Fix recently introduced bugs about parsing ispell/hunspell files.

feodor · feodor · commit 53ef36cb4aab · 2007-09-10T20:27:12.000Z
In most cases it cause because of unneeded lowercasing of flags.

Per experiment with regression checks with ispell dictionary.
diff --git a/src/backend/tsearch/spell.c b/src/backend/tsearch/spell.c
@@ -7,7 +7,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/tsearch/spell.c,v 1.3 2007/09/10 10:39:56 teodor Exp $
+ *	  $PostgreSQL: pgsql/src/backend/tsearch/spell.c,v 1.4 2007/09/10 20:27:12 teodor Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -647,12 +647,10 @@ NIImportOOAffixes(IspellDict * Conf, const char *filename)
 				goto nextline;
 			flag = *sflag;
 			isSuffix = (STRNCMP(ptype, "sfx") == 0) ? true : false;
-			pfind = lowerstr_ctx(find);
-			if (t_iseq(find, 'y'))
+			if (t_iseq(find, 'y') || t_iseq(find, 'Y'))
 				flagflags = FF_CROSSPRODUCT;
 			else
 				flagflags = 0;
-			pfree(pfind);
 		}
 		else
 		{
@@ -666,7 +664,7 @@ NIImportOOAffixes(IspellDict * Conf, const char *filename)
 			if ((ptr = strchr(prepl, '/')) != NULL)
 			{
 				*ptr = '\0';
-				ptr++;
+				ptr = repl + (ptr-prepl) + 1;
 				while (*ptr)
 				{
 					aflg |= Conf->flagval[(unsigned int) *ptr];
@@ -704,8 +702,7 @@ NIImportOOAffixes(IspellDict * Conf, const char *filename)
 void
 NIImportAffixes(IspellDict * Conf, const char *filename)
 {
-	char		str[BUFSIZ],
-			   *pstr = NULL;
+	char	   *pstr = NULL;
 	char		mask[BUFSIZ];
 	char		find[BUFSIZ];
 	char		repl[BUFSIZ];
@@ -742,13 +739,15 @@ NIImportAffixes(IspellDict * Conf, const char *filename)
 
 		if (STRNCMP(pstr, "compoundwords") == 0)
 		{
-			s = findchar(str, 'l');
+			s = findchar(pstr, 'l');
 			if (s)
 			{
+				s = recoded + ( s-pstr ); /* we need non-lowercased string */
 				while (*s && !t_isspace(s))
 					s++;
 				while (*s && t_isspace(s))
 					s++;
+
 				if (*s && pg_mblen(s) == 1)
 				{
 					Conf->flagval[(unsigned int) *s] = FF_COMPOUNDFLAG;
@@ -774,7 +773,7 @@ NIImportAffixes(IspellDict * Conf, const char *filename)
 		}
 		if (STRNCMP(pstr, "flag") == 0)
 		{
-			s = str + 4;
+			s = recoded + 4; /* we need non-lowercased string */
 			flagflags = 0;
 
 			while (*s && t_isspace(s))
@@ -1333,7 +1332,7 @@ addToResult(char **forms, char **cur, char *word)
 }
 
 static char **
-NormalizeSubWord(IspellDict * Conf, char *word, char flag)
+NormalizeSubWord(IspellDict * Conf, char *word, int flag)
 {
 	AffixNodeData *suffix = NULL,
 			   *prefix = NULL;

