Fix #7243 - Some UNICODE characters can lead to wrong CONTAINING evaluation when lower/upper uses different number of bytes in its encoding.

asfernandes · asfernandes · commit 6d594b5ada11 · 2022-08-03T23:01:26.000-03:00
diff --git a/src/jrd/intl_classes.h b/src/jrd/intl_classes.h
@@ -87,7 +87,10 @@ class UpcaseConverter : public PrevConverter
 	UpcaseConverter(MemoryPool& pool, TextType* obj, const UCHAR*& str, SLONG& len)
 		: PrevConverter(pool, obj, str, len)
 	{
-		obj->str_to_upper(len, str, len, tempBuffer.getBuffer(len, false));
+		const auto charSet = obj->getCharSet();
+		const auto bufferSize = len / charSet->minBytesPerChar() * charSet->maxBytesPerChar();
+
+		len = obj->str_to_upper(len, str, bufferSize, tempBuffer.getBuffer(bufferSize, false));
 		str = tempBuffer.begin();
 	}
 

Original file line number	Diff line number	Diff line change
`@@ -87,7 +87,10 @@ class UpcaseConverter : public PrevConverter`
`87`	`87`	`UpcaseConverter(MemoryPool& pool, TextType* obj, const UCHAR*& str, SLONG& len)`
`88`	`88`	`: PrevConverter(pool, obj, str, len)`
`89`	`89`	`{`
`90`		`- obj->str_to_upper(len, str, len, tempBuffer.getBuffer(len, false));`
	`90`	`+ const auto charSet = obj->getCharSet();`
	`91`	`+ const auto bufferSize = len / charSet->minBytesPerChar() * charSet->maxBytesPerChar();`
	`92`	`+`
	`93`	`+ len = obj->str_to_upper(len, str, bufferSize, tempBuffer.getBuffer(bufferSize, false));`
`91`	`94`	`str = tempBuffer.begin();`
`92`	`95`	`}`
`93`	`96`