moved date/time handling functions to their own object

1 files changed, 628 insertions, 0 deletions

diff --git a/datetime.c b/datetime.c
new file mode 100644
index 00000000..9ecd32b3
--- /dev/null
+++ b/datetime.c
@@ -0,0 +1,628 @@
+/* The datetime object. It contains date and time related functions.
+ *
+ * Module begun 2008-03-05 by Rainer Gerhards, based on some code
+ * from syslogd.c. The main intension was to move code out of syslogd.c
+ * in a useful manner. It is still undecided if all functions will continue
+ * to stay here or some will be moved into parser modules (once we have them).
+ *
+ * Copyright 2008 Rainer Gerhards and Adiscon GmbH.
+ *
+ * This file is part of rsyslog.
+ *
+ * Rsyslog is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * Rsyslog is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with Rsyslog.  If not, see <http://www.gnu.org/licenses/>.
+ *
+ * A copy of the GPL can be found in the file "COPYING" in this distribution.
+ */
+
+#include "config.h"
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdarg.h>
+#include <ctype.h>
+#include <assert.h>
+#ifdef HAVE_SYS_TIME_H
+#	include <sys/time.h>
+#endif
+
+#include "rsyslog.h"
+#include "obj.h"
+#include "datetime.h"
+#include "sysvar.h"
+#include "srUtils.h"
+#include "stringbuf.h"
+#include "errmsg.h"
+
+/* static data */
+DEFobjStaticHelpers
+DEFobjCurrIf(errmsg)
+
+
+/* ------------------------------ methods ------------------------------ */
+
+
+/**
+ * Get the current date/time in the best resolution the operating
+ * system has to offer (well, actually at most down to the milli-
+ * second level.
+ *
+ * The date and time is returned in separate fields as this is
+ * most portable and removes the need for additional structures
+ * (but I have to admit it is somewhat "bulky";)).
+ *
+ * Obviously, all caller-provided pointers must not be NULL...
+ */
+void getCurrTime(struct syslogTime *t)
+{
+	struct timeval tp;
+	struct tm *tm;
+	struct tm tmBuf;
+	long lBias;
+#	if defined(__hpux)
+	struct timezone tz;
+#	endif
+
+	assert(t != NULL);
+#	if defined(__hpux)
+		/* TODO: check this: under HP UX, the tz information is actually valid
+		 * data. So we need to obtain and process it there.
+		 */
+		gettimeofday(&tp, &tz);
+#	else
+		gettimeofday(&tp, NULL);
+#	endif
+	tm = localtime_r((time_t*) &(tp.tv_sec), &tmBuf);
+
+	t->year = tm->tm_year + 1900;
+	t->month = tm->tm_mon + 1;
+	t->day = tm->tm_mday;
+	t->hour = tm->tm_hour;
+	t->minute = tm->tm_min;
+	t->second = tm->tm_sec;
+	t->secfrac = tp.tv_usec;
+	t->secfracPrecision = 6;
+
+#	if __sun
+		/* Solaris uses a different method of exporting the time zone.
+		 * It is UTC - localtime, which is the opposite sign of mins east of GMT.
+		 */
+		lBias = -(daylight ? altzone : timezone);
+#	elif defined(__hpux)
+		lBias = tz.tz_dsttime ? - tz.tz_minuteswest : 0;
+#	else
+		lBias = tm->tm_gmtoff;
+#	endif
+	if(lBias < 0)
+	{
+		t->OffsetMode = '-';
+		lBias *= -1;
+	}
+	else
+		t->OffsetMode = '+';
+	t->OffsetHour = lBias / 3600;
+	t->OffsetMinute = lBias % 3600;
+}
+
+
+
+
+/*******************************************************************
+ * BEGIN CODE-LIBLOGGING                                           *
+ *******************************************************************
+ * Code in this section is borrowed from liblogging. This is an
+ * interim solution. Once liblogging is fully integrated, this is
+ * to be removed (see http://www.monitorware.com/liblogging for
+ * more details. 2004-11-16 rgerhards
+ *
+ * Please note that the orginal liblogging code is modified so that
+ * it fits into the context of the current version of syslogd.c.
+ *
+ * DO NOT PUT ANY OTHER CODE IN THIS BEGIN ... END BLOCK!!!!
+ */
+
+/**
+ * Parse a 32 bit integer number from a string.
