/** Check if we are still in the number */
@Override
public boolean isWordPart(char c) {
// ok, we have to test for scientific notation e.g.: 10.9e10
if ((c == 'x' || c == 'X') && buffer.length() == 1 && buffer.charAt(0) == '0') {
// it is an hexadecimal
buffer.append(c);
isInHexa = true;
return true;
} else {
buffer.append(c);
}
if (isInHexa) {
return Character.isDigit(c)
|| c == 'a'
|| c == 'A'
|| c == 'b'
|| c == 'B'
|| c == 'c'
|| c == 'C'
|| c == 'd'
|| c == 'D'
|| c == 'e'
|| c == 'E'
|| c == 'f'
|| c == 'F';
} else {
return Character.isDigit(c) || c == 'e' || c == '.';
}
}
/**
* Adds spaces after the '#' according to the configured settings. The first char of the buffer
* passed (which is also the output) should always start with a '#'.
*/
public static void formatComment(FormatStd std, FastStringBuffer bufWithComment) {
if (std.spacesInStartComment > 0) {
Assert.isTrue(bufWithComment.charAt(0) == '#');
int len = bufWithComment.length();
char firstCharFound = '\n';
String bufAsString = bufWithComment.toString();
// handle cases where the code-formatting should not take place
if (FileUtils.isPythonShebangLine(bufAsString)) {
return;
}
int spacesFound = 0;
String remainingStringContent = "";
for (int j = 1; j < len; j++) { // start at 1 because 0 should always be '#'
if ((firstCharFound = bufWithComment.charAt(j)) != ' ') {
remainingStringContent = bufAsString.substring(j).trim();
break;
}
spacesFound += 1;
}
if (firstCharFound != '\r'
&& firstCharFound != '\n') { // Only add spaces if it wasn't an empty line.
// handle cases where the code-formatting should not take place
for (String s : BLOCK_COMMENT_SKIPS) {
if (remainingStringContent.endsWith(s) || remainingStringContent.startsWith(s)) {
return;
}
}
int diff = std.spacesInStartComment - spacesFound;
if (diff > 0) {
bufWithComment.insertN(1, ' ', diff);
}
}
}
}
public static Map<Integer, String> loadDictFrom(
FastBufferedReader reader, FastStringBuffer buf, ObjectsPoolMap objectsPoolMap)
throws IOException {
int size = StringUtils.parsePositiveInt(reader.readLine());
HashMap<Integer, String> map = new HashMap<Integer, String>(size + 5);
FastStringBuffer line;
int val = 0;
while (true) {
line = reader.readLine();
if (line == null) {
return map;
} else if (line.startsWith("-- ")) {
if (line.startsWith("-- END DICTIONARY")) {
return map;
}
throw new RuntimeException("Unexpected line: " + line);
} else {
int length = line.length();
// line is str=int
for (int i = 0; i < length; i++) {
char c = line.charAt(i);
if (c == '=') {
val = StringUtils.parsePositiveInt(buf);
buf.clear();
} else {
buf.appendResizeOnExc(c);
}
}
String bufStr = buf.toString();
String interned = objectsPoolMap.get(bufStr);
if (interned == null) {
interned = bufStr;
objectsPoolMap.put(bufStr, bufStr);
}
map.put(val, interned);
buf.clear();
}
}
}
/**
* Formats the contents for when a parenthesis is found (so, go until the closing parens and
* format it accordingly)
*
* @param throwSyntaxError
* @throws SyntaxErrorException
*/
private int formatForPar(
final ParsingUtils parsingUtils,
final char[] cs,
final int i,
final FormatStd std,
final FastStringBuffer buf,
final int parensLevel,
final String delimiter,
boolean throwSyntaxError)
throws SyntaxErrorException {
char c = ' ';
FastStringBuffer locBuf = new FastStringBuffer();
int j = i + 1;
int start = j;
int end = start;
while (j < cs.length && (c = cs[j]) != ')') {
j++;
if (c == '\'' || c == '"') { // ignore comments or multiline comments...
