llstring.cpp

void tokenizeStringToArray(const std::string& data, std::vector<std::string>& output)
{
	output.clear();
	size_t length = data.size();
	
	// tokenize it and put it in the array
	std::string cur_word;
	for(size_t i = 0; i < length; ++i)
	{
		if(data[i] == ':')
		{
			output.push_back(cur_word);
			cur_word.clear();
		}
		else
		{
			cur_word.append(1, data[i]);
		}
	}
	output.push_back(cur_word);
}

void LLStringOps::setupWeekDaysNames(const std::string& data)
{
	tokenizeStringToArray(data,sWeekDayList);
}
void LLStringOps::setupWeekDaysShortNames(const std::string& data)
{
	tokenizeStringToArray(data,sWeekDayShortList);
}
void LLStringOps::setupMonthNames(const std::string& data)
{
	tokenizeStringToArray(data,sMonthList);
}
void LLStringOps::setupMonthShortNames(const std::string& data)
{
	tokenizeStringToArray(data,sMonthShortList);
}
void LLStringOps::setupDayFormat(const std::string& data)
{
	sDayFormat = data;
}


std::string LLStringOps::getDatetimeCode (std::string_view key)
{
	auto iter = datetimeToCodes.find (key);
	if (iter != datetimeToCodes.end())
	{
		return iter->second;
	}
	else
	{
		return std::string();
	}
}

std::string LLStringOps::getReadableNumber(F64 num)
{
    if (fabs(num)>=1e9)
    {
		return absl::StrFormat("%.2lfB", num / 1e9);
    }
    else if (fabs(num)>=1e6)
    {
		return absl::StrFormat("%.2lfM", num / 1e6);
    }
    else if (fabs(num)>=1e3)
    {
		return absl::StrFormat("%.2lfK", num / 1e3);
    }
    else
    {
		return absl::StrFormat("%.2lf", num);
    }
}

namespace LLStringFn
{
	// NOTE - this restricts output to ascii
	void replace_nonprintable_in_ascii(std::basic_string<char>& string, char replacement)
	{
		const char MIN = 0x20;
		std::basic_string<char>::size_type len = string.size();
		for(std::basic_string<char>::size_type ii = 0; ii < len; ++ii)
		{
			if(string[ii] < MIN)
			{
				string[ii] = replacement;
			}
		}
	}


	// NOTE - this restricts output to ascii
	void replace_nonprintable_and_pipe_in_ascii(std::basic_string<char>& str,
									   char replacement)
	{
		const char MIN  = 0x20;
		const char PIPE = 0x7c;
		std::basic_string<char>::size_type len = str.size();
		for(std::basic_string<char>::size_type ii = 0; ii < len; ++ii)
		{
			if( (str[ii] < MIN) || (str[ii] == PIPE) )
			{
				str[ii] = replacement;
			}
		}
	}

	// https://wiki.lindenlab.com/wiki/Unicode_Guidelines has details on
	// allowable code points for XML. Specifically, they are:
	// 0x09, 0x0a, 0x0d, and 0x20 on up.  JC
	std::string strip_invalid_xml(const std::string& instr)
	{
		std::string output;
		output.reserve( instr.size() );
		std::string::const_iterator it = instr.begin();
		while (it != instr.end())
		{
			// Must compare as unsigned for >=
			// Test most likely match first
			const unsigned char c = (unsigned char)*it;
			if (   c >= (unsigned char)0x20   // SPACE
				|| c == (unsigned char)0x09   // TAB
				|| c == (unsigned char)0x0a   // LINE_FEED
				|| c == (unsigned char)0x0d ) // CARRIAGE_RETURN
			{
				output.push_back(c);
			}
			++it;
		}
		return output;
	}