+ *
+ * \param ppsz Pointer to the Pointer to the string being parsed. It
+ *             must be positioned at the first digit. Will be updated 
+ *             so that on return it points to the first character AFTER
+ *             the integer parsed.
+ * \retval The number parsed.
+ */
+
+static int srSLMGParseInt32(char** ppsz)
+{
+	int i;
+
+	i = 0;
+	while(isdigit((int) **ppsz))
+	{
+		i = i * 10 + **ppsz - '0';
+		++(*ppsz);
+	}
+
+	return i;
+}
+
+
+/**
+ * Parse a TIMESTAMP-3339.
+ * updates the parse pointer position.
+ */
+static int
+ParseTIMESTAMP3339(struct syslogTime *pTime, char** ppszTS)
+{
+	char *pszTS = *ppszTS;
+
+	assert(pTime != NULL);
+	assert(ppszTS != NULL);
+	assert(pszTS != NULL);
+
+	pTime->year = srSLMGParseInt32(&pszTS);
+
+	/* We take the liberty to accept slightly malformed timestamps e.g. in 
+	 * the format of 2003-9-1T1:0:0. This doesn't hurt on receiving. Of course,
+	 * with the current state of affairs, we would never run into this code
+	 * here because at postion 11, there is no "T" in such cases ;)
+	 */
+	if(*pszTS++ != '-')
+		return FALSE;
+	pTime->month = srSLMGParseInt32(&pszTS);
+	if(pTime->month < 1 || pTime->month > 12)
+		return FALSE;
+
+	if(*pszTS++ != '-')
+		return FALSE;
+	pTime->day = srSLMGParseInt32(&pszTS);
+	if(pTime->day < 1 || pTime->day > 31)
+		return FALSE;
+
+	if(*pszTS++ != 'T')
+		return FALSE;
+
+	pTime->hour = srSLMGParseInt32(&pszTS);
+	if(pTime->hour < 0 || pTime->hour > 23)
+		return FALSE;
+
+	if(*pszTS++ != ':')
+		return FALSE;
+	pTime->minute = srSLMGParseInt32(&pszTS);
+	if(pTime->minute < 0 || pTime->minute > 59)
+		return FALSE;
+
+	if(*pszTS++ != ':')
+		return FALSE;
+	pTime->second = srSLMGParseInt32(&pszTS);
+	if(pTime->second < 0 || pTime->second > 60)
+		return FALSE;
+
+	/* Now let's see if we have secfrac */
+	if(*pszTS == '.')
+	{
+		char *pszStart = ++pszTS;
+		pTime->secfrac = srSLMGParseInt32(&pszTS);
+		pTime->secfracPrecision = (int) (pszTS - pszStart);
+	}
+	else
+	{
+		pTime->secfracPrecision = 0;
+		pTime->secfrac = 0;
+	}
+
+	/* check the timezone */
+	if(*pszTS == 'Z')
+	{
+		pszTS++; /* eat Z */
+		pTime->OffsetMode = 'Z';
+		pTime->OffsetHour = 0;
+		pTime->OffsetMinute = 0;
+	}
+	else if((*pszTS == '+') || (*pszTS == '-'))
+	{
+		pTime->OffsetMode = *pszTS;
+		pszTS++;
+
+		pTime->OffsetHour = srSLMGParseInt32(&pszTS);
+		if(pTime->OffsetHour < 0 || pTime->OffsetHour > 23)
+			return FALSE;
+
+		if(*pszTS++ != ':')
+			return FALSE;
+		pTime->OffsetMinute = srSLMGParseInt32(&pszTS);
+		if(pTime->OffsetMinute < 0 || pTime->OffsetMinute > 59)
+			return FALSE;
+	}
+	else
+		/* there MUST be TZ information */
+		return FALSE;
+
+	/* OK, we actually have a 3339 timestamp, so let's indicated this */
+	if(*pszTS == ' ')
+		++pszTS;
+	else
+		return FALSE;
+
+	/* update parse pointer */
+	*ppszTS = pszTS;
+
+	return TRUE;
+}
+
+
+/**
+ * Parse a TIMESTAMP-3164.
+ * Returns TRUE on parse OK, FALSE on parse error.
