From 1ba1b5511ce578095e1153f36ba820cd170866ef Mon Sep 17 00:00:00 2001
From: danielk1977 <danielk1977@noemail.net>
Date: Wed, 23 Jun 2004 13:46:32 +0000
Subject: [PATCH] Improve test coverage of utf.c. (CVS 1678)

FossilOrigin-Name: 53c553c760a84e213c2277588bcc292892c2b0ce
---
 manifest      | 16 ++++++++--------
 manifest.uuid |  2 +-
 src/test5.c   | 26 +++++++++++++++++++++-----
 src/utf.c     | 36 ++++++++++++++----------------------
 test/enc.test |  4 ++--
 5 files changed, 46 insertions(+), 38 deletions(-)

diff --git a/manifest b/manifest
index 6fbd6974c0..9fc90af207 100644
--- a/manifest
+++ b/manifest
@@ -1,5 +1,5 @@
-C Fix\sa\sbug\sin\sthe\sprevious\scheckin.\s(CVS\s1677)
-D 2004-06-23T12:35:15
+C Improve\stest\scoverage\sof\sutf.c.\s(CVS\s1678)
+D 2004-06-23T13:46:32
 F Makefile.in 0a3d7aaefa50717bd550b0cf568a51072c4c103c
 F Makefile.linux-gcc a9e5a0d309fa7c38e7c14d3ecf7690879d3a5457
 F README f1de682fbbd94899d50aca13d387d1b3fd3be2dd
@@ -66,11 +66,11 @@ F src/test1.c 2504259e29a250190581aaac73d31e71e155df4b
 F src/test2.c dafd8bd314a554bf376c6d3a8c83fd69219f5a40
 F src/test3.c 7247090d15a5a43823079b6fd8dad1ed3cccdedf
 F src/test4.c a921a69821fd30209589228e64f94e9f715b6fe2
-F src/test5.c 1b7c275b2929dbd034a567255d2aca339410d1d6
+F src/test5.c 605e9cf8ce768a51e5938c9e64c05d5430d5ed02
 F src/tokenize.c 900374b6b37f04748bcd48c2d29a41c251542935
 F src/trigger.c 296e888ae931e9774e1761996b2b66db40f7d216
 F src/update.c b66b1896c9da54678ba3eff2bf0b4d291a95986a
-F src/utf.c 9c66605c5eac4c4aa999207634fabe7885e5b0b3
+F src/utf.c a8fb39c94ae4ef6606d5216b6f05d9b86f6901b2
 F src/util.c e31e35d3d76cab7a02045095064897eca49cbce3
 F src/vacuum.c fcb930215a3f6c50087300782555f61ad11dd80c
 F src/vdbe.c 0d8f2fea386c173997fb8eab555ab999497591fe
@@ -108,7 +108,7 @@ F test/crash.test 01b4a1cf195678138810f973ec9e2e6cef731d3e
 F test/crashtest1.c 09c1c7d728ccf4feb9e481671e29dda5669bbcc2
 F test/date.test aed5030482ebc02bd8d386c6c86a29f694ab068d
 F test/delete.test 4f0c86e2bebdc822d179c80697b1ceabe6bbcd07
-F test/enc.test e29a0d0558f4597ac1652a7bbac03b0e4ce17416
+F test/enc.test 2f5463af488d50aef60c6110bec6b21b5efba961
 F test/enc2.test 8a3b62254292c8ed6bb0671fcecb032b6fcbc356
 F test/enc3.test 315f302ed9a6042be76710eb6aa70e4551e9aa73
 F test/expr.test b4e945265c4c697bf5213b72558914ba10a989cc
@@ -229,7 +229,7 @@ F www/tclsqlite.tcl 19191cf2a1010eaeff74c51d83fd5f5a4d899075
 F www/vdbe.tcl 59288db1ac5c0616296b26dce071c36cb611dfe9
 F www/version3.tcl 563ba3ac02f64da27ab17f3edbe8e56bfd0293fb
 F www/whentouse.tcl a8335bce47cc2fddb07f19052cb0cb4d9129a8e4
-P db6bab574869fde49a147d4e19dd73005f247092
-R 5a1f0cbd459d0c1b045d2cff0431415e
+P f31c84a64f53cf2e75a5e988fbb6ad82537f2403
+R f27d37e6f5bb430f35de9b0eadf438e6
 U danielk1977
-Z 013707b4c596e9bfbd71b5c3bc4ddcc6
+Z 8f088a4eb584d07b7963f42469cbadf6
diff --git a/manifest.uuid b/manifest.uuid
index bb52b53b97..6f379fc5b1 100644
--- a/manifest.uuid
+++ b/manifest.uuid
@@ -1 +1 @@
-f31c84a64f53cf2e75a5e988fbb6ad82537f2403
\ No newline at end of file
+53c553c760a84e213c2277588bcc292892c2b0ce
\ No newline at end of file
diff --git a/src/test5.c b/src/test5.c
index 8ce0053237..cd6d877f3b 100644
--- a/src/test5.c
+++ b/src/test5.c
@@ -15,7 +15,7 @@
 ** is used for testing the SQLite routines for converting between
 ** the various supported unicode encodings.
 **
-** $Id: test5.c,v 1.11 2004/06/18 04:24:55 danielk1977 Exp $
+** $Id: test5.c,v 1.12 2004/06/23 13:46:32 danielk1977 Exp $
 */
 #include "sqliteInt.h"
 #include "vdbeInt.h"
@@ -116,6 +116,10 @@ static u8 name_to_enc(Tcl_Interp *interp, Tcl_Obj *pObj){
   return pEnc->enc;
 }
 
