}
    return ((mask & 0x0007)==0x0005);
  }
}

/* porter rule condition: (m > 1 and (*S or *T)) */
static int fts5Porter_MGt1_and_S_or_T(char *zStem, int nStem){
  return nStem>0
      && (zStem[nStem-1]=='s' || zStem[nStem-1]=='t')
      && fts5Porter_MGt1(zStem, nStem);
}

/* porter rule condition: (*v*) */
static int fts5Porter_Vowel(char *zStem, int nStem){
  int i;
  for(i=0; i<nStem; i++){
................................................................................

  /* Steps 2 through 4. */
  fts5PorterStep2(aBuf, &nBuf);
  fts5PorterStep3(aBuf, &nBuf);
  fts5PorterStep4(aBuf, &nBuf);

  /* Step 5a. */

  if( nBuf>0 && aBuf[nBuf-1]=='e' ){
    if( fts5Porter_MGt1(aBuf, nBuf-1) 
     || (fts5Porter_MEq1(aBuf, nBuf-1) && !fts5Porter_Ostar(aBuf, nBuf-1))
    ){
      nBuf--;
    }
  }

    }
    return ((mask & 0x0007)==0x0005);
  }
}

/* porter rule condition: (m > 1 and (*S or *T)) */
static int fts5Porter_MGt1_and_S_or_T(char *zStem, int nStem){
  assert( nStem>0 );
  return (zStem[nStem-1]=='s' || zStem[nStem-1]=='t') 
      && fts5Porter_MGt1(zStem, nStem);
}

/* porter rule condition: (*v*) */
static int fts5Porter_Vowel(char *zStem, int nStem){
  int i;
  for(i=0; i<nStem; i++){
................................................................................

  /* Steps 2 through 4. */
  fts5PorterStep2(aBuf, &nBuf);
  fts5PorterStep3(aBuf, &nBuf);
  fts5PorterStep4(aBuf, &nBuf);

  /* Step 5a. */
  assert( nBuf>0 );
  if( aBuf[nBuf-1]=='e' ){
    if( fts5Porter_MGt1(aBuf, nBuf-1) 
     || (fts5Porter_MEq1(aBuf, nBuf-1) && !fts5Porter_Ostar(aBuf, nBuf-1))
    ){
      nBuf--;
    }
  }

# 2014 Dec 20
#
# The author disclaims copyright to this source code.  In place of
# a legal notice, here is a blessing:
#
#    May you do good and not evil.
#    May you find forgiveness for yourself and forgive others.
#    May you share freely, never taking more than you give.
#
#***********************************************************************
#
# Tests focusing on the fts5 porter stemmer implementation.
#
# These are extra tests added to those in fts5porter.test in order to
# improve test coverage of the porter stemmer implementation.
#

source [file join [file dirname [info script]] fts5_common.tcl]
set testprefix fts5porter2

set test_vocab {
  tion          tion
  ation         ation
  vation        vation
  avation       avat
  vion          vion
  ion           ion
  relational    relat
  relation      relat
  relate        relat
  zzz           zzz
  ii            ii
  iiing         ii
  xtional       xtional
  xenci         xenci
  xlogi         xlogi
  realization   realiz
  realize       realiz
  xization      xizat
  capitalism    capit
  talism        talism
  xiveness      xive
  xfulness      xful
  xousness      xous
  xical         xical
  xicate        xicat
  xicity        xiciti
  ies           ie
  eed           e
  eing           e
  s             s
}

set i 0
foreach {in out} $test_vocab {
  do_test "1.$i.($in -> $out)" {
    lindex [sqlite3_fts5_tokenize db porter $in] 0
  } $out
  incr i
}


finish_test

  INSERT INTO e5 VALUES($c || ' ' || $a);
}

do_execsql_test 7.1 {SELECT rowid FROM e5 WHERE e5 MATCH $a} { 1 3 }
do_execsql_test 7.2 {SELECT rowid FROM e5 WHERE e5 MATCH $b} { 1 2 }
do_execsql_test 7.3 {SELECT rowid FROM e5 WHERE e5 MATCH $c} { 2 3 }

































finish_test

  INSERT INTO e5 VALUES($c || ' ' || $a);
}

do_execsql_test 7.1 {SELECT rowid FROM e5 WHERE e5 MATCH $a} { 1 3 }
do_execsql_test 7.2 {SELECT rowid FROM e5 WHERE e5 MATCH $b} { 1 2 }
do_execsql_test 7.3 {SELECT rowid FROM e5 WHERE e5 MATCH $c} { 2 3 }

