From 67a0bf383f00d4f25d9ca27864caabf2a0f572b2 Mon Sep 17 00:00:00 2001
From: drh <>
Date: Mon, 18 Jul 2022 11:44:16 +0000
Subject: [PATCH] Enhance the ext/misc/regexp.c code so that when it is
 compiled with SQLITE_DEBUG, a new function named regexp_bytecode() is
 available that prints out the compiled NFA as human-readable text, for
 debugging purposes.

FossilOrigin-Name: cb5c08978fe8f074e6ae16953575213709e98b8bbae4359e0d2e6de67a7ea9e5
---
 ext/misc/regexp.c | 95 ++++++++++++++++++++++++++++++++++++++++++++++-
 manifest          | 14 +++----
 manifest.uuid     |  2 +-
 3 files changed, 102 insertions(+), 9 deletions(-)

diff --git a/ext/misc/regexp.c b/ext/misc/regexp.c
index d5b3872058..54c505233d 100644
--- a/ext/misc/regexp.c
+++ b/ext/misc/regexp.c
@@ -94,6 +94,31 @@ SQLITE_EXTENSION_INIT1
 #define RE_OP_NOTSPACE   16    /* Not a digit */
 #define RE_OP_BOUNDARY   17    /* Boundary between word and non-word */
 
+#if defined(SQLITE_DEBUG)
+/* Opcode names used for symbolic debugging */
+static const char *ReOpName[] = {
+  "EOF",
+  "MATCH",
+  "ANY",
+  "ANYSTAR",
+  "FORK",
+  "GOTO",
+  "ACCEPT",
+  "CC_INC",
+  "CC_EXC",
+  "CC_VALUE",
+  "CC_RANGE",
+  "WORD",
+  "NOTWORD",
+  "DIGIT",
+  "NOTDIGIT",
+  "SPACE",
+  "NOTSPACE",
+  "BOUNDARY",
+};
+#endif /* SQLITE_DEBUG */
+
+
 /* Each opcode is a "state" in the NFA */
 typedef unsigned short ReStateNumber;
 
@@ -127,7 +152,7 @@ struct ReCompiled {
   int *aArg;                  /* Arguments to each operator */
   unsigned (*xNextChar)(ReInput*);  /* Next character function */
   unsigned char zInit[12];    /* Initial text to match */
-  int nInit;                  /* Number of characters in zInit */
+  int nInit;                  /* Number of bytes in zInit */
   unsigned nState;            /* Number of entries in aOp[] and aArg[] */
   unsigned nAlloc;            /* Slots allocated for aOp[] and aArg[] */
 };
@@ -745,6 +770,67 @@ static void re_sql_func(
   }
 }
 