+ */
+static int
+ParseTIMESTAMP3164(struct syslogTime *pTime, char* pszTS)
+{
+	assert(pTime != NULL);
+	assert(pszTS != NULL);
+
+	getCurrTime(pTime);	/* obtain the current year and UTC offsets! */
+
+	/* If we look at the month (Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec),
+	 * we may see the following character sequences occur:
+	 *
+	 * J(an/u(n/l)), Feb, Ma(r/y), A(pr/ug), Sep, Oct, Nov, Dec
+	 *
+	 * We will use this for parsing, as it probably is the
+	 * fastest way to parse it.
+	 *
+	 * 2005-07-18, well sometimes it pays to be a bit more verbose, even in C...
+	 * Fixed a bug that lead to invalid detection of the data. The issue was that
+	 * we had an if(++pszTS == 'x') inside of some of the consturcts below. However,
+	 * there were also some elseifs (doing the same ++), which than obviously did not
+	 * check the orginal character but the next one. Now removed the ++ and put it
+	 * into the statements below. Was a really nasty bug... I didn't detect it before
+	 * june, when it first manifested. This also lead to invalid parsing of the rest
+	 * of the message, as the time stamp was not detected to be correct. - rgerhards
+	 */
+	switch(*pszTS++)
+	{
+	case 'J':
+		if(*pszTS == 'a') {
+			++pszTS;
+			if(*pszTS == 'n') {
+				++pszTS;
+				pTime->month = 1;
+			} else
+				return FALSE;
+		} else if(*pszTS == 'u') {
+			++pszTS;
+			if(*pszTS == 'n') {
+				++pszTS;
+				pTime->month = 6;
+			} else if(*pszTS == 'l') {
+				++pszTS;
+				pTime->month = 7;
+			} else
+				return FALSE;
+		} else
+			return FALSE;
+		break;
+	case 'F':
+		if(*pszTS == 'e') {
+			++pszTS;
+			if(*pszTS == 'b') {
+				++pszTS;
+				pTime->month = 2;
+			} else
+				return FALSE;
+		} else
+			return FALSE;
+		break;
+	case 'M':
+		if(*pszTS == 'a') {
+			++pszTS;
+			if(*pszTS == 'r') {
+				++pszTS;
+				pTime->month = 3;
+			} else if(*pszTS == 'y') {
+				++pszTS;
+				pTime->month = 5;
+			} else
+				return FALSE;
+		} else
+			return FALSE;
+		break;
+	case 'A':
+		if(*pszTS == 'p') {
+			++pszTS;
+			if(*pszTS == 'r') {
+				++pszTS;
+				pTime->month = 4;
+			} else
+				return FALSE;
+		} else if(*pszTS == 'u') {
+			++pszTS;
+			if(*pszTS == 'g') {
+				++pszTS;
+				pTime->month = 8;
+			} else
+				return FALSE;
+		} else
+			return FALSE;
+		break;
+	case 'S':
+		if(*pszTS == 'e') {
+			++pszTS;
+			if(*pszTS == 'p') {
+				++pszTS;
+				pTime->month = 9;
+			} else
+				return FALSE;
+		} else
+			return FALSE;
+		break;
+	case 'O':
+		if(*pszTS == 'c') {
+			++pszTS;
+			if(*pszTS == 't') {
+				++pszTS;
+				pTime->month = 10;
+			} else
+				return FALSE;
+		} else
+			return FALSE;
+		break;
+	case 'N':
+		if(*pszTS == 'o') {
+			++pszTS;
+			if(*pszTS == 'v') {
+				++pszTS;
+				pTime->month = 11;
+			} else
+				return FALSE;
+		} else
+			return FALSE;
+		break;
+	case 'D':
+		if(*pszTS == 'e') {
+			++pszTS;
+			if(*pszTS == 'c') {
+				++pszTS;
+				pTime->month = 12;
+			} else
+				return FALSE;
+		} else
+			return FALSE;
+		break;
+	default:
+		return FALSE;
+	}
+
+	/* done month */
+
+	if(*pszTS++ != ' ')
+		return FALSE;
+
+	/* we accept a slightly malformed timestamp when receiving. This is
+	 * we accept one-digit days
+	 */
+	if(*pszTS == ' ')
+		++pszTS;
+
+	pTime->day = srSLMGParseInt32(&pszTS);
+	if(pTime->day < 1 || pTime->day > 31)
+		return FALSE;
+
+	if(*pszTS++ != ' ')
+		return FALSE;
+	pTime->hour = srSLMGParseInt32(&pszTS);
+	if(pTime->hour < 0 || pTime->hour > 23)
+		return FALSE;
+
+	if(*pszTS++ != ':')
+		return FALSE;
+	pTime->minute = srSLMGParseInt32(&pszTS);
+	if(pTime->minute < 0 || pTime->minute > 59)
+		return FALSE;
+
+	if(*pszTS++ != ':')
+		return FALSE;
+	pTime->second = srSLMGParseInt32(&pszTS);
+	if(pTime->second < 0 || pTime->second > 60)
+		return FALSE;
+	if(*pszTS++ != ':')
+
+	/* OK, we actually have a 3164 timestamp, so let's indicate this
+	 * and fill the rest of the properties. */
+	pTime->timeType = 1;
+ 	pTime->secfracPrecision = 0;
+	pTime->secfrac = 0;
+	return TRUE;
+}
+
+/*******************************************************************
+ * END CODE-LIBLOGGING                                             *
+ *******************************************************************/
+
+/**
+ * Format a syslogTimestamp into format required by MySQL.
