static void testMonthParsing() {
const char *months[] = {
"Jan/", "Feb/", "Mar/", "Apr/", "May/", "Jun/",
"Jul/", "Aug/", "Sep/", "Oct/", "Nov/", "Dec/"
};
for(int i=0; i<12; i++) {
for(int j=0; j<10; j++) {
int rv=parseMonth(months[i]);
if(i != rv) {
std::cerr << "Expected " << i << " for "
<< months[i] << " got " << rv << std::endl;
abort();
}
}
}
for(int j=0; j<10; j++) {
for(int i=0; i<12; i++) {
int rv=parseMonth(months[i]);
if(i != rv) {
std::cerr << "Expected " << i << " for "
<< months[i] << " got " << rv << std::endl;
abort();
}
}
}
}
const char* parseTime(const char* start, const char* const end, unsigned& t)
{
t = 0;
tm tm;
auto comma = std::find(start, end, ',');
if (comma != end) {
start = comma + 1;
auto hyphen = std::find(start, end, '-');
if (hyphen != end) {
// RFC 850: Weekday, 00-Mon-00 00:00:00 GMT
start = parseDigits(start, end, tm.tm_mday);
start = parseMonth(++start, end, tm.tm_mon);
start = parseDigits(++start, end, tm.tm_year);
start = parseDigits(start, end, tm.tm_hour);
start = parseDigits(++start, end, tm.tm_min);
start = parseDigits(++start, end, tm.tm_sec);
if (tm.tm_year < 70)
tm.tm_year += 100;
} else {
// RFC 1123: Wkd, 00 Mon 0000 00:00:00 GMT
start = parseDigits(start, end, tm.tm_mday);
start = parseMonth(start, end, tm.tm_mon);
start = parseDigits(start, end, tm.tm_year);
tm.tm_year -= 1900;
start = parseDigits(start, end, tm.tm_hour);
start = parseDigits(++start, end, tm.tm_min);
start = parseDigits(++start, end, tm.tm_sec);
}
} else {
// asctime: Wkd Mon 00 00:00:00 0000 GMT
start = skipSpace(start, end);
start = std::find(start, end, ' ');
if (start == end)
return end;
start = parseMonth(start, end, tm.tm_mon);
start = parseDigits(start, end, tm.tm_mday);
start = parseDigits(start, end, tm.tm_hour);
start = parseDigits(++start, end, tm.tm_min);
start = parseDigits(++start, end, tm.tm_sec);
start = parseDigits(start, end, tm.tm_year);
tm.tm_year -= 1900;
}
if (tm.tm_sec < 0 || 59 < tm.tm_sec ||
tm.tm_min < 0 || 59 < tm.tm_min ||
tm.tm_hour < 0 || 23 < tm.tm_hour ||
tm.tm_mday < 1 || 31 < tm.tm_mday ||
tm.tm_mon < 0 || 11 < tm.tm_mon ||
tm.tm_year < 70 || 120 < tm.tm_year) {
return start;
}
t = timegm(&tm);
return start;
}
CommandBase* ParserList::makeCmdListSpecificMonth(string monthStr) {
int parsedMonth;
// Parse month, if month is valid
try {
parsedMonth = parseMonth(monthStr);
} catch (invalid_argument) {
return NULL;
}
// Ensure that the month is the coming month, and not a month which
// has already passed
date currentDate = second_clock::local_time().date();
date startDate = date(currentDate.year(), parsedMonth, 1);
if (currentDate.month() > parsedMonth) {
startDate += years(1);
}
ptime startTime = ptime(startDate);
ptime endTime = getLastOfMonth(startTime);
return makeCmdListPeriod(startTime, endTime);
}
bool DateComponents::parseDate(const String& src, unsigned start, unsigned& end)
{
unsigned index;
if (!parseMonth(src, start, index))
return false;
// '-' and 2-digits are needed.
