+/*
+The software in this package is distributed under the GNU General
+Public License version 2 (with a special exception described below).
+
+A copy of GNU General Public License (GPL) is included in this distribution,
+in the file COPYING.GPL.
+
+As a special exception, if other files instantiate templates or use macros
+or inline functions from this file, or you compile this file and link it
+with other works to produce a work based on this file, this file
+does not by itself cause the resulting work to be covered
+by the GNU General Public License.
+
+However the source code for this file must still be made available
+in accordance with section (3) of the GNU General Public License.
+
+This exception does not invalidate any other reasons why a work based
+on this file might be covered by the GNU General Public License.
+*/
/** @file
*
* (c) Copyright 2007-2008 by Intra2net AG
- *
- * info@intra2net.com
*/
#include <iostream>
#include <sstream>
#include <stdexcept>
#include <algorithm>
+#include <cmath> // for round()
#include <wchar.h>
#include <stdlib.h>
#include <iconv.h>
#include <i18n.h>
+#include <boost/numeric/conversion/cast.hpp>
+#include <boost/foreach.hpp>
+
#include <stringfunc.hxx>
using namespace std;
} // eo join_string(const std::list< std::string >&,const std::string&)
+/** @brief same as join_string for list, except uses a vector */
+std::string join_string(
+ const std::vector< std::string >& parts,
+ const std::string& delimiter
+)
+{
+ std::string result;
+ if (! parts.empty() )
+ {
+ std::vector< std::string >::const_iterator it= parts.begin();
+ result = *it;
+ while ( ++it != parts.end() )
+ {
+ result+= delimiter;
+ result+= *it;
+ }
+ }
+ return result;
+} // eo join_string(const std::vector< std::string >&,const std::string&)
+
+
/*
** conversions
char *in = (char *)utf8string.c_str();
char *out = (char*) buf;
- if (iconv(utf82wstr, &in, &in_size, &out, &out_size) == -1)
+ if (iconv(utf82wstr, &in, &in_size, &out, &out_size) == (size_t)-1)
throw runtime_error("error converting char encodings");
buf[ ( (utf8string.size()+1)*sizeof(wchar_t)-out_size) /sizeof(wchar_t) ]=0;
string output;
vector<pair<string,bool> >::const_iterator token, tokens_end = tokenized.end();
- for (token = tokenized.begin(); token != tokens_end; token++)
+ for (token = tokenized.begin(); token != tokens_end; ++token)
if (!token->second)
output += token->first;
string output;
vector<pair<string,bool> >::const_iterator token, tokens_end = tokenized.end();
- for (token = tokenized.begin(); token != tokens_end; token++)
+ for (token = tokenized.begin(); token != tokens_end; ++token)
{
// keep HTML tags as they are
if (token->second)
return str;
} // eo html_entities(std::string)
+// convert HTML entities to something that can be viewed on a basic text console (restricted to ASCII-7)
+string html_entities_to_console(std::string str)
+{
+ // Normal chars
+ replace_all (str, "&", "&");
+ replace_all (str, "<", "<");
+ replace_all (str, ">", ">");
+ replace_all (str, """, "\"");
+ replace_all (str, "'", "'");
+ replace_all (str, "/", "/");
+
+ // Umlauts
+ replace_all (str, "ä", "ae");
+ replace_all (str, "ö", "oe");
+ replace_all (str, "ü", "ue");
+ replace_all (str, "Ä", "Ae");
+ replace_all (str, "Ö", "Oe");
+ replace_all (str, "Ü", "Ue");
+
+ // Misc
+ replace_all (str, "ß", "ss");
+
+ return str;
+}
+
+// find_html_comments + remove_html_comments(str, comments)
+void remove_html_comments(string &str)
+{
+ vector<CommentZone> comments;
+ find_html_comments(str, comments);
+ remove_html_comments(str, comments);
+}
+
+// find all html comments, behaving correctly if they are nested; ignores comment tags ("<!--FOO .... BAR-->")