+#if defined(SQLITE_DEBUG)
+/*
+** This function is used for testing and debugging only.  It is only available
+** if the SQLITE_DEBUG compile-time option is used.
+**
+** Compile a regular expression and then convert the compiled expression into
+** text and return that text.
+*/
+static void re_bytecode_func(
+  sqlite3_context *context,
+  int argc,
+  sqlite3_value **argv
+){
+  const char *zPattern;
+  const char *zErr;
+  ReCompiled *pRe;
+  sqlite3_str *pStr;
+  int i;
+  int n;
+  char *z;
+
+  zPattern = (const char*)sqlite3_value_text(argv[0]);
+  if( zPattern==0 ) return;
+  zErr = re_compile(&pRe, zPattern, sqlite3_user_data(context)!=0);
+  if( zErr ){
+    re_free(pRe);
+    sqlite3_result_error(context, zErr, -1);
+    return;
+  }
+  if( pRe==0 ){
+    sqlite3_result_error_nomem(context);
+    return;
+  }
+  pStr = sqlite3_str_new(0);
+  if( pStr==0 ) goto re_bytecode_func_err;
+  if( pRe->nInit>0 ){
+    sqlite3_str_appendf(pStr, "INIT     ");
+    for(i=0; i<pRe->nInit; i++){
+      sqlite3_str_appendf(pStr, "%02x", pRe->zInit[i]);
+    }
+    sqlite3_str_appendf(pStr, "\n");
+  }
+  for(i=0; i<pRe->nState; i++){
+    sqlite3_str_appendf(pStr, "%-8s %4d\n",
+         ReOpName[(unsigned char)pRe->aOp[i]], pRe->aArg[i]);
+  }
+  n = sqlite3_str_length(pStr);
+  z = sqlite3_str_finish(pStr);
+  if( n==0 ){
+    sqlite3_free(z);
+  }else{
+    sqlite3_result_text(context, z, n-1, sqlite3_free);
+  }
+
+re_bytecode_func_err:
+  re_free(pRe);
+}
+
+#endif /* SQLITE_DEBUG */
+
+
 /*
 ** Invoke this routine to register the regexp() function with the
 ** SQLite database connection.
@@ -769,6 +855,13 @@ int sqlite3_regexp_init(
     rc = sqlite3_create_function(db, "regexpi", 2,
                             SQLITE_UTF8|SQLITE_INNOCUOUS|SQLITE_DETERMINISTIC,
                             (void*)db, re_sql_func, 0, 0);
+#if defined(SQLITE_DEBUG)
+    if( rc==SQLITE_OK ){
+      rc = sqlite3_create_function(db, "regexp_bytecode", 1,
+                            SQLITE_UTF8|SQLITE_INNOCUOUS|SQLITE_DETERMINISTIC,
+                            0, re_bytecode_func, 0, 0);
+    }
+#endif /* SQLITE_DEBUG */
   }
   return rc;
 }
diff --git a/manifest b/manifest
index 7922441c62..e2ba6ba112 100644
--- a/manifest
+++ b/manifest
@@ -1,5 +1,5 @@
-C Fixes\sfor\sthe\sgenerated\s"mallocs.tcl"\sand\s"leaks.tcl"\sscripts\sgenerated\sby\srunning\stcl\stests\swith\sthe\s--malloctrace=1\soption.
-D 2022-07-16T18:08:48.116
+C Enhance\sthe\sext/misc/regexp.c\scode\sso\sthat\swhen\sit\sis\scompiled\swith\nSQLITE_DEBUG,\sa\snew\sfunction\snamed\sregexp_bytecode()\sis\savailable\sthat\nprints\sout\sthe\scompiled\sNFA\sas\shuman-readable\stext,\sfor\sdebugging\spurposes.
+D 2022-07-18T11:44:16.316
 F .fossil-settings/empty-dirs dbb81e8fc0401ac46a1491ab34a7f2c7c0452f2f06b54ebb845d024ca8283ef1
 F .fossil-settings/ignore-glob 35175cdfcf539b2318cb04a9901442804be81cd677d8b889fcc9149c21f239ea
 F LICENSE.md df5091916dbb40e6e9686186587125e1b2ff51f022cc334e886c19a0e9982724
@@ -334,7 +334,7 @@ F ext/misc/normalize.c bd84355c118e297522aba74de34a4fd286fc775524e0499b14473918d
 F ext/misc/percentile.c b9086e223d583bdaf8cb73c98a6539d501a2fc4282654adbfea576453d82e691
 F ext/misc/prefixes.c 0f4f8cff5aebc00a7e3ac4021fd59cfe1a8e17c800ceaf592859ecb9cbc38196
 F ext/misc/qpvtab.c 09738419e25f603a35c0ac8bd0a04daab794f48d08a9bc07a6085b9057b99009
-F ext/misc/regexp.c c1fb4f0639ad70f180fa7ebca3b9830c1c648cdcd072e28747f727699f3a5071
+F ext/misc/regexp.c 554bbbc06980fac640618da2401b9b1a5da722f60fbc9ecd0de0244c96173a23
 F ext/misc/remember.c add730f0f7e7436cd15ea3fd6a90fd83c3f706ab44169f7f048438b7d6baa69c
 F ext/misc/rot13.c 51ac5f51e9d5fd811db58a9c23c628ad5f333c173f1fc53c8491a3603d38556c
 F ext/misc/scrub.c 2a44b0d44c69584c0580ad2553f6290a307a49df4668941d2812135bfb96a946
@@ -1980,8 +1980,8 @@ F vsixtest/vsixtest.tcl 6a9a6ab600c25a91a7acc6293828957a386a8a93
 F vsixtest/vsixtest.vcxproj.data 2ed517e100c66dc455b492e1a33350c1b20fbcdc
 F vsixtest/vsixtest.vcxproj.filters 37e51ffedcdb064aad6ff33b6148725226cd608e
 F vsixtest/vsixtest_TemporaryKey.pfx e5b1b036facdb453873e7084e1cae9102ccc67a0
-P 92d60b64ebfc2d1f0a9cabaa88e7bf0d11737ed01a77b627af10dd1b96a5321c
-R c2826cbb8b9c28cc73030d164bbfa10e
-U dan
-Z 0beb54d1e8fdf448050ee0b11342583e
+P 449799e2d5902464540e8fda53ab429e0518278dab3b17c86911759114cddea0
+R c5fe187a7b0d702467f87ef01664ca12
+U drh
+Z c7627cc664c2932a41026727705de830
 # Remove this line to create a well-formed Fossil manifest.
diff --git a/manifest.uuid b/manifest.uuid
index 496a3b3050..5a0612c077 100644
--- a/manifest.uuid
+++ b/manifest.uuid
@@ -1 +1 @@
-449799e2d5902464540e8fda53ab429e0518278dab3b17c86911759114cddea0
\ No newline at end of file
+cb5c08978fe8f074e6ae16953575213709e98b8bbae4359e0d2e6de67a7ea9e5
\ No newline at end of file