Fix pg_restore's direct-to-database mode for standard_conforming_strings.

pg_backup_db.c contained a mini SQL lexer with which it tried to identify
boundaries between SQL commands, but that code was not designed to cope
with standard_conforming_strings, and would get the wrong answer if a
backslash immediately precedes a closing single quote in such a string,
as per report from Julian Mehnle.  The bug only affects direct-to-database
restores from archive files made with standard_conforming_strings = on.

Rather than complicating the code some more to try to fix that, let's just
rip it all out.  The only reason it was needed was to cope with COPY data
embedded into ordinary archive entries, which was a layout that was used
only for about the first three weeks of the archive format's existence,
and never in any production release of pg_dump.  Instead, just rely on the
archive file layout to tell us whether we're printing COPY data or not.

This bug represents a data corruption hazard in all releases in which
standard_conforming_strings can be turned on, ie 8.2 and later, so
back-patch to all supported branches.

Author: tglsfdc

src/bin/pg_dump/pg_backup_archiver.c

@@ -63,6 +63,7 @@ static TocEntry *getTocEntryByDumpId(ArchiveHandle *AH, DumpId id);
 static void _moveAfter(ArchiveHandle *AH, TocEntry *pos, TocEntry *te);
 static int	_discoverArchiveFormat(ArchiveHandle *AH);
 
+static int	RestoringToDB(ArchiveHandle *AH);
 static void dump_lo_buf(ArchiveHandle *AH);
 static void _write_msg(const char *modulename, const char *fmt, va_list ap);
 static void _die_horribly(ArchiveHandle *AH, const char *modulename, const char *fmt, va_list ap);
@@ -379,14 +380,7 @@ RestoreArchive(Archive *AHX, RestoreOptions *ropt)
 							  te->tag);
 
 						/*
-						 * If we have a copy statement, use it. As of V1.3,
-						 * these are separate to allow easy import from
-						 * withing a database connection. Pre 1.3 archives can
-						 * not use DB connections and are sent to output only.
-						 *
-						 * For V1.3+, the table data MUST have a copy
-						 * statement so that we can go into appropriate mode
-						 * with libpq.
+						 * If we have a copy statement, use it.
 						 */
 						if (te->copyStmt && strlen(te->copyStmt) > 0)
 						{
@@ -396,7 +390,15 @@ RestoreArchive(Archive *AHX, RestoreOptions *ropt)
 
 						(*AH->PrintTocDataPtr) (AH, te, ropt);
 
-						AH->writingCopyData = false;
+						/*
+						 * Terminate COPY if needed.
+						 */
+						if (AH->writingCopyData)
+						{
+							if (RestoringToDB(AH))
+								EndDBCopyMode(AH, te);
+							AH->writingCopyData = false;
+						}
 
 						_enableTriggersIfNecessary(AH, te, ropt);
 					}
@@ -1038,17 +1040,13 @@ ahprintf(ArchiveHandle *AH, const char *fmt,...)
 {
 	char	   *p = NULL;
 	va_list		ap;
-	int			bSize = strlen(fmt) + 256;		/* Should be enough */
+	int			bSize = strlen(fmt) + 256;		/* Usually enough */
 	int			cnt = -1;
 
 	/*
 	 * This is paranoid: deal with the possibility that vsnprintf is willing
-	 * to ignore trailing null
-	 */
-
-	/*
-	 * or returns > 0 even if string does not fit. It may be the case that it
-	 * returns cnt = bufsize
+	 * to ignore trailing null or returns > 0 even if string does not fit.
+	 * It may be the case that it returns cnt = bufsize.
 	 */
 	while (cnt < 0 || cnt >= (bSize - 1))
 	{
@@ -1128,7 +1126,7 @@ dump_lo_buf(ArchiveHandle *AH)
 
 
 /*
- *	Write buffer to the output file (usually stdout). This is user for
+ *	Write buffer to the output file (usually stdout). This is used for
  *	outputting 'restore' scripts etc. It is even possible for an archive
  *	format to create a custom output routine to 'fake' a restore if it
  *	wants to generate a script (see TAR output).
@@ -1180,7 +1178,7 @@ ahwrite(const void *ptr, size_t size, size_t nmemb, ArchiveHandle *AH)
 		 * connected then send it to the DB.
 		 */
 		if (RestoringToDB(AH))
-			return ExecuteSqlCommandBuf(AH, (void *) ptr, size * nmemb);		/* Always 1, currently */
+			return ExecuteSqlCommandBuf(AH, (const char *) ptr, size * nmemb);
 		else
 		{
 			res = fwrite((void *) ptr, size, nmemb, AH->OF);
@@ -1730,9 +1728,6 @@ _allocAH(const char *FileSpec, const ArchiveFormat fmt,
 	AH->mode = mode;
 	AH->compression = compression;
 
-	AH->pgCopyBuf = createPQExpBuffer();
-	AH->sqlBuf = createPQExpBuffer();
-
 	/* Open stdout with no compression for AH output handle */
 	AH->gzOut = 0;
 	AH->OF = stdout;

src/bin/pg_dump/pg_backup_archiver.h

@@ -128,28 +128,6 @@ typedef struct _outputContext
 	int			gzOut;
 } OutputContext;
 
-typedef enum
-{
-	SQL_SCAN = 0,				/* normal */
-	SQL_IN_SQL_COMMENT,			/* -- comment */
-	SQL_IN_EXT_COMMENT,			/* slash-star comment */
-	SQL_IN_SINGLE_QUOTE,		/* '...' literal */
-	SQL_IN_E_QUOTE,				/* E'...' literal */
-	SQL_IN_DOUBLE_QUOTE,		/* "..." identifier */
-	SQL_IN_DOLLAR_TAG,			/* possible dollar-quote starting tag */
-	SQL_IN_DOLLAR_QUOTE			/* body of dollar quote */
-} sqlparseState;
-
-typedef struct
-{
-	sqlparseState state;		/* see above */
-	char		lastChar;		/* preceding char, or '\0' initially */
-	bool		backSlash;		/* next char is backslash quoted? */
-	int			braceDepth;		/* parenthesis nesting depth */
-	PQExpBuffer tagBuf;			/* dollar quote tag (NULL if not created) */
-	int			minTagEndPos;	/* first possible end position of $-quote */
-} sqlparseInfo;
-
 typedef enum
 {
 	STAGE_NONE = 0,
@@ -185,9 +163,6 @@ typedef struct _archiveHandle
 								 * Added V1.7 */
 	ArchiveFormat format;		/* Archive format */
 
-	sqlparseInfo sqlparse;
-	PQExpBuffer sqlBuf;
-
 	time_t		createDate;		/* Date archive created */
 
 	/*
@@ -235,8 +210,6 @@ typedef struct _archiveHandle
 								 * required */
 	bool		writingCopyData;	/* True when we are sending COPY data */
 	bool		pgCopyIn;		/* Currently in libpq 'COPY IN' mode. */
-	PQExpBuffer pgCopyBuf;		/* Left-over data from incomplete lines in
-								 * COPY IN */
 
 	int			loFd;			/* BLOB fd */
 	int			writingBlob;	/* Flag */