	/**
	 * @brief Replace all control characters (c < 0x20) with replacement in
	 * string.
	 */
	void replace_ascii_controlchars(std::basic_string<char>& string, char replacement)
	{
		const unsigned char MIN = 0x20;
		std::basic_string<char>::size_type len = string.size();
		for(std::basic_string<char>::size_type ii = 0; ii < len; ++ii)
		{
			const unsigned char c = (unsigned char) string[ii];
			if(c < MIN)
			{
				string[ii] = replacement;
			}
		}
	}
}

////////////////////////////////////////////////////////////

// Forward specialization of LLStringUtil::format before use in LLStringUtil::formatDatetime.
template<>
S32 LLStringUtil::format(std::string& s, const format_map_t& substitutions);

//static
template<> 
void LLStringUtil::getTokens(const std::string& instr, std::vector<std::string >& tokens, const std::string& delims)
{
	// Starting at offset 0, scan forward for the next non-delimiter. We're
	// done when the only characters left in 'instr' are delimiters.
	for (std::string::size_type begIdx, endIdx = 0;
		 (begIdx = instr.find_first_not_of (delims, endIdx)) != std::string::npos; )
	{
		// Found a non-delimiter. After that, find the next delimiter.
		endIdx = instr.find_first_of (delims, begIdx);
		if (endIdx == std::string::npos)
		{
			// No more delimiters: this token extends to the end of the string.
			endIdx = instr.length();
		}

		// extract the token between begIdx and endIdx; substr() needs length
		std::string currToken(instr.substr(begIdx, endIdx - begIdx));
		LLStringUtil::trim (currToken);
		tokens.push_back(currToken);
		// next scan past delimiters starts at endIdx
	}
}

template<> 
LLStringUtil::size_type LLStringUtil::getSubstitution(const std::string& instr, size_type& start, std::vector<std::string>& tokens)
{
	const std::string delims (",");
	
	// Find the first [
	size_type pos1 = instr.find('[', start);
	if (pos1 == std::string::npos)
		return std::string::npos;

	//Find the first ] after the initial [
	size_type pos2 = instr.find(']', pos1);
	if (pos2 == std::string::npos)
		return std::string::npos;

	// Find the last [ before ] in case of nested [[]]
	pos1 = instr.find_last_of('[', pos2-1);
	if (pos1 == std::string::npos || pos1 < start)
		return std::string::npos;
	
	getTokens(std::string(instr,pos1+1,pos2-pos1-1), tokens, delims);
	start = pos2+1;
	
	return pos1;
}

// static
template<> 
bool LLStringUtil::simpleReplacement(std::string &replacement, std::string token, const format_map_t& substitutions)
{
	// see if we have a replacement for the bracketed string (without the brackets)
	// test first using has() because if we just look up with operator[] we get back an
	// empty string even if the value is missing. We want to distinguish between 
	// missing replacements and deliberately empty replacement strings.
	format_map_t::const_iterator iter = substitutions.find(token);
	if (iter != substitutions.end())
	{
		replacement = iter->second;
		return true;
	}
	// if not, see if there's one WITH brackets
	iter = substitutions.find(std::string("[" + token + "]"));
	if (iter != substitutions.end())
	{
		replacement = iter->second;
		return true;
	}

	return false;
}

// static
template<> 
bool LLStringUtil::simpleReplacement(std::string &replacement, std::string token, const LLSD& substitutions)
{
	// see if we have a replacement for the bracketed string (without the brackets)
	// test first using has() because if we just look up with operator[] we get back an
	// empty string even if the value is missing. We want to distinguish between 
	// missing replacements and deliberately empty replacement strings.
	if (substitutions.has(token))
	{
		replacement = substitutions[token].asString();
		return true;
	}
	// if not, see if there's one WITH brackets
	else if (substitutions.has(std::string("[" + token + "]")))
	{
		replacement = substitutions[std::string("[" + token + "]")].asString();
		return true;
	}

	return false;
}

//static
template<>
void LLStringUtil::setLocale(std::string inLocale)
{
	sLocale = std::move(inLocale);
};