+/*
+** Usage:   test_translate <string/blob> <from enc> <to enc> ?<transient>?
+**
+*/
 static int test_translate(
   void * clientData,
   Tcl_Interp *interp,
@@ -126,16 +130,20 @@ static int test_translate(
   u8 enc_to;
   sqlite3_value *pVal;
 
-  const char *z;
+  char *z;
   int len;
+  void (*xDel)(void *p) = SQLITE_STATIC;
 
-  if( objc!=4 ){
+  if( objc!=4 && objc!=5 ){
     Tcl_AppendResult(interp, "wrong # args: should be \"",
         Tcl_GetStringFromObj(objv[0], 0), 
         " <string/blob> <from enc> <to enc>", 0
     );
     return TCL_ERROR;
   }
+  if( objc==5 ){
+    xDel = sqlite3FreeX;
+  }
 
   enc_from = name_to_enc(interp, objv[2]);
   if( !enc_from ) return TCL_ERROR;
@@ -146,10 +154,18 @@ static int test_translate(
 
   if( enc_from==SQLITE_UTF8 ){
     z = Tcl_GetString(objv[1]);
-    sqlite3ValueSetStr(pVal, -1, z, enc_from, SQLITE_STATIC);
+    if( objc==5 ){
+      z = sqliteStrDup(z);
+    }
+    sqlite3ValueSetStr(pVal, -1, z, enc_from, xDel);
   }else{
     z = Tcl_GetByteArrayFromObj(objv[1], &len);
-    sqlite3ValueSetStr(pVal, -1, z, enc_from, SQLITE_STATIC);
+    if( objc==5 ){
+      char *zTmp = z;
+      z = sqliteMalloc(len);
+      memcpy(z, zTmp, len);
+    }
+    sqlite3ValueSetStr(pVal, -1, z, enc_from, xDel);
   }
 
   z = sqlite3ValueText(pVal, enc_to);
diff --git a/src/utf.c b/src/utf.c
index 4bb08b5e07..326e9bde96 100644
--- a/src/utf.c
+++ b/src/utf.c
@@ -12,7 +12,7 @@
 ** This file contains routines used to translate between UTF-8, 
 ** UTF-16, UTF-16BE, and UTF-16LE.
 **
-** $Id: utf.c,v 1.24 2004/06/23 00:23:49 danielk1977 Exp $
+** $Id: utf.c,v 1.25 2004/06/23 13:46:32 danielk1977 Exp $
 **
 ** Notes on UTF-8:
 **
@@ -292,11 +292,7 @@ int sqlite3VdbeMemTranslate(Mem *pMem, u8 desiredEnc){
     zOut = sqliteMallocRaw(len);
     if( !zOut ) return SQLITE_NOMEM;
   }else{
-    if( pMem->z==pMem->zShort ){
-      zOut = zShort;
-    }else{
-      zOut = pMem->zShort;
-    }
+    zOut = zShort;
   }
   z = zOut;
 