+// If there are invalid comments ("-->" before "<!--" or different number of closing and opening tags),
+// then the unknown index of corresponding start/end tag will be represented by a string::npos
+// Indices are from start of start tag until first index after closing tag
+void find_html_comments(const std::string &str, vector<CommentZone> &comments)
+{
+ static const string START = "<!--";
+ static const string CLOSE = "-->";
+ static const string::size_type START_LEN = START.length();
+ static const string::size_type CLOSE_LEN = CLOSE.length();
+
+ // in order to find nested comments, need either recursion or a stack
+ vector<string::size_type> starts; // stack of start tags
+
+ string::size_type pos = 0;
+ string::size_type len = str.length();
+ string::size_type next_start, next_close;
+
+ while (pos < len) // not really needed but just in case
+ {
+ next_start = str.find(START, pos);
+ next_close = str.find(CLOSE, pos);
+
+ if ( (next_start == string::npos) && (next_close == string::npos) )
+ break; // we are done
+
+ else if ( (next_start == string::npos) || (next_close < next_start) ) // close one comment (pop)
+ {
+ if (starts.empty()) // closing tag without a start
+ comments.push_back(CommentZone(string::npos, next_close+CLOSE_LEN));
+ else
+ {
+ comments.push_back(CommentZone(starts.back(), next_close+CLOSE_LEN));
+ starts.pop_back();
+ }
+ pos = next_close + CLOSE_LEN;
+ }
+
+ else if ( (next_close == string::npos) || (next_start < next_close) ) // start a new comment (push)
+ {
+ starts.push_back(next_start);
+ pos = next_start + START_LEN;
+ }
+ }
+
+ // add comments that have no closing tag from back to front (important for remove_html_comments!)
+ while (!starts.empty())
+ {
+ comments.push_back(CommentZone(starts.back(), string::npos));
+ starts.pop_back();
+ }
+}
+
+// remove all html comments foundby find_html_comments
+void remove_html_comments(std::string &str, const vector<CommentZone> &comments)
+{
+ // remember position where last removal started
+ string::size_type last_removal_start = str.length();
+
+ // Go from back to front to not mess up indices.
+ // This requires that bigger comments, that contain smaller comments, come AFTER
+ // the small contained comments in the comments vector (i.e. comments are ordered by
+ // their closing tag, not their opening tag). This is true for results from find_html_comments
+ BOOST_REVERSE_FOREACH(const CommentZone &comment, comments)
+ {
+ if (comment.first == string::npos)
+ {
+ str = str.replace(0, comment.second, ""); // comment starts "before" str --> delete from start
+ break; // there can be no more
+ }
+ else if (comment.first >= last_removal_start)
+ {
+ continue; // this comment is inside another comment that we have removed already
+ }
+ else if (comment.second == string::npos) // comment ends "after" str --> delete until end
+ {
+ str = str.replace(comment.first, string::npos, "");
+ last_removal_start = comment.first;
+ }
+ else
+ {
+ str = str.replace(comment.first, comment.second-comment.first, "");
+ last_removal_start = comment.first;
+ }
+ }
+}
bool replace_all(string &base, const char *ist, const char *soll)
{
return found_ist;
}
+/**
+ * @brief replaces all characters that could be problematic or impose a security risk when being logged
+ * @param str the original string
+ * @param replace_with the character to replace the unsafe chars with
+ * @return a string that is safe to send to syslog or other logfiles
+ *
+ * All chars between 0x20 (space) and 0x7E (~) (including) are considered safe for logging.
+ * See e.g. RFC 5424, section 8.2 or the posix character class "printable".
+ * This eliminates all possible problems with NUL, control characters, 8 bit chars, UTF8.