//static
template<>
std::string LLStringUtil::getLocale(void)
{
	return sLocale;
};

// static
template<> 
void LLStringUtil::formatNumber(std::string& numStr, std::string decimals)
{
	std::stringstream strStream;
	S32 intDecimals = 0;

	convertToS32 (decimals, intDecimals);
	if (!sLocale.empty())
	{
		// std::locale() throws if the locale is unknown! (EXT-7926)
		try
		{
			strStream.imbue(std::locale(sLocale.c_str()));
		} catch (const std::exception &)
		{
			LL_WARNS_ONCE("Locale") << "Cannot set locale to " << sLocale << LL_ENDL;
		}
	}

	if (!intDecimals)
	{
		S32 intStr;

		if (convertToS32(numStr, intStr))
		{
			strStream << intStr;
			numStr = strStream.str();
		}
	}
	else
	{
		F32 floatStr;

		if (convertToF32(numStr, floatStr))
		{
			strStream << std::fixed << std::showpoint << std::setprecision(intDecimals) << floatStr;
			numStr = strStream.str();
		}
	}
}

// static
template<> 
bool LLStringUtil::formatDatetime(std::string& replacement, std::string token,
								  std::string param, S32 secFromEpoch)
{
	if (param == "local")   // local
	{
		secFromEpoch -= LLStringOps::getLocalTimeOffset();
	}
	else if (param != "utc") // slt
	{
		secFromEpoch -= LLStringOps::getPacificTimeOffset();
	}
		
	// if never fell into those two ifs above, param must be utc
	if (secFromEpoch < 0) secFromEpoch = 0;

	std::string code = LLStringOps::getDatetimeCode (token);

	// special case to handle timezone
	if (code == "%Z") {
		if (param == "utc")
		{
			replacement = "GMT";
		}
		else if (param == "local")
		{
			replacement.clear();		// user knows their own timezone
		}
		else
		{
			// "slt" = Second Life Time, which is deprecated.
			// If not utc or user local time, fallback to Pacific time
			replacement = LLStringOps::getPacificDaylightTime() ? "PDT" : "PST";
		}
		return true;
	}

	//EXT-7013
	//few codes are not suppotred by strtime function (example - weekdays for Japanise)
	//so use predefined ones
	
	//if sWeekDayList is not empty than current locale doesn't support
        //weekday name.
	time_t loc_seconds = (time_t) secFromEpoch;
	if(LLStringOps::sWeekDayList.size() == 7 && code == "%A")
	{
		struct tm * gmt = gmtime (&loc_seconds);
		replacement = LLStringOps::sWeekDayList[gmt->tm_wday];
	}
	else if(LLStringOps::sWeekDayShortList.size() == 7 && code == "%a")
	{
		struct tm * gmt = gmtime (&loc_seconds);
		replacement = LLStringOps::sWeekDayShortList[gmt->tm_wday];
	}
	else if(LLStringOps::sMonthList.size() == 12 && code == "%B")
	{
		struct tm * gmt = gmtime (&loc_seconds);
		replacement = LLStringOps::sMonthList[gmt->tm_mon];
	}
	else if( !LLStringOps::sDayFormat.empty() && code == "%d" )
	{
		struct tm * gmt = gmtime (&loc_seconds);
		LLStringUtil::format_map_t args;
		args["[MDAY]"] = absl::StrFormat("%d", gmt->tm_mday);
		replacement = LLStringOps::sDayFormat;
		LLStringUtil::format(replacement, args);
	}
	else if (code == "%-d")
	{
		struct tm * gmt = gmtime (&loc_seconds);
		replacement = absl::StrFormat("%d", gmt->tm_mday); // day of the month without leading zero
	}
	else if( !LLStringOps::sAM.empty() && !LLStringOps::sPM.empty() && code == "%p" )
	{
		struct tm * gmt = gmtime (&loc_seconds);
		if(gmt->tm_hour<12)
		{
			replacement = LLStringOps::sAM;
		}
		else
		{
			replacement = LLStringOps::sPM;
		}
	}
	else
	{
		LLDate datetime((F64)secFromEpoch);
		replacement = datetime.toHTTPDateString(code);
	}