@@ -343,9 +339,7 @@ int sqlite3VdbeMemTranslate(Mem *pMem, u8 desiredEnc){
   sqlite3VdbeMemRelease(pMem);
   pMem->flags &= ~(MEM_Static|MEM_Dyn|MEM_Ephem|MEM_Short);
   pMem->enc = desiredEnc;
-  if( (char *)zOut==pMem->zShort ){
-    pMem->flags |= (MEM_Term|MEM_Short);
-  }else if( zOut==zShort ){
+  if( zOut==zShort ){
     memcpy(pMem->zShort, zOut, len);
     zOut = pMem->zShort;
     pMem->flags |= (MEM_Term|MEM_Short);
@@ -390,22 +384,20 @@ int sqlite3VdbeMemHandleBom(Mem *pMem){
   }
   
   if( bom ){
-    if( pMem->flags & MEM_Short ){
-      memmove(pMem->zShort, &pMem->zShort[2], NBFS-2);
-      pMem->n -= 2;
-      pMem->enc = bom;
-    }
-    else if( pMem->flags & MEM_Dyn ){
+    /* This function is called as soon as a string is stored in a Mem*,
+    ** from within sqlite3VdbeMemSetStr(). At that point it is not possible
+    ** for the string to be stored in Mem.zShort, or for it to be stored
+    ** in dynamic memory with no destructor.
+    */
+    assert( !(pMem->flags&MEM_Short) );
+    assert( !(pMem->flags&MEM_Dyn) || pMem->xDel );
+    if( pMem->flags & MEM_Dyn ){
       void (*xDel)(void*) = pMem->xDel;
       char *z = pMem->z;
       pMem->z = 0;
       pMem->xDel = 0;
       rc = sqlite3VdbeMemSetStr(pMem, &z[2], pMem->n-2, bom, SQLITE_TRANSIENT);
-      if( xDel ){
-        xDel(z);
-      }else{
-        sqliteFree(z);
-      }
+      xDel(z);
     }else{
       rc = sqlite3VdbeMemSetStr(pMem, &pMem->z[2], pMem->n-2, bom, 
           SQLITE_TRANSIENT);
@@ -424,7 +416,7 @@ int sqlite3VdbeMemHandleBom(Mem *pMem){
 int sqlite3utf8CharLen(const char *z, int nByte){
   int r = 0;
   const char *zTerm;
-  if( nByte>0 ){
+  if( nByte>=0 ){
     zTerm = &z[nByte];
   }else{
     zTerm = (const char *)(-1);
@@ -528,7 +520,7 @@ void sqlite3utfSelfTest(){
   int n;
   int c;
 
-  for(i=0; 0 && i<0x00110000; i++){
+  for(i=0; i<0x00110000; i++){
     z = zBuf;
     WRITE_UTF8(z, i);
     n = z-zBuf;
diff --git a/test/enc.test b/test/enc.test
index 6b4575e6b2..7242257744 100644
--- a/test/enc.test
+++ b/test/enc.test
@@ -13,7 +13,7 @@
 # various suported unicode encodings (UTF-8, UTF-16, UTF-16le and
 # UTF-16be).
 #
-# $Id: enc.test,v 1.3 2004/06/19 00:16:31 drh Exp $
+# $Id: enc.test,v 1.4 2004/06/23 13:46:33 danielk1977 Exp $
 
 set testdir [file dirname $argv0]
 source $testdir/tester.tcl
@@ -104,7 +104,7 @@ proc test_conversion {testname str} {
 
   # Step 4 (little endian).
   append utf16le_bom "\xFF\xFE" $utf16le
-  set utf8_sqlite3 [test_translate $utf16le_bom UTF16 UTF8]
+  set utf8_sqlite3 [test_translate $utf16le_bom UTF16 UTF8 1]
   do_bincmp_test $testname.4.le $utf8_sqlite3 [binarize $str]
 
   # Step 4 (big endian).