From: danielk1977 <danielk1977@noemail.net>
Date: Thu, 1 Jan 2009 07:08:54 +0000 (+0000)
Subject: Add pseudo-random tests of the fts3 expression parser. Revise the fix in (6091).... 
X-Git-Tag: version-3.6.10~92
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=5973e6a30b13e481a26422279a0a5a9884f4ad9d;p=thirdparty%2Fsqlite.git

Add pseudo-random tests of the fts3 expression parser. Revise the fix in (6091). (CVS 6092)

FossilOrigin-Name: 11c2d4686197fb3f0d601651d5bbb3492af8f0dd
---

diff --git a/ext/fts3/fts3_expr.c b/ext/fts3/fts3_expr.c
index fe6cf7af78..c98dc72100 100644
--- a/ext/fts3/fts3_expr.c
+++ b/ext/fts3/fts3_expr.c
@@ -532,6 +532,12 @@ static int fts3ExprParse(
         int eType = p->eType;
         assert( eType!=FTSQUERY_PHRASE || !p->pPhrase->isNot );
         isPhrase = (eType==FTSQUERY_PHRASE || p->pLeft);
+
+        /* The isRequirePhrase variable is set to true if a phrase or
+        ** an expression contained in parenthesis is required. If a
+        ** binary operator (AND, OR, NOT or NEAR) is encounted when
+        ** isRequirePhrase is set, this is a syntax error.
+        */
         if( !isPhrase && isRequirePhrase ){
           sqlite3Fts3ExprFree(p);
           rc = SQLITE_ERROR;
@@ -554,14 +560,19 @@ static int fts3ExprParse(
           pPrev = pAnd;
         }
 
+        /* This test catches attempts to make either operand of a NEAR
+        ** operator something other than a phrase. For example, either of
+        ** the following:
+        **
+        **    (bracketed expression) NEAR phrase
+        **    phrase NEAR (bracketed expression)
+        **
+        ** Return an error in either case.
+        */
         if( pPrev && (
-            (pPrev->eType==FTSQUERY_NEAR && eType!=FTSQUERY_PHRASE)
-         || (eType==FTSQUERY_NEAR && pPrev->eType!=FTSQUERY_PHRASE && !isPhrase)
+            (eType==FTSQUERY_NEAR && !isPhrase && pPrev->eType!=FTSQUERY_PHRASE)
+         || (eType!=FTSQUERY_PHRASE && isPhrase && pPrev->eType==FTSQUERY_NEAR)
         )){
-          /* This is an attempt to do "phrase NEAR (bracketed expression)"
-          ** or "(bracketed expression) NEAR phrase", both of which are
-          ** illegal. Return an error.
-          */
           sqlite3Fts3ExprFree(p);
           rc = SQLITE_ERROR;
           goto exprparse_out;
diff --git a/manifest b/manifest
index e4bcd8dc65..a9efbb5463 100644
--- a/manifest
+++ b/manifest
@@ -1,5 +1,5 @@
-C Fix\sa\sbug\sparsing\s"<expr>\sAND\s(abc\sNEAR\sdef)"\sin\sfts3_expr.c.\s(CVS\s6091)
-D 2009-01-01T04:19:51
+C Add\spseudo-random\stests\sof\sthe\sfts3\sexpression\sparser.\sRevise\sthe\sfix\sin\s(6091).\s(CVS\s6092)
+D 2009-01-01T07:08:55
 F Makefile.arm-wince-mingw32ce-gcc fcd5e9cd67fe88836360bb4f9ef4cb7f8e2fb5a0
 F Makefile.in 77635d0909c2067cee03889a1e04ce910d8fb809
 F Makefile.linux-gcc d53183f4aa6a9192d249731c90dbdffbd2c68654
@@ -55,7 +55,7 @@ F ext/fts3/README.tokenizers 998756696647400de63d5ba60e9655036cb966e9
 F ext/fts3/README.txt 8c18f41574404623b76917b9da66fcb0ab38328d
 F ext/fts3/fts3.c 3aa6aef1eadc44606f6ed3c841062735a5210077
 F ext/fts3/fts3.h 3a10a0af180d502cecc50df77b1b22df142817fe
-F ext/fts3/fts3_expr.c 4f00c773a3f7ebf2c7079b97c992f3f6ebb8cd6f
+F ext/fts3/fts3_expr.c ac57b3ae142da80527cccdc4ab4e100fac1737fb
 F ext/fts3/fts3_expr.h 4dad4d87cf5d41ea924a815fe89a6f87dc76f277
 F ext/fts3/fts3_hash.c e15e84d18f8df149ab290029872d4559c4c7c15a
 F ext/fts3/fts3_hash.h 004b759e1602ff16dfa02fea3ca1c77336ad6798
@@ -368,7 +368,8 @@ F test/fts3b.test b3a25180a633873d37d86e1ccd00ed690d37237a
 F test/fts3c.test 4c7ef29b37aca3e8ebb6a39b57910caa6506034e
 F test/fts3d.test d92a47fe8ed59c9e53d2d8e6d2685bb380aadadc
 F test/fts3e.test 1f6c6ac9cc8b772ca256e6b22aaeed50c9350851
-F test/fts3expr.test fa7bd31721cb3520521d43d5e4df3fbb9faa6426
+F test/fts3expr.test 40a2fff17a2487e9389234f7304b174064155acc
+F test/fts3expr2.test 8501de895a4c0631e7226c9bac055cd49c9f6646
 F test/fts3near.test e8a9b4e16c63a795918b334b74d4aec14815bf8b
 F test/func.test a50f0a4b69ac251debe1dce3ba29da7476dc8c52
 F test/fuzz.test 62fc19dd36a427777fd671b569df07166548628a
@@ -689,7 +690,7 @@ F tool/speedtest16.c c8a9c793df96db7e4933f0852abb7a03d48f2e81
 F tool/speedtest2.tcl ee2149167303ba8e95af97873c575c3e0fab58ff
 F tool/speedtest8.c 2902c46588c40b55661e471d7a86e4dd71a18224
 F tool/speedtest8inst1.c 293327bc76823f473684d589a8160bde1f52c14e
-P 73958060aaf641d93bede3a42851e5b3451f5432
-R b18c53d48a15a59b2ef3ca97050c92ee
+P d1a6a2edd799d65ff88510df951e909919e35b6b
+R 7818503c72e44fd985399b599f4925a2
 U danielk1977
-Z ecd0fc691e306c5b472f9b36b016cb1f
+Z e4f3adbf5a486d51a8277e421fd369a1
diff --git a/manifest.uuid b/manifest.uuid
index 985de7b7e0..08bc41a06a 100644
--- a/manifest.uuid
+++ b/manifest.uuid
@@ -1 +1 @@
-d1a6a2edd799d65ff88510df951e909919e35b6b
\ No newline at end of file
+11c2d4686197fb3f0d601651d5bbb3492af8f0dd
\ No newline at end of file
diff --git a/test/fts3expr.test b/test/fts3expr.test
index 31a328a6c1..3e70d5efc0 100644
--- a/test/fts3expr.test
+++ b/test/fts3expr.test
@@ -11,7 +11,7 @@
 # This file implements regression tests for SQLite library.  The
 # focus of this script is testing the FTS3 module.
 #
-# $Id: fts3expr.test,v 1.2 2009/01/01 04:19:51 danielk1977 Exp $
+# $Id: fts3expr.test,v 1.3 2009/01/01 07:08:55 danielk1977 Exp $
 #
 