if (index + 2 >= src.length())
return false;
if (src[index] != '-')
return false;
++index;
int day;
if (!toInt(src, index, 2, day) || day < 1 || day > maxDayOfMonth(m_year, m_month))
return false;
if (!withinHTMLDateLimits(m_year, m_month, day))
return false;
m_monthDay = day;
end = index + 2;
m_type = Date;
return true;
}
void DateTimeParser::parse(const std::string& fmt, const std::string& str, DateTime& dateTime, int& timeZoneDifferential)
{
int year = 0;
int month = 0;
int day = 0;
int hour = 0;
int minute = 0;
int second = 0;
int millis = 0;
int micros = 0;
int tzd = 0;
std::string::const_iterator it = str.begin();
std::string::const_iterator end = str.end();
std::string::const_iterator itf = fmt.begin();
std::string::const_iterator endf = fmt.end();
while (itf != endf && it != end)
{
if (*itf == '%')
{
if (++itf != endf)
{
switch (*itf)
{
case 'w':
case 'W':
while (it != end && std::isspace(*it)) ++it;
while (it != end && std::isalpha(*it)) ++it;
break;
case 'b':
case 'B':
month = parseMonth(it, end);
break;
case 'd':
case 'e':
case 'f':
SKIP_JUNK();
PARSE_NUMBER_N(day, 2);
break;
case 'm':
case 'n':
case 'o':
SKIP_JUNK();
PARSE_NUMBER_N(month, 2);
break;
case 'y':
SKIP_JUNK();
PARSE_NUMBER_N(year, 2);
if (year >= 70)
year += 1900;
else
year += 2000;
break;
case 'Y':
SKIP_JUNK();
PARSE_NUMBER_N(year, 4);
break;
case 'H':
case 'h':
SKIP_JUNK();
PARSE_NUMBER_N(hour, 2);
break;
case 'a':
case 'A':
hour = parseAMPM(it, end, hour);
break;
case 'M':
SKIP_JUNK();
PARSE_NUMBER_N(minute, 2);
break;
case 'S':
SKIP_JUNK();
PARSE_NUMBER_N(second, 2);
break;
case 'i':
SKIP_JUNK();
PARSE_NUMBER_N(millis, 3);
break;
case 'c':
SKIP_JUNK();
PARSE_NUMBER_N(millis, 1);
millis *= 100;
break;
case 'F':
SKIP_JUNK();
PARSE_NUMBER_N(millis, 3);
PARSE_NUMBER_N(micros, 3);
break;
case 'z':
case 'Z':
tzd = parseTZD(it, end);
break;
}
++itf;
}
}
else ++itf;
}
if (month == 0) month = 1;
if (day == 0) day = 1;
if (DateTime::isValid(year, month, day, hour, minute, second, millis, micros))
dateTime.assign(year, month, day, hour, minute, second, millis, micros);
else
throw SyntaxException("date/time component out of range");
timeZoneDifferential = tzd;
}
time_t LogFile::parseTimestamp()
{
const char *p;
assert(!line.empty());
timestamp=-1;
p=line.c_str();
try {
/* The shortest line I can parse is about 32 characters. */
if(line.length() < 32) {
/* This is a broken entry */
fprintf(stderr, "Broken log entry (too short): %s\n", p);
} else if(index(p, '[') != NULL) {
struct tm tm;
memset(&tm, 0x00, sizeof(tm));
p=index(p, '[');
/* Input validation */
if(p == NULL || line.length() < 32) {
std::cerr << "Invalid log line: " << line << std::endl;
throw BadTimestamp();
}
/* fprintf(stderr, "**** Parsing %s\n", p); */
p++;
tm.tm_mday=atoi(p);
p+=3;
tm.tm_mon=parseMonth(p);
p+=4;
tm.tm_year=atoi(p);
p+=5;
tm.tm_hour=atoi(p);
p+=3;
tm.tm_min=atoi(p);
p+=3;
tm.tm_sec=atoi(p);
/* Make sure it still looks like CLF */
if(p[2] != ' ') {
std::cerr << "log line is starting to not look like CLF: "
<< line
<< std::endl;
throw BadTimestamp();
}
tm.tm_year-=1900;
/* Let mktime guess the timezone */
tm.tm_isdst=-1;
timestamp=mktime(&tm);
} else {
fprintf(stderr, "Unknown log format: %s\n", p);
}
} catch(BadTimestamp e) {
// Damn.
}
if(timestamp < 0) {
std::cerr << "* Error parsing timestamp from " << line << std::endl;
}
return(timestamp);
}