	// *HACK: delete leading zero from hour string in case 'hour12' (code = %I) time format
	// to show time without leading zero, e.g. 08:16 -> 8:16 (EXT-2738).
	// We could have used '%l' format instead, but it's not supported by Windows.
	if(code == "%I" && token == "hour12" && replacement.at(0) == '0')
	{
		replacement = replacement.at(1);
	}

	return !code.empty();
}

// LLStringUtil::format recogizes the following patterns.
// All substitutions *must* be encased in []'s in the input string.
// The []'s are optional in the substitution map.
// [FOO_123]
// [FOO,number,precision]
// [FOO,datetime,format]


// static
template<> 
S32 LLStringUtil::format(std::string& s, const format_map_t& substitutions)
{
	LL_RECORD_BLOCK_TIME(FT_STRING_FORMAT);
	S32 res = 0;

	std::string output;
	std::vector<std::string> tokens;

	std::string::size_type start = 0;
	std::string::size_type prev_start = 0;
	std::string::size_type key_start = 0;
	while ((key_start = getSubstitution(s, start, tokens)) != std::string::npos)
	{
		output += std::string(s, prev_start, key_start-prev_start);
		prev_start = start;
		
		bool found_replacement = false;
		std::string replacement;

		if (tokens.size() == 0)
		{
			found_replacement = false;
		}
		else if (tokens.size() == 1)
		{
			found_replacement = simpleReplacement (replacement, tokens[0], substitutions);
		}
		else if (tokens[1] == "number")
		{
			std::string param = "0";

			if (tokens.size() > 2) param = tokens[2];
			found_replacement = simpleReplacement (replacement, tokens[0], substitutions);
			if (found_replacement) formatNumber (replacement, param);
		}
		else if (tokens[1] == "datetime")
		{
			std::string param;
			if (tokens.size() > 2) param = tokens[2];
			
			format_map_t::const_iterator iter = substitutions.find("datetime");
			if (iter != substitutions.end())
			{
				S32 secFromEpoch = 0;
				BOOL r = LLStringUtil::convertToS32(iter->second, secFromEpoch);
				if (r)
				{
					found_replacement = formatDatetime(replacement, tokens[0], param, secFromEpoch);
				}
			}
		}

		if (found_replacement)
		{
			output += replacement;
			res++;
		}
		else
		{
			// we had no replacement, use the string as is
			// e.g. "hello [MISSING_REPLACEMENT]" or "-=[Stylized Name]=-"
			output += std::string(s, key_start, start-key_start);
		}
		tokens.clear();
	}
	// send the remainder of the string (with no further matches for bracketed names)
	output += std::string(s, start);
	s = output;
	return res;
}

//static
template<> 
S32 LLStringUtil::format(std::string& s, const LLSD& substitutions)
{
	LL_RECORD_BLOCK_TIME(FT_STRING_FORMAT);
	S32 res = 0;

	if (!substitutions.isMap()) 
	{
		return res;
	}

	std::string output;
	std::vector<std::string> tokens;

	std::string::size_type start = 0;
	std::string::size_type prev_start = 0;
	std::string::size_type key_start = 0;
	while ((key_start = getSubstitution(s, start, tokens)) != std::string::npos)
	{
		output += std::string(s, prev_start, key_start-prev_start);
		prev_start = start;
		
		bool found_replacement = false;
		std::string replacement;

		if (tokens.size() == 0)
		{
			found_replacement = false;
		}
		else if (tokens.size() == 1)
		{
			found_replacement = simpleReplacement (replacement, tokens[0], substitutions);
		}
		else if (tokens[1] == "number")
		{
			std::string param = "0";

			if (tokens.size() > 2) param = tokens[2];
			found_replacement = simpleReplacement (replacement, tokens[0], substitutions);
			if (found_replacement) formatNumber (replacement, param);
		}
		else if (tokens[1] == "datetime")
		{
			std::string param;
			if (tokens.size() > 2) param = tokens[2];
			
