/*
* Perform an unescape operation based on char[].
*/
static void unescape(
final char[] text,
final int offset,
final int len,
final Writer writer,
final UriEscapeType escapeType,
final String encoding)
throws IOException {
if (text == null) {
return;
}
final int max = (offset + len);
int readOffset = offset;
for (int i = offset; i < max; i++) {
final char c = text[i];
/*
* Check the need for an unescape operation at this point
*/
if (c != ESCAPE_PREFIX && (c != '+' || !escapeType.canPlusEscapeWhitespace())) {
continue;
}
/*
* At this point we know for sure we will need some kind of unescape, so we
* can increase the offset copy all the contents pending up to this point.
*/
if (i - readOffset > 0) {
writer.write(text, readOffset, (i - readOffset));
}
/*
* Deal with possible '+'-escaped whitespace (application/x-www-form-urlencoded)
*/
if (c == '+') {
// if we reached this point with c == '+', it's escaping a whitespace
writer.write(' ');
readOffset = i + 1;
continue;
}
/*
* ESCAPE PROCESS
* --------------
* If there are more than one percent-encoded/escaped sequences together, we will
* need to unescape them all at once (because they might be bytes --up to 4-- of
* the same char).
*/
// Max possible size will be the remaining amount of chars / 3
final byte[] bytes = new byte[(max - i) / 3];
char aheadC = c;
int pos = 0;
while (((i + 2) < max) && aheadC == ESCAPE_PREFIX) {
bytes[pos++] = parseHexa(text[i + 1], text[i + 2]);
i += 3;
if (i < max) {
aheadC = text[i];
}
}
if (i < max && aheadC == ESCAPE_PREFIX) {
// Incomplete escape sequence!
throw new IllegalArgumentException("Incomplete escaping sequence in input");
}
try {
writer.write(new String(bytes, 0, pos, encoding));
} catch (final UnsupportedEncodingException e) {
throw new IllegalArgumentException(
"Exception while escaping URI: Bad encoding '" + encoding + "'", e);
}
readOffset = i;
}
/*
* -----------------------------------------------------------------------------------------------
* Final cleaning: return the original String object if no unescape was actually needed. Otherwise
* append the remaining escaped text and return.
* -----------------------------------------------------------------------------------------------
*/
if (max - readOffset > 0) {
writer.write(text, readOffset, (max - readOffset));
}
}
/*
* Perform an escape operation, based on char[], according to the specified type
*/
static void escape(
final char[] text,
final int offset,
final int len,
final Writer writer,
final UriEscapeType escapeType,
final String encoding)
throws IOException {
if (text == null || text.length == 0) {
return;
}
final int max = (offset + len);
int readOffset = offset;
for (int i = offset; i < max; i++) {
final int codepoint = Character.codePointAt(text, i);
/*
* Shortcut: most characters will be alphabetic, and we won't need to do anything at
* all for them. No need to use the complete UriEscapeType check system at all.
*/
if (UriEscapeType.isAlpha(codepoint)) {
continue;
}
/*
* Check whether the character is allowed or not
*/
if (escapeType.isAllowed(codepoint)) {
continue;
}
/*
* At this point we know for sure we will need some kind of escape, so we
* can write all the contents pending up to this point.
*/
if (i - readOffset > 0) {
writer.write(text, readOffset, (i - readOffset));
}
if (Character.charCount(codepoint) > 1) {
// This is to compensate that we are actually reading two char[] positions with a single
// codepoint.
i++;
}
readOffset = i + 1;
/*
* -----------------------------------------------------------------------------------------
*
* Peform the real escape
*
* -----------------------------------------------------------------------------------------
*/
final byte[] charAsBytes;
try {
charAsBytes = new String(Character.toChars(codepoint)).getBytes(encoding);
} catch (final UnsupportedEncodingException e) {
throw new IllegalArgumentException(
"Exception while escaping URI: Bad encoding '" + encoding + "'", e);
}
for (final byte b : charAsBytes) {
writer.write('%');
writer.write(printHexa(b));
}
}
/*
* -----------------------------------------------------------------------------------------------
* Final cleaning: return the original String object if no escape was actually needed. Otherwise
* append the remaining unescaped text to the string builder and return.
* -----------------------------------------------------------------------------------------------
*/
if (max - readOffset > 0) {
writer.write(text, readOffset, (max - readOffset));
}
}