 set testdir [file dirname $argv0]
@@ -122,6 +122,28 @@ do_test fts3expr-3.4 {
 do_test fts3expr-3.5 {
   test_fts3expr2 "one AND (two NEAR three)"
 } {AND one {NEAR/10 two three}}
+do_test fts3expr-3.6 {
+  test_fts3expr2 "one (two NEAR three)"
+} {AND one {NEAR/10 two three}}
+do_test fts3expr-3.7 {
+  test_fts3expr2 "(two NEAR three) one"
+} {AND {NEAR/10 two three} one}
+do_test fts3expr-3.8 {
+  test_fts3expr2 "(two NEAR three) AND one"
+} {AND {NEAR/10 two three} one}
+do_test fts3expr-3.9 {
+  test_fts3expr2 "(two NEAR three) (four five)"
+} {AND {NEAR/10 two three} {AND four five}}
+do_test fts3expr-3.10 {
+  test_fts3expr2 "(two NEAR three) AND (four five)"
+} {AND {NEAR/10 two three} {AND four five}}
+do_test fts3expr-3.11 {
+  test_fts3expr2 "(two NEAR three) (four NEAR five)"
+} {AND {NEAR/10 two three} {NEAR/10 four five}}
+do_test fts3expr-3.12 {
+  test_fts3expr2 "(two NEAR three) OR (four NEAR five)"
+} {OR {NEAR/10 two three} {NEAR/10 four five}}
+
 
