/*
- * Copyright 2012 Facebook, Inc.
+ * Copyright 2013 Facebook, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
#ifndef FOLLY_BASE_STRING_H_
#define FOLLY_BASE_STRING_H_
+#include <exception>
#include <string>
#include <boost/type_traits.hpp>
return out;
}
+/**
+ * URI-escape a string. Appends the result to the output string.
+ *
+ * Alphanumeric characters and other characters marked as "unreserved" in RFC
+ * 3986 ( -_.~ ) are left unchanged. In PATH mode, the forward slash (/) is
+ * also left unchanged. In QUERY mode, spaces are replaced by '+'. All other
+ * characters are percent-encoded.
+ */
+enum class UriEscapeMode : unsigned char {
+ // The values are meaningful, see generate_escape_tables.py
+ ALL = 0,
+ QUERY = 1,
+ PATH = 2
+};
+template <class String>
+void uriEscape(StringPiece str,
+ String& out,
+ UriEscapeMode mode = UriEscapeMode::ALL);
+
+/**
+ * Similar to uriEscape above, but returns the escaped string.
+ */
+template <class String>
+String uriEscape(StringPiece str, UriEscapeMode mode = UriEscapeMode::ALL) {
+ String out;
+ uriEscape(str, out, mode);
+ return out;
+}
+
+/**
+ * URI-unescape a string. Appends the result to the output string.
+ *
+ * In QUERY mode, '+' are replaced by space. %XX sequences are decoded if
+ * XX is a valid hex sequence, otherwise we throw invalid_argument.
+ */
+template <class String>
+void uriUnescape(StringPiece str,
+ String& out,
+ UriEscapeMode mode = UriEscapeMode::ALL);
+
+/**
+ * Similar to uriUnescape above, but returns the unescaped string.
+ */
+template <class String>
+String uriUnescape(StringPiece str, UriEscapeMode mode = UriEscapeMode::ALL) {
+ String out;
+ uriUnescape(str, out, mode);
+ return out;
+}
+
/**
* stringPrintf is much like printf but deposits its result into a
* string. Two signatures are supported: the first simply returns the
*
* Current types are:
* PRETTY_TIME - s, ms, us, ns, etc.
- * PRETTY_BYTES - kb, MB, GB, etc (goes up by 2^10 = 1024 each time)
- * PRETTY_BYTES_METRIC - kb, MB, GB, etc (goes up by 10^3 = 1000 each time)
+ * PRETTY_BYTES_METRIC - kB, MB, GB, etc (goes up by 10^3 = 1000 each time)
+ * PRETTY_BYTES - kB, MB, GB, etc (goes up by 2^10 = 1024 each time)
+ * PRETTY_BYTES_IEC - KiB, MiB, GiB, etc
* PRETTY_UNITS_METRIC - k, M, G, etc (goes up by 10^3 = 1000 each time)
* PRETTY_UNITS_BINARY - k, M, G, etc (goes up by 2^10 = 1024 each time)
+ * PRETTY_UNITS_BINARY_IEC - Ki, Mi, Gi, etc
*
* @author Mark Rabkin <mrabkin@fb.com>
*/
enum PrettyType {
PRETTY_TIME,
- PRETTY_BYTES,
+
PRETTY_BYTES_METRIC,
+ PRETTY_BYTES_BINARY,
+ PRETTY_BYTES = PRETTY_BYTES_BINARY,
+ PRETTY_BYTES_BINARY_IEC,
+ PRETTY_BYTES_IEC = PRETTY_BYTES_BINARY_IEC,
+
PRETTY_UNITS_METRIC,
PRETTY_UNITS_BINARY,
+ PRETTY_UNITS_BINARY_IEC,
PRETTY_NUM_TYPES
};
return folly::to<fbstring>(demangle(typeid(e)), ": ", e.what());
}
+inline fbstring exceptionStr(std::exception_ptr ep) {
+ try {
+ std::rethrow_exception(ep);
+ } catch (const std::exception& e) {
+ return exceptionStr(e);
+ } catch (...) {
+ return "<unknown exception>";
+ }
+}
+
/*
* Split a string into a list of tokens by delimiter.
*
* The split interface here supports different output types, selected
* at compile time: StringPiece, fbstring, or std::string. If you are
* using a vector to hold the output, it detects the type based on
- * what your vector contains.
+ * what your vector contains. If the output vector is not empty, split
+ * will append to the end of the vector.
*
* You can also use splitTo() to write the output to an arbitrary
* OutputIterator (e.g. std::inserter() on a std::set<>), in which
* case you have to tell the function the type. (Rationale:
* OutputIterators don't have a value_type, so we can't detect the
- * type in split without being told.)
+ * type in splitTo without being told.)
*
* Examples:
*
* folly::splitTo<StringPiece>(":", "asd:bsd:asd:csd",
* std::inserter(s, s.begin()));
*
- * Split also takes a flag (ignoreEmpty) that indicates whether
- * adjacent tokens should be treated as one separator or not. Note
- * that unlikely strtok() the default is to treat them as separators.
+ * Split also takes a flag (ignoreEmpty) that indicates whether adjacent
+ * delimiters should be treated as one single separator (ignoring empty tokens)
+ * or not (generating empty tokens).
*/
template<class Delim, class String, class OutputType>
OutputIterator out,
bool ignoreEmpty = false);
+/*
+ * Split a string into a fixed number of pieces by delimiter. Returns 'true' if
+ * the fields were all successfully populated.
+ *
+ * Example:
+ *
+ * folly::StringPiece name, key, value;
+ * if (folly::split('\t', line, name, key, value))
+ * ...
+ *
+ * The 'exact' template paremeter specifies how the function behaves when too
+ * many fields are present in the input string. When 'exact' is set to its
+ * default value of 'true', a call to split will fail if the number of fields in
+ * the input string does not exactly match the number of output parameters
+ * passed. If 'exact' is overridden to 'false', all remaining fields will be
+ * stored, unsplit, in the last field, as shown below:
+ *
+ * folly::StringPiece x, y.
+ * if (folly::split<false>(':', "a:b:c", x, y))
+ * assert(x == "a" && y == "b:c");
+ */
+template<bool exact = true,
+ class Delim,
+ class... StringPieces>
+bool split(const Delim& delimiter,
+ StringPiece input,
+ StringPiece& outHead,
+ StringPieces&... outTail);
+
+/*
+ * Join list of tokens.
+ *
+ * Stores a string representation of tokens in the same order with
+ * deliminer between each element.
+ */
+
+template <class Delim, class Iterator, class String>
+void join(const Delim& delimiter,
+ Iterator begin,
+ Iterator end,
+ String& output);
+
+template <class Delim, class Container, class String>
+void join(const Delim& delimiter,
+ const Container& container,
+ String& output) {
+ join(delimiter, container.begin(), container.end(), output);
+}
+
+template <class Delim, class Container>
+std::string join(const Delim& delimiter,
+ const Container& container) {
+ std::string output;
+ join(delimiter, container.begin(), container.end(), output);
+ return output;
+}
+
} // namespace folly
// Hash functions for string and fbstring usable with e.g. hash_map