/** Handles the case where we found a '#' in the code. */
private int handleComment(
FormatStd std,
char[] cs,
FastStringBuffer buf,
FastStringBuffer tempBuf,
ParsingUtils parsingUtils,
int i) {
if (std.spacesBeforeComment != FormatStd.DONT_HANDLE_SPACES) {
for (int j = i - 1; j >= 0; j--) {
char cj = cs[j];
if (cj == '\t' || cj == ' ') {
continue;
}
// Ok, found a non-whitespace -- if it's not a new line, we're after some
// code, in which case we have to put the configured amount of spaces.
if (cj != '\r' && cj != '\n') {
buf.rightTrim();
buf.appendN(' ', std.spacesBeforeComment);
}
break;
}
}
tempBuf.clear();
i = parsingUtils.eatComments(tempBuf, i);
if (std.trimLines) {
String endLine = "";
if (tempBuf.endsWith("\r\n")) {
endLine = "\r\n";
tempBuf.deleteLastChars(2);
} else if (tempBuf.endsWith('\r') || tempBuf.endsWith('\n')) {
endLine += tempBuf.lastChar();
tempBuf.deleteLast();
}
tempBuf.rightTrim();
tempBuf.append(endLine);
}
formatComment(std, tempBuf);
buf.append(tempBuf);
return i;
}
/**
* We just want to trim whitespaces, not newlines!
*
* @param locBuf the buffer to be trimmed
* @return the same buffer passed as a parameter
*/
private FastStringBuffer rtrim(FastStringBuffer locBuf) {
while (locBuf.length() > 0 && (locBuf.lastChar() == ' ' || locBuf.lastChar() == '\t')) {
locBuf.deleteLast();
}
return locBuf;
}
/**
* Handles having an operator
*
* @param std the coding standard to be used
* @param cs the contents of the string
* @param buf the buffer where the contents should be added
* @param parsingUtils helper to get the contents
* @param i current index
* @param c current char
* @return the new index after handling the operator
*/
private int handleOperator(
FormatStd std, char[] cs, FastStringBuffer buf, ParsingUtils parsingUtils, int i, char c) {
// let's discover if it's an unary operator (~ + -)
boolean isUnaryWithContents = true;
boolean isUnary = false;
boolean changeWhitespacesBefore = true;
if (c == '~' || c == '+' || c == '-') {
// could be an unary operator...
String trimmedLastWord = buf.getLastWord().trim();
isUnary =
trimmedLastWord.length() == 0 || PySelection.ALL_KEYWORD_TOKENS.contains(trimmedLastWord);
if (!isUnary) {
for (char itChar : buf.reverseIterator()) {
if (itChar == ' ' || itChar == '\t') {
continue;
}
switch (itChar) {
case '[':
case '{':
case '=':
changeWhitespacesBefore = false;
case '(':
case ':':
isUnaryWithContents = false;
case '>':
case '<':
case '-':
case '+':
case '~':
case '*':
case '/':
case '%':
case '!':
case '&':
case '^':
case '|':
case ',':
isUnary = true;
}
break;
}
} else {
isUnaryWithContents = buf.length() > 0;
}
}
if (!isUnary) {
// We don't want to change whitespaces before in a binary operator that is in a new line.
for (char ch : buf.reverseIterator()) {
if (!Character.isWhitespace(ch)) {
break;
}
if (ch == '\r' || ch == '\n') {
changeWhitespacesBefore = false;
break;
}
}
}
if (changeWhitespacesBefore) {
while (buf.length() > 0 && (buf.lastChar() == ' ' || buf.lastChar() == ' ')) {
buf.deleteLast();
}
}
boolean surroundWithSpaces = std.operatorsWithSpace;
if (changeWhitespacesBefore) {
// add spaces before
if (isUnaryWithContents && surroundWithSpaces) {
buf.append(' ');
}
}
char localC = c;
char prev = '\0';
boolean backOne = true;
while (isOperatorPart(localC, prev)) {
buf.append(localC);
prev = localC;
i++;
if (i == cs.length) {
break;
}
localC = cs[i];
if (localC == '=') {
// when we get to an assign, we have found a full stmt (with assign) -- e.g.: a \\= a += a
// ==
buf.append(localC);
backOne = false;
break;
}
}
if (backOne) {
i--;
}
// add space after only if it's not unary
if (!isUnary && surroundWithSpaces) {
buf.append(' ');
}
i = parsingUtils.eatWhitespaces(null, i + 1);
return i;
}
/**
* 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();
}