 #------------------------------------------------------------------------
 # The following tests, fts3expr-4.*, test the parsers response to syntax
diff --git a/test/fts3expr2.test b/test/fts3expr2.test
new file mode 100644
index 0000000000..48f7194ef3
--- /dev/null
+++ b/test/fts3expr2.test
@@ -0,0 +1,167 @@
+# 2009 January 1
+#
+# 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.
+#
+#*************************************************************************
+# This file implements regression tests for SQLite library.  The
+# focus of this script is testing the FTS3 module syntax parser.
+#
+# $Id: fts3expr2.test,v 1.1 2009/01/01 07:08:55 danielk1977 Exp $
+#
+
+set testdir [file dirname $argv0]
+source $testdir/tester.tcl
+
+# If SQLITE_ENABLE_FTS3 is defined, omit this file.
+ifcapable !fts3 {
+  finish_test
+  return
+}
+
+# Test overview:
+# 
+#   The tests in this file are pseudo-randomly generated. They test
+#   the fts3 match expression parser via the test interface
+#   SQL function "fts3_exprtest" (see comments in fts3_expr.c).
+#
+#   Each test case works as follows:
+#   
+#     1. A random expression tree is generated using proc [random_expr_tree].
+#     2. The expression tree is converted to the text of an equivalent
+#        fts3 expression using proc [tree_to_expr].
+#     3. The test SQL function "fts3_exprtest" is used to parse the 
+#        expression text generated in step (2), returning a parsed expression
+#        tree.
+#     4. Test that the tree returned in step (3) matches that generated in 
+#        step (1).
+#
+#   In step (2), 4 different fts3 expressions are created from each 
+#   expression tree by varying the following boolean properties:
+#
+#     * Whether or not superflous parenthesis are included. i.e. if
+#       "a OR b AND (c OR d)" or "a OR (b AND (c OR d))" is generated.
+#
+#     * Whether or not explict AND operators are used. i.e. if
+#     "a OR b AND c" or "a OR b c" is generated.
+#
+
+set sqlite_fts3_enable_parentheses 1
+
+proc strip_phrase_data {L} {
+  if {[lindex $L 0] eq "PHRASE"} {
+    return [list P [lrange $L 3 end]]
+  }
+  return [list \
+    [lindex $L 0] \
+    [strip_phrase_data [lindex $L 1]] \
+    [strip_phrase_data [lindex $L 2]] \
+  ]
+}
+proc test_fts3expr2 {expr} {
+  strip_phrase_data [
+    db one {SELECT fts3_exprtest('simple', $expr, 'a', 'b', 'c')}
+  ]
+}
+
+proc rnd {nMax} { expr {int(rand()*$nMax)} }
+
+proc random_phrase {} {
+  set phrases [list one two three four "one two" "three four"]
+  list P [lindex $phrases [rnd [llength $phrases]]]
+}
+
+# Generate and return a pseudo-random expression tree. Using the same 
+# format returned by the [test_fts3expr2] proc.
+#
+proc random_expr_tree {iHeight} {
+  if {$iHeight==0 || [rnd 3]==0} {
+    return [random_phrase]
+  }
+
+  set operators [list NEAR NOT AND OR]
+  set op [lindex $operators [rnd 4]]
+
+  if {$op eq "NEAR"} {
+    set iDistance [rnd 15]
+    return [list $op/$iDistance [random_phrase] [random_phrase]]
+  }
+
+  set iNH [expr {$iHeight - 1}]
+  return [list $op [random_expr_tree $iNH] [random_expr_tree $iNH]]
+}
+
+# Given an expression tree, generate a corresponding expression.
+#
+proc tree_to_expr {tree all_brackets implicit_and} {
+  set prec(NOT) 2
+  set prec(AND) 3
+  set prec()    3
+  set prec(OR)  4
+
+  set op [lindex $tree 0]
+
+  if {$op eq "P"} {
+    set phrase [lindex $tree 1]
+    if {[llength $phrase]>1} {
+      return "\"$phrase\""
+    } else {
+      return $phrase
+    }
+  }
+
+  if {$op eq "NEAR/10"} {
+    set op "NEAR"
+  }
+  if {$op eq "AND" && $implicit_and} {
+    set op ""
+  }
+
+  set lhs [lindex $tree 1]
+  set rhs [lindex $tree 2]
+  set zLeft  [tree_to_expr $lhs $all_brackets $implicit_and]
+  set zRight [tree_to_expr $rhs $all_brackets $implicit_and]
+
+  set iPrec 5
+  set iLeftPrec 0
+  set iRightPrec 0
+
+  catch {set iPrec      $prec($op)}
+  catch {set iLeftPrec  $prec([lindex $lhs 0])}
+  catch {set iRightPrec $prec([lindex $rhs 0])}
+
+  if {$iLeftPrec > $iPrec || $all_brackets} {
+    set zLeft "($zLeft)"
+  } 
+  if {$iRightPrec >= $iPrec || $all_brackets} {
+    set zRight "($zRight)"
+  } 
+
+  return "$zLeft $op $zRight"
+}
+
+proc do_exprparse_test {name expr tree} {
+  uplevel do_test $name [list "test_fts3expr2 {$expr}"] [list $tree]
+}
+
+for {set iTest 1} {$iTest<500} {incr iTest} {
+  set t [random_expr_tree 4]
+
+  set e1 [tree_to_expr $t 0 0]
+  set e2 [tree_to_expr $t 0 1]
+  set e3 [tree_to_expr $t 1 0]
+  set e4 [tree_to_expr $t 1 1]
+
+  do_exprparse_test fts3expr2-$iTest.1 $e1 $t
+  do_exprparse_test fts3expr2-$iTest.2 $e2 $t
+  do_exprparse_test fts3expr2-$iTest.3 $e3 $t
+  do_exprparse_test fts3expr2-$iTest.4 $e4 $t
+}
+
+set sqlite_fts3_enable_parentheses 0
+finish_test
+