j = parsingUtils.eatLiterals(null, j - 1, std.trimMultilineLiterals) + 1;
end = j;
} else if (c == '#') {
j = parsingUtils.eatComments(null, j - 1) + 1;
end = j;
} else if (c == '(') { // open another par.
if (end > start) {
locBuf.append(cs, start, end - start);
start = end;
}
j =
formatForPar(
parsingUtils,
cs,
j - 1,
std,
locBuf,
parensLevel + 1,
delimiter,
throwSyntaxError)
+ 1;
start = j;
} else {
end = j;
}
}
if (end > start) {
locBuf.append(cs, start, end - start);
start = end;
}
if (c == ')') {
// Now, when a closing parens is found, let's see the contents of the line where that parens
// was found
// and if it's only whitespaces, add all those whitespaces (to handle the following case:
// a(a,
// b
// ) <-- we don't want to change this one.
char c1;
FastStringBuffer buf1 = new FastStringBuffer();
if (locBuf.indexOf('\n') != -1 || locBuf.indexOf('\r') != -1) {
for (int k = locBuf.length();
k > 0 && (c1 = locBuf.charAt(k - 1)) != '\n' && c1 != '\r';
k--) {
buf1.insert(0, c1);
}
}
String formatStr =
formatStr(trim(locBuf).toString(), std, parensLevel, delimiter, throwSyntaxError);
FastStringBuffer formatStrBuf = trim(new FastStringBuffer(formatStr, 10));
String closing = ")";
if (buf1.length() > 0 && PySelection.containsOnlyWhitespaces(buf1.toString())) {
formatStrBuf.append(buf1);
} else if (std.parametersWithSpace) {
closing = " )";
}
if (std.parametersWithSpace) {
if (formatStrBuf.length() == 0) {
buf.append("()");
} else {
buf.append("( ");
buf.append(formatStrBuf);
buf.append(closing);
}
} else {
buf.append('(');
buf.append(formatStrBuf);
buf.append(closing);
}
return j;
} else {
if (throwSyntaxError) {
throw new SyntaxErrorException("No closing ')' found.");
}
// we found no closing parens but we finished looking already, so, let's just add anything
// without
// more formatting...
buf.append('(');
buf.append(locBuf);
return j;
}
}
/**
* This method formats a string given some standard.
*
* @param str the string to be formatted
* @param std the standard to be used
* @param parensLevel the level of the parenthesis available.
* @return a new (formatted) string
* @throws SyntaxErrorException
*/
private String formatStr(
String str, FormatStd std, int parensLevel, String delimiter, boolean throwSyntaxError)
throws SyntaxErrorException {
char[] cs = str.toCharArray();
FastStringBuffer buf = new FastStringBuffer();
// Temporary buffer for some operations. Must always be cleared before it's used.
FastStringBuffer tempBuf = new FastStringBuffer();
ParsingUtils parsingUtils = ParsingUtils.create(cs, throwSyntaxError);
char lastChar = '\0';
for (int i = 0; i < cs.length; i++) {
char c = cs[i];
switch (c) {
case '\'':
case '"':
// ignore literals and multi-line literals, including comments...