#-------------------------------------------------------------------------
# Test the 'separators' option with the unicode61 tokenizer.
#
do_execsql_test 8.1 {
  BEGIN;
  CREATE VIRTUAL TABLE e6 USING fts5(x,
    tokenize="unicode61 separators ABCDEFGHIJKLMNOPQRSTUVWXYZ"
  );
  INSERT INTO e6 VALUES('theAquickBbrownCfoxDjumpedWoverXtheYlazyZdog');
  CREATE VIRTUAL TABLE e7 USING fts5vocab(e6, 'row');
  SELECT term FROM e7;
  ROLLBACK;
} {
  brown dog fox jumped lazy over quick the
}

do_execsql_test 8.2 [subst {
  BEGIN;
  CREATE VIRTUAL TABLE e6 USING fts5(x,
    tokenize="unicode61 separators '\u0E01\u0E02\u0E03\u0E04\u0E05\u0E06\u0E07'"
  );
  INSERT INTO e6 VALUES('the\u0E01quick\u0E01brown\u0E01fox\u0E01' 
                     || 'jumped\u0E01over\u0E01the\u0E01lazy\u0E01dog'
  );
  INSERT INTO e6 VALUES('\u0E08\u0E07\u0E09');
  CREATE VIRTUAL TABLE e7 USING fts5vocab(e6, 'row');
  SELECT term FROM e7;
  ROLLBACK;
}] [subst {
  brown dog fox jumped lazy over quick the \u0E08 \u0E09
}]

finish_test

# Check that diacritics are removed if remove_diacritics=1 is specified.
# And that they do not break tokens.
do_unicode_token_test2 1.10 "xx\u0301xx" "xxxx xx\u301xx"

# Title-case mappings work
do_unicode_token_test 1.11 "\u01c5" "\u01c6 \u01c5"







#-------------------------------------------------------------------------
#
set docs [list {
  Enhance the INSERT syntax to allow multiple rows to be inserted via the
  VALUES clause.
} {
................................................................................
  execsql {
    CREATE VIRTUAL TABLE t1 USING fts5(tokenize=unicode61, x);
    INSERT INTO t1 VALUES($a);
    INSERT INTO t1 VALUES($b);
    INSERT INTO t1 VALUES($c);
    INSERT INTO t1 VALUES($d);
  }




} {}

do_test 4.2 {
  set a [binary format c* {0x61 0xF7 0xBF 0xBF 0xBF 0x62}]
  set b [binary format c* {0x61 0xF7 0xBF 0xBF 0xBF 0xBF 0x62}]
  set c [binary format c* {0x61 0xF7 0xBF 0xBF 0xBF 0xBF 0xBF 0x62}]
  set d [binary format c* {0x61 0xF7 0xBF 0xBF 0xBF 0xBF 0xBF 0xBF 0x62}]
................................................................................
  execsql {
    INSERT INTO t1 VALUES($a);
    INSERT INTO t1 VALUES($b);
    INSERT INTO t1 VALUES($c);
    INSERT INTO t1 VALUES($d);
  }
} {}











#-------------------------------------------------------------------------

breakpoint
do_unicode_token_test3 5.1 {tokenchars {}} {
  sqlite3_reset sqlite3_column_int

# Check that diacritics are removed if remove_diacritics=1 is specified.
# And that they do not break tokens.
do_unicode_token_test2 1.10 "xx\u0301xx" "xxxx xx\u301xx"

# Title-case mappings work
do_unicode_token_test 1.11 "\u01c5" "\u01c6 \u01c5"

do_unicode_token_test 1.12 "\u00C1abc\u00C2 \u00D1def\u00C3" \
    "\u00E1abc\u00E2 \u00C1abc\u00C2 \u00F1def\u00E3 \u00D1def\u00C3"

do_unicode_token_test 1.13 "\u00A2abc\u00A3 \u00A4def\u00A5" \
    "abc abc def def"

#-------------------------------------------------------------------------
#
set docs [list {
  Enhance the INSERT syntax to allow multiple rows to be inserted via the
  VALUES clause.
} {
................................................................................
  execsql {
    CREATE VIRTUAL TABLE t1 USING fts5(tokenize=unicode61, x);
    INSERT INTO t1 VALUES($a);
    INSERT INTO t1 VALUES($b);
    INSERT INTO t1 VALUES($c);
    INSERT INTO t1 VALUES($d);
  }

  execsql "CREATE VIRTUAL TABLE t8 USING fts5(
      a, b, tokenize=\"unicode61 separators '\uFFFE\uD800\u00BF'\"
  )"
} {}

do_test 4.2 {
  set a [binary format c* {0x61 0xF7 0xBF 0xBF 0xBF 0x62}]
  set b [binary format c* {0x61 0xF7 0xBF 0xBF 0xBF 0xBF 0x62}]
  set c [binary format c* {0x61 0xF7 0xBF 0xBF 0xBF 0xBF 0xBF 0x62}]
  set d [binary format c* {0x61 0xF7 0xBF 0xBF 0xBF 0xBF 0xBF 0xBF 0x62}]
................................................................................
  execsql {
    INSERT INTO t1 VALUES($a);
    INSERT INTO t1 VALUES($b);
    INSERT INTO t1 VALUES($c);
    INSERT INTO t1 VALUES($d);
  }
} {}

do_test 4.4 {
  sqlite3_exec_hex db {
    CREATE VIRTUAL TABLE t9 USING fts5(a, b, 
      tokenize="unicode61 separators '%C09004'"
    );
    INSERT INTO t9(a) VALUES('abc%88def %89ghi%90');
  }
} {0 {}}


#-------------------------------------------------------------------------

breakpoint
do_unicode_token_test3 5.1 {tokenchars {}} {
  sqlite3_reset sqlite3_column_int