Original file line number	Diff line number	Diff line change
`@@ -7,7 +7,7 @@`
`7`	`7`	`*`
`8`	`8`	`*`
`9`	`9`	`* IDENTIFICATION`
`10`		`- * $PostgreSQL: pgsql/src/backend/tsearch/spell.c,v 1.3 2007/09/10 10:39:56 teodor Exp $`
	`10`	`+ * $PostgreSQL: pgsql/src/backend/tsearch/spell.c,v 1.4 2007/09/10 20:27:12 teodor Exp $`
`11`	`11`	`*`
`12`	`12`	`*-------------------------------------------------------------------------`
`13`	`13`	`*/`
`@@ -647,12 +647,10 @@ NIImportOOAffixes(IspellDict * Conf, const char *filename)`
`647`	`647`	`goto nextline;`
`648`	`648`	`flag = *sflag;`
`649`	`649`	`isSuffix = (STRNCMP(ptype, "sfx") == 0) ? true : false;`
`650`		`- pfind = lowerstr_ctx(find);`
`651`		`- if (t_iseq(find, 'y'))`
	`650`	`+ if (t_iseq(find, 'y') \|\| t_iseq(find, 'Y'))`
`652`	`651`	`flagflags = FF_CROSSPRODUCT;`
`653`	`652`	`else`
`654`	`653`	`flagflags = 0;`
`655`		`- pfree(pfind);`
`656`	`654`	`}`
`657`	`655`	`else`
`658`	`656`	`{`
`@@ -666,7 +664,7 @@ NIImportOOAffixes(IspellDict * Conf, const char *filename)`
`666`	`664`	`if ((ptr = strchr(prepl, '/')) != NULL)`
`667`	`665`	`{`
`668`	`666`	`*ptr = '\0';`
`669`		`- ptr++;`
	`667`	`+ ptr = repl + (ptr-prepl) + 1;`
`670`	`668`	`while (*ptr)`
`671`	`669`	`{`
`672`	`670`	`aflg \|= Conf->flagval[(unsigned int) *ptr];`
`@@ -704,8 +702,7 @@ NIImportOOAffixes(IspellDict * Conf, const char *filename)`
`704`	`702`	`void`
`705`	`703`	`NIImportAffixes(IspellDict * Conf, const char *filename)`
`706`	`704`	`{`
`707`		`- char str[BUFSIZ],`
`708`		`- *pstr = NULL;`
	`705`	`+ char *pstr = NULL;`
`709`	`706`	`char mask[BUFSIZ];`
`710`	`707`	`char find[BUFSIZ];`
`711`	`708`	`char repl[BUFSIZ];`
`@@ -742,13 +739,15 @@ NIImportAffixes(IspellDict * Conf, const char *filename)`
`742`	`739`
`743`	`740`	`if (STRNCMP(pstr, "compoundwords") == 0)`
`744`	`741`	`{`
`745`		`- s = findchar(str, 'l');`
	`742`	`+ s = findchar(pstr, 'l');`
`746`	`743`	`if (s)`
`747`	`744`	`{`
	`745`	`+ s = recoded + ( s-pstr ); /* we need non-lowercased string */`
`748`	`746`	`while (*s && !t_isspace(s))`
`749`	`747`	`s++;`
`750`	`748`	`while (*s && t_isspace(s))`
`751`	`749`	`s++;`
	`750`	`+`
`752`	`751`	`if (*s && pg_mblen(s) == 1)`
`753`	`752`	`{`
`754`	`753`	`Conf->flagval[(unsigned int) *s] = FF_COMPOUNDFLAG;`
`@@ -774,7 +773,7 @@ NIImportAffixes(IspellDict * Conf, const char *filename)`
`774`	`773`	`}`
`775`	`774`	`if (STRNCMP(pstr, "flag") == 0)`
`776`	`775`	`{`
`777`		`- s = str + 4;`
	`776`	`+ s = recoded + 4; /* we need non-lowercased string */`
`778`	`777`	`flagflags = 0;`
`779`	`778`
`780`	`779`	`while (*s && t_isspace(s))`
`@@ -1333,7 +1332,7 @@ addToResult(char forms, char cur, char *word)`
`1333`	`1332`	`}`
`1334`	`1333`
`1335`	`1334`	`static char **`
`1336`		`-NormalizeSubWord(IspellDict * Conf, char *word, char flag)`
	`1335`	`+NormalizeSubWord(IspellDict * Conf, char *word, int flag)`
`1337`	`1336`	`{`
`1338`	`1337`	`AffixNodeData *suffix = NULL,`
`1339`	`1338`	`*prefix = NULL;`