improving charset implementation maintainability and performance

While reviewing the fix for

6233303 nio ISO2022JP.Encoder is broken unmappableCharacterAction is

I noticed a number of cases of dead code, duplicated code, and unnecessary
code calls in ISO2022JP.Encoder.encodeLoop.

This code would be much more maintainable and better performing if
these kinds of changes were applied to all NIO charsets.

There is no need to check for surrogateness, as far as I can see,
except possibly for better error diagnosis. And certainly there
is no need to check for it unless the char is unmappable, i.e.
encodeDouble fails.

outputSize appears to be unused.

The code for backslash and tilde can be unified.

(These changes are completely untested)

Similar changes can be applied to other encodeLoop methods and
other encodings.

--- /tmp/geta24581 2005-02-28 09:18:30.337098600 -0800
+++ ISO2022_JP.java 2005-02-26 17:17:57.951819000 -0800
@@ -383,131 +383,89 @@
  private CoderResult encodeArrayLoop(CharBuffer src,
  ByteBuffer dst)
  {
  char[] sa = src.array();
  int sp = src.arrayOffset() + src.position();
  int sl = src.arrayOffset() + src.limit();
  assert (sp <= sl);
- sp = (sp <= sl ? sp : sl);
+
  byte[] da = dst.array();
  int dp = dst.arrayOffset() + dst.position();
  int dl = dst.arrayOffset() + dst.limit();
  assert (dp <= dl);
- dp = (dp <= dl ? dp : dl);
 
- int outputSize = 0;
+ int index;
 
  try {
  while (sp < sl) {
- int newMode = currentMode;
  char c = sa[sp];
 
                     if (c <= '\u007F') {
  if (currentMode != ASCII) {
- if (dl - dp < 4)
+ if (dl - dp < 3 + 1)
  return CoderResult.OVERFLOW;
  da[dp++] = (byte)0x1b;
  da[dp++] = (byte)0x28;
  da[dp++] = (byte)0x42;
- da[dp++] = (byte)c;
- newMode = ASCII;
- sp++;
+ currentMode = ASCII;
  }
- else {
- if (dl - dp < 1)
- return CoderResult.OVERFLOW;
- da[dp++] = (byte)c;
- sp++;
- }
+ if (dl - dp < 1)
+ return CoderResult.OVERFLOW;
+ da[dp++] = (byte)c;
+ sp++;
  }
  // Is it a single byte kana?
  else if (c >= 0xff61 && c <= 0xff9f) {
  if (currentMode != JISX0201_1976_KANA) {
- if (dl - dp < 4)
+ if (dl - dp < 3 + 1)
  return CoderResult.OVERFLOW;
  da[dp++] = (byte)0x1b;
  da[dp++] = (byte)0x28;
  da[dp++] = (byte)0x49;
- da[dp++] = (byte)(c - 0xff40);
- newMode = JISX0201_1976_KANA;
- sp++;
- } else {
- if (dl - dp < 1)
+ currentMode = JISX0201_1976_KANA;
+ }
+ if (dl - dp < 1)
+ return CoderResult.OVERFLOW;
+ da[dp++] = (byte)(c - 0xff40);
+ sp++;
+ }
+ else if ((index = encodeDouble(c)) != 0) {
+ if (currentMode != JISX0208_1983) {
+ if (dl - dp < 3 + 2)
  return CoderResult.OVERFLOW;
- da[dp++] = (byte)(c - 0xff40);
- sp++;
+ da[dp++] = (byte)0x1b;
+ da[dp++] = (byte)0x24;
+ da[dp++] = (byte)0x42;
+ currentMode = JISX0208_1983;
  }
+ if (dl - dp < 2)
+ return CoderResult.OVERFLOW;
+ da[dp++] = (byte)(index >> 8);
+ da[dp++] = (byte)(index & 0xff);
+ sp++;
  }
- else if (c == '\u00A5') {
+ else if (c == '\u00A5' /* backslash */ ||
+ c == '\u203E' /* tilde */) {
  if (currentMode != JISX0201_1976) {
- if (dl - dp < 4)
+ if (dl - dp < 3 + 1)
  return CoderResult.OVERFLOW;
  da[dp++] = (byte)0x1b;
  da[dp++] = (byte)0x28;
  da[dp++] = (byte)0x4a;
- da[dp++] = (byte)0x5c;
- newMode = JISX0201_1976;
- sp++;
- } else {
- da[dp++] = (byte)0x5C;
- sp++;
+ currentMode = JISX0201_1976;
  }
- }
- else if (c == '\u203E') { // is it a tilde?
- if (currentMode != JISX0201_1976) {
- if (dl - dp < 4)
- return CoderResult.OVERFLOW;
- da[dp++] = (byte)0x1b;
- da[dp++] = (byte)0x28;
- da[dp++] = (byte)0x4a;
- da[dp++] = (byte)0x7e;
- newMode = JISX0201_1976;
- sp++;
- } else {
- if (dl - dp < 1)
- return CoderResult.OVERFLOW;
- da[dp++] = (byte)0x7e;
- sp++;
- }
- }
- else {
- if (dl - dp < 5)
+ if (dl - dp < 1)
  return CoderResult.OVERFLOW;
- int index = encodeDouble(c);
- if (index != 0) {
- if (currentMode != JISX0208_1983) {
- da[dp++] = (byte)0x1b;
- da[dp++] = (byte)0x24;
- da[dp++] = (byte)0x42;
- da[dp++] = (byte)(index >> 8);
- da[dp++] = (byte)(index & 0xff);
- newMode = JISX0208_1983;
- sp++;
- } else {
- da[dp++] = (byte)(index >> 8);
- da[dp++] = (byte)(index & 0xff);
- sp++;
- }
- }
- else {
- return CoderResult.unmappableForLength(1);
- }
+ da[dp++] = (byte) (c == '\u00A5' ? 0x5c : 0x7e);
+ sp++;
  }
-
- if (Surrogate.is(c)) {
- if (sgp.parse(c, sa, sp, sl) < 0)
- return sgp.error();
- return sgp.unmappableResult();
+ else {
+ return CoderResult.unmappableForLength(1);
  }
-
- if (dl - dp < outputSize)
- return CoderResult.OVERFLOW;
- currentMode = newMode;
-
- }
+ }
  return CoderResult.UNDERFLOW;
  } finally {
  src.position(sp - src.arrayOffset());
  dst.position(dp - dst.arrayOffset());
  }
  }
 

###@###.### 2005-2-28 17:23:47 GMT