			S32 secFromEpoch = (S32) substitutions["datetime"].asInteger();
			found_replacement = formatDatetime (replacement, tokens[0], param, secFromEpoch);
		}

		if (found_replacement)
		{
			output += replacement;
			res++;
		}
		else
		{
			// we had no replacement, use the string as is
			// e.g. "hello [MISSING_REPLACEMENT]" or "-=[Stylized Name]=-"
			output += std::string(s, key_start, start-key_start);
		}
		tokens.clear();
	}
	// send the remainder of the string (with no further matches for bracketed names)
	output += std::string(s, start);
	s = output;
	return res;
}

////////////////////////////////////////////////////////////
// Testing

#ifdef _DEBUG

template<class T> 
void LLStringUtilBase<T>::testHarness()
{
	std::string s1;
	
	llassert( s1.c_str() == NULL );
	llassert( s1.size() == 0 );
	llassert( s1.empty() );
	
	std::string s2( "hello");
	llassert( !strcmp( s2.c_str(), "hello" ) );
	llassert( s2.size() == 5 ); 
	llassert( !s2.empty() );
	std::string s3( s2 );

	llassert( "hello" == s2 );
	llassert( s2 == "hello" );
	llassert( s2 > "gello" );
	llassert( "gello" < s2 );
	llassert( "gello" != s2 );
	llassert( s2 != "gello" );

	std::string s4 = s2;
	llassert( !s4.empty() );
	s4.empty();
	llassert( s4.empty() );
	
	std::string s5("");
	llassert( s5.empty() );
	
	llassert( isValidIndex(s5, 0) );
	llassert( !isValidIndex(s5, 1) );
	
	s3 = s2;
	s4 = "hello again";
	
	s4 += "!";
	s4 += s4;
	llassert( s4 == "hello again!hello again!" );
	
	
	std::string s6 = s2 + " " + s2;
	std::string s7 = s6;
	llassert( s6 == s7 );
	llassert( !( s6 != s7) );
	llassert( !(s6 < s7) );
	llassert( !(s6 > s7) );
	
	llassert( !(s6 == "hi"));
	llassert( s6 == "hello hello");
	llassert( s6 < "hi");
	
	llassert( s6[1] == 'e' );
	s6[1] = 'f';
	llassert( s6[1] == 'f' );
	
	s2.erase( 4, 1 );
	llassert( s2 == "hell");
	s2.insert( 0, "y" );
	llassert( s2 == "yhell");
	s2.erase( 1, 3 );
	llassert( s2 == "yl");
	s2.insert( 1, "awn, don't yel");
	llassert( s2 == "yawn, don't yell");
	
	std::string s8 = s2.substr( 6, 5 );
	llassert( s8 == "don't"  );
	
	std::string s9 = "   \t\ntest  \t\t\n  ";
	trim(s9);
	llassert( s9 == "test"  );

	s8 = "abc123&*(ABC";

	s9 = s8;
	toUpper(s9);
	llassert( s9 == "ABC123&*(ABC"  );

	s9 = s8;
	toLower(s9);
	llassert( s9 == "abc123&*(abc"  );


	std::string s10( 10, 'x' );
	llassert( s10 == "xxxxxxxxxx" );

	std::string s11( "monkey in the middle", 7, 2 );
	llassert( s11 == "in" );

	std::string s12;  //empty
	s12 += "foo";
	llassert( s12 == "foo" );

	std::string s13;  //empty
	s13 += 'f';
	llassert( s13 == "f" );
}


#endif  // _DEBUG