+        if (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)
+            return 1;
+        z->c++;
     } while (repeat);
     return 0;
 }
 
 extern int in_grouping_b(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) {
     do {
-	int ch;
-	if (z->c <= z->lb) return -1;
-	ch = z->p[z->c - 1];
-	if (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)
-	    return 1;
-	z->c--;
+        int ch;
+        if (z->c <= z->lb) return -1;
+        ch = z->p[z->c - 1];
+        if (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)
+            return 1;
+        z->c--;
     } while (repeat);
     return 0;
 }
 
 extern int out_grouping(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) {
     do {
-	int ch;
-	if (z->c >= z->l) return -1;
-	ch = z->p[z->c];
-	unless (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)
-	    return 1;
-	z->c++;
+        int ch;
+        if (z->c >= z->l) return -1;
+        ch = z->p[z->c];
+        if (!(ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0))
+            return 1;
+        z->c++;
     } while (repeat);
     return 0;
 }
 
 extern int out_grouping_b(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) {
     do {
-	int ch;
-	if (z->c <= z->lb) return -1;
-	ch = z->p[z->c - 1];
-	unless (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)
-	    return 1;
-	z->c--;
+        int ch;
+        if (z->c <= z->lb) return -1;
+        ch = z->p[z->c - 1];
+        if (!(ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0))
+            return 1;
+        z->c--;
     } while (repeat);
     return 0;
 }
@@ -215,7 +231,7 @@ extern int find_among(struct SN_env * z, const struct among * v, int v_size) {
     int j = v_size;
 
     int c = z->c; int l = z->l;
-    symbol * q = z->p + c;
+    const symbol * q = z->p + c;
 
     const struct among * w;
 
@@ -224,7 +240,7 @@ extern int find_among(struct SN_env * z, const struct among * v, int v_size) {
 
     int first_key_inspected = 0;
 
-    while(1) {
+    while (1) {
         int k = i + ((j - i) >> 1);
         int diff = 0;
         int common = common_i < common_j ? common_i : common_j; /* smaller */
@@ -237,8 +253,13 @@ extern int find_among(struct SN_env * z, const struct among * v, int v_size) {
                 common++;
             }
         }
-        if (diff < 0) { j = k; common_j = common; }
-                 else { i = k; common_i = common; }
+        if (diff < 0) {
+            j = k;
+            common_j = common;
+        } else {
+            i = k;
+            common_i = common;
+        }
         if (j - i <= 1) {
             if (i > 0) break; /* v->s has been inspected */
             if (j == i) break; /* only one item in v */
@@ -251,7 +272,7 @@ extern int find_among(struct SN_env * z, const struct among * v, int v_size) {
             first_key_inspected = 1;
         }
     }
-    while(1) {
+    while (1) {
         w = v + i;
         if (common_i >= w->s_size) {
             z->c = c + w->s_size;
@@ -275,7 +296,7 @@ extern int find_among_b(struct SN_env * z, const struct among * v, int v_size) {
     int j = v_size;
 
     int c = z->c; int lb = z->lb;
-    symbol * q = z->p + c - 1;
+    const symbol * q = z->p + c - 1;
 
     const struct among * w;
 
@@ -284,7 +305,7 @@ extern int find_among_b(struct SN_env * z, const struct among * v, int v_size) {
 
     int first_key_inspected = 0;
 
-    while(1) {
+    while (1) {
         int k = i + ((j - i) >> 1);
         int diff = 0;
         int common = common_i < common_j ? common_i : common_j;
@@ -306,7 +327,7 @@ extern int find_among_b(struct SN_env * z, const struct among * v, int v_size) {
             first_key_inspected = 1;
         }
     }
-    while(1) {
+    while (1) {
         w = v + i;
         if (common_i >= w->s_size) {
             z->c = c - w->s_size;
@@ -367,11 +388,10 @@ extern int replace_s(struct SN_env * z, int c_bra, int c_ket, int s_size, const
         z->l += adjustment;
         if (z->c >= c_ket)
             z->c += adjustment;
-        else
-            if (z->c > c_bra)
-                z->c = c_bra;
+        else if (z->c > c_bra)
+            z->c = c_bra;
     }
-    unless (s_size == 0) memmove(z->p + c_bra, s, s_size * sizeof(symbol));
+    if (s_size) memmove(z->p + c_bra, s, s_size * sizeof(symbol));
     if (adjptr != NULL)
         *adjptr = adjustment;
     return 0;
@@ -417,12 +437,7 @@ extern int insert_s(struct SN_env * z, int bra, int ket, int s_size, const symbo
 }
 
 extern int insert_v(struct SN_env * z, int bra, int ket, const symbol * p) {
-    int adjustment;
-    if (replace_s(z, bra, ket, SIZE(p), p, &adjustment))
-        return -1;
-    if (bra <= z->bra) z->bra += adjustment;
-    if (bra <= z->ket) z->ket += adjustment;
-    return 0;
+    return insert_s(z, bra, ket, SIZE(p), p);
 }
 
 extern symbol * slice_to(struct SN_env * z, symbol * p) {
@@ -455,6 +470,16 @@ extern symbol * assign_to(struct SN_env * z, symbol * p) {
     return p;
 }
 
+extern int len_utf8(const symbol * p) {
+    int size = SIZE(p);
+    int len = 0;
+    while (size--) {
+        symbol b = *p++;
+        if (b >= 0xC0 || b < 0x80) ++len;
+    }
+    return len;
+}
+
 #if 0
 extern void debug(struct SN_env * z, int number, int line_count) {
     int i;