i = parsingUtils.eatLiterals(buf, i, std.trimMultilineLiterals);
break;
case '#':
i = handleComment(std, cs, buf, tempBuf, parsingUtils, i);
break;
case ',':
i = formatForComma(std, cs, buf, i, tempBuf);
break;
case '(':
i =
formatForPar(
parsingUtils, cs, i, std, buf, parensLevel + 1, delimiter, throwSyntaxError);
break;
// Things to treat:
// +, -, *, /, %
// ** // << >>
// <, >, !=, <>, <=, >=, //=, *=, /=,
// & ^ ~ |
case '*':
// for *, we also need to treat when it's used in varargs, kwargs and list expansion
boolean isOperator = false;
for (int j = buf.length() - 1; j >= 0; j--) {
char localC = buf.charAt(j);
if (Character.isWhitespace(localC)) {
continue;
}
if (localC == '(' || localC == ',') {
// it's not an operator, but vararg. kwarg or list expansion
}
if (Character.isJavaIdentifierPart(localC)) {
// ok, there's a chance that it can be an operator, but we still have to check
// the chance that it's a wild import
tempBuf.clear();
while (Character.isJavaIdentifierPart(localC)) {
tempBuf.append(localC);
j--;
if (j < 0) {
break; // break while
}
localC = buf.charAt(j);
}
String reversed = tempBuf.reverse().toString();
if (!reversed.equals("import") && !reversed.equals("lambda")) {
isOperator = true;
}
}
if (localC == '\'' || localC == ')' || localC == ']') {
isOperator = true;
}
// If it got here (i.e.: not whitespace), get out of the for loop.
break;
}
if (!isOperator) {
buf.append('*');
break; // break switch
}
// Otherwise, FALLTHROUGH
case '+':
case '-':
if (c == '-' || c == '+') { // could also be *
// handle exponentials correctly: e.g.: 1e-6 cannot have a space
tempBuf.clear();
boolean started = false;
for (int j = buf.length() - 1; ; j--) {
if (j < 0) {
break;
}
char localC = buf.charAt(j);
if (localC == ' ' || localC == '\t') {
if (!started) {
continue;
} else {
break;
}
}
started = true;
if (Character.isJavaIdentifierPart(localC) || localC == '.') {
tempBuf.append(localC);
} else {
break; // break for
}
}
boolean isExponential = true;
String partialNumber = tempBuf.reverse().toString();
int partialLen = partialNumber.length();
if (partialLen < 2 || !Character.isDigit(partialNumber.charAt(0))) {
// at least 2 chars: the number and the 'e'
isExponential = false;
} else {
// first char checked... now, if the last is an 'e', we must leave it together no
// matter what
if (partialNumber.charAt(partialLen - 1) != 'e'
&& partialNumber.charAt(partialLen - 1) != 'E') {
isExponential = false;
}
}
if (isExponential) {
buf.rightTrim();
buf.append(c);
// skip the next whitespaces from the buffer
int initial = i;
do {
i++;
} while (i < cs.length && (c = cs[i]) == ' ' || c == '\t');
if (i > initial) {
i--; // backup 1 because we walked 1 too much.
}
break; // break switch
}
// Otherwise, FALLTHROUGH
}
case '/':
case '%':
case '<':
case '>':
case '!':
case '&':
case '^':
case '~':
case '|':
i = handleOperator(std, cs, buf, parsingUtils, i, c);
c = cs[i];
break;
// check for = and == (other cases that have an = as the operator should already be
// treated)
case '=':
if (i < cs.length - 1 && cs[i + 1] == '=') {
// if == handle as if a regular operator
i = handleOperator(std, cs, buf, parsingUtils, i, c);
c = cs[i];
break;
}
while (buf.length() > 0 && buf.lastChar() == ' ') {
buf.deleteLast();
}
boolean surroundWithSpaces = std.operatorsWithSpace;
if (parensLevel > 0) {
surroundWithSpaces = std.assignWithSpaceInsideParens;
}
// add space before
if (surroundWithSpaces) {
buf.append(' ');
}
// add the operator and the '='
buf.append('=');
// add space after
if (surroundWithSpaces) {
buf.append(' ');
}
i = parsingUtils.eatWhitespaces(null, i + 1);
break;
default:
if (c == '\r' || c == '\n') {
if (lastChar == ',' && std.spaceAfterComma && buf.lastChar() == ' ') {
buf.deleteLast();
}
if (std.trimLines) {
buf.rightTrim();
}
}
buf.append(c);
}
lastChar = c;
}
if (parensLevel == 0 && std.trimLines) {
buf.rightTrim();
}
return buf.toString();
}