String getScriptText(Charset[] charsetGuess) {
if (getHunks().isEmpty()) {
// If we have no hunks then we can safely assume the entire
// patch is a binary style patch, or a meta-data only style
// patch. Either way the encoding of the headers should be
// strictly 7-bit US-ASCII and the body is either 7-bit ASCII
// (due to the base 85 encoding used for a BinaryHunk) or is
// arbitrary noise we have chosen to ignore and not understand
// (e.g. the message "Binary files ... differ").
//
return extractBinaryString(buf, startOffset, endOffset);
}
if (charsetGuess != null && charsetGuess.length != getParentCount() + 1)
throw new IllegalArgumentException(
MessageFormat.format(
JGitText.get().expectedCharacterEncodingGuesses,
Integer.valueOf(getParentCount() + 1)));
if (trySimpleConversion(charsetGuess)) {
Charset cs = charsetGuess != null ? charsetGuess[0] : null;
if (cs == null) cs = Constants.CHARSET;
try {
return decodeNoFallback(cs, buf, startOffset, endOffset);
} catch (CharacterCodingException cee) {
// Try the much slower, more-memory intensive version which
// can handle a character set conversion patch.
}
}
final StringBuilder r = new StringBuilder(endOffset - startOffset);
// Always treat the headers as US-ASCII; Git file names are encoded
// in a C style escape if any character has the high-bit set.
//
final int hdrEnd = getHunks().get(0).getStartOffset();
for (int ptr = startOffset; ptr < hdrEnd; ) {
final int eol = Math.min(hdrEnd, nextLF(buf, ptr));
r.append(extractBinaryString(buf, ptr, eol));
ptr = eol;
}
final String[] files = extractFileLines(charsetGuess);
final int[] offsets = new int[files.length];
for (final HunkHeader h : getHunks()) h.extractFileLines(r, files, offsets);
return r.toString();
}
private String[] extractFileLines(final Charset[] csGuess) {
final TemporaryBuffer[] tmp = new TemporaryBuffer[getParentCount() + 1];
try {
for (int i = 0; i < tmp.length; i++) tmp[i] = new TemporaryBuffer.LocalFile();
for (final HunkHeader h : getHunks()) h.extractFileLines(tmp);
final String[] r = new String[tmp.length];
for (int i = 0; i < tmp.length; i++) {
Charset cs = csGuess != null ? csGuess[i] : null;
if (cs == null) cs = Constants.CHARSET;
r[i] = RawParseUtils.decode(cs, tmp[i].toByteArray());
}
return r;
} catch (IOException ioe) {
throw new RuntimeException(JGitText.get().cannotConvertScriptToText, ioe);
} finally {
for (final TemporaryBuffer b : tmp) {
if (b != null) b.destroy();
}
}
}