+ *
+ */
+std::string sanitize_for_logging(const std::string &str, const char replace_with)
+{
+ std::string output=str;
+
+ const string::size_type len = output.size();
+ for (std::string::size_type p=0; p < len; p++)
+ if (output[p] < 0x20 || output[p] > 0x7E)
+ output[p]=replace_with;
+
+ return output;
+}
+
#if 0
string to_lower(const string &src)
{
}
#endif
+const int MAX_UNIT_FORMAT_SYMBOLS = 6;
+
+const string shortUnitFormatSymbols[MAX_UNIT_FORMAT_SYMBOLS] = {
+ " B",
+ " KB",
+ " MB",
+ " GB",
+ " TB",
+ " PB"
+};
+
+const string longUnitFormatSymbols[MAX_UNIT_FORMAT_SYMBOLS] = {
+ i18n_noop(" Bytes"),
+ i18n_noop(" KBytes"),
+ i18n_noop(" MBytes"),
+ i18n_noop(" GBytes"),
+ i18n_noop(" TBytes"),
+ i18n_noop(" PBytes")
+};
+
+
+long double rounding_upwards(
+ const long double number,
+ const int rounding_multiplier
+)
+{
+ long double rounded_number;
+ rounded_number = number * rounding_multiplier;
+ rounded_number += 0.5;
+ rounded_number = (int64_t) (rounded_number);
+ rounded_number = (long double) (rounded_number) / (long double) (rounding_multiplier);
+
+ return rounded_number;
+}
+
+
string nice_unit_format(
const int64_t input,
- const UnitSystem system
+ const UnitFormat format,
+ const UnitBase base
)
{
+ // select the system of units (decimal or binary)
int multiple = 0;
-
- if (system == US_SI)
+ if (base == UnitBase1000)
{
multiple = 1000;
}
}
long double size = input;
- int sizecount = 0;
+ // check the size of the input number to fit in the appropriate symbol
+ int sizecount = 0;
while (size > multiple)
{
size = size / multiple;
sizecount++;
+
+ // rollback to the previous values and stop the loop when cannot
+ // represent the number length.
+ if (sizecount >= MAX_UNIT_FORMAT_SYMBOLS)
+ {
+ size = size * multiple;
+ sizecount--;
+ break;
+ }
}
- long double tmp; // round
- tmp = size * 10;
- tmp += 0.5;
- tmp = (int64_t) (tmp);
- tmp = (long double) (tmp) / (long double) (10);
- size = tmp;
+ // round the input number "half up" to multiples of 10
+ const int rounding_multiplier = 10;
+ size = rounding_upwards(size, rounding_multiplier);
+ // format the input number, placing the appropriate symbol
ostringstream out;
-
out.setf (ios::fixed);
- out.precision (2);
- switch (sizecount)
+ if (format == ShortUnitFormat)
{
- case 0:
- out << size << i18n (" Bytes");
- break;
- case 1:
- out << size << i18n (" KBytes");
- break;
- case 2:
- out << size << i18n (" MBytes");
- break;
- case 3:
- out << size << i18n (" GBytes");
- break;
- case 4:
- out << size << i18n (" TBytes");
- break;
- case 5:
- out << size << i18n (" PBytes");
- break;
- case 6:
- out << size << i18n (" EBytes");
- break;
- case 7:
- out << size << i18n (" ZBytes");
- break;
- case 8:
- out << size << i18n (" YBytes");
- break;
- default:
- out << size << "*10^" << (sizecount*3)<< i18n (" Bytes");
- break;
+ out.precision(1);
+ out << size << i18n( shortUnitFormatSymbols[sizecount].c_str() );
+ }
+ else
+ {
+ out.precision (2);
+ out << size << i18n( longUnitFormatSymbols[sizecount].c_str() );
}
return out.str();
} // eo nice_unit_format(int input)
+string nice_unit_format(
+ const double input,
+ const UnitFormat format,
+ const UnitBase base
+)
+{
+ // round as double and cast to int64_t
+ // cast raised overflow error near max val of int64_t (~9.2e18, see unittest)
+ int64_t input_casted_and_rounded =
+ boost::numeric_cast<int64_t>( round(input) );
+
+ // now call other
+ return nice_unit_format( input_casted_and_rounded, format, base );
+} // eo nice_unit_format(double input)
+
+
string escape(const string &s)
{
string out(s);
{
string output = "'";
string::const_iterator it, it_end = input.end();
- for (it = input.begin(); it != it_end; it++)
+ for (it = input.begin(); it != it_end; ++it)
{
if ( (*it) == '\'')
output += "'\\'";