+ * We are using the 14 digits format. For example 20041111122600 
+ * is interpreted as '2004-11-11 12:26:00'. 
+ * The caller must provide the timestamp as well as a character
+ * buffer that will receive the resulting string. The function
+ * returns the size of the timestamp written in bytes (without
+ * the string terminator). If 0 is returend, an error occured.
+ */
+int formatTimestampToMySQL(struct syslogTime *ts, char* pDst, size_t iLenDst)
+{
+	/* currently we do not consider localtime/utc. This may later be
+	 * added. If so, I recommend using a property replacer option
+	 * and/or a global configuration option. However, we should wait
+	 * on user requests for this feature before doing anything.
+	 * rgerhards, 2007-06-26
+	 */
+	assert(ts != NULL);
+	assert(pDst != NULL);
+
+	if (iLenDst < 15) /* we need at least 14 bytes
+			     14 digits for timestamp + '\n' */
+		return(0); 
+
+	return(snprintf(pDst, iLenDst, "%4.4d%2.2d%2.2d%2.2d%2.2d%2.2d", 
+		ts->year, ts->month, ts->day, ts->hour, ts->minute, ts->second));
+
+}
+
+int formatTimestampToPgSQL(struct syslogTime *ts, char *pDst, size_t iLenDst)
+{
+       /* see note in formatTimestampToMySQL, applies here as well */
+       assert(ts != NULL);
+       assert(pDst != NULL);
+
+       if (iLenDst < 21) /* we need 20 bytes + '\n' */
+               return(0);
+
+       return(snprintf(pDst, iLenDst, "%4.4d-%2.2d-%2.2d %2.2d:%2.2d:%2.2d",
+                               ts->year, ts->month, ts->day, ts->hour, ts->minute, ts->second));
+}
+
+/**
+ * Format a syslogTimestamp to a RFC3339 timestamp string (as
+ * specified in syslog-protocol).
+ * The caller must provide the timestamp as well as a character
+ * buffer that will receive the resulting string. The function
+ * returns the size of the timestamp written in bytes (without
+ * the string terminator). If 0 is returend, an error occured.
+ */
+int formatTimestamp3339(struct syslogTime *ts, char* pBuf, size_t iLenBuf)
+{
+	int iRet;
+	char szTZ[7]; /* buffer for TZ information */
+
+	assert(ts != NULL);
+	assert(pBuf != NULL);
+	
+	if(iLenBuf < 20)
+		return(0); /* we NEED at least 20 bytes */
+
+	/* do TZ information first, this is easier to take care of "Z" zone in rfc3339 */
+	if(ts->OffsetMode == 'Z') {
+		szTZ[0] = 'Z';
+		szTZ[1] = '\0';
+	} else {
+		snprintf(szTZ, sizeof(szTZ) / sizeof(char), "%c%2.2d:%2.2d",
+			ts->OffsetMode, ts->OffsetHour, ts->OffsetMinute);
+	}
+
+	if(ts->secfracPrecision > 0)
+	{	/* we now need to include fractional seconds. While doing so, we must look at
+		 * the precision specified. For example, if we have millisec precision (3 digits), a
+		 * secFrac value of 12 is not equivalent to ".12" but ".012". Obviously, this
+		 * is a huge difference ;). To avoid this, we first create a format string with
+		 * the specific precision and *then* use that format string to do the actual
+		 * formating (mmmmhhh... kind of self-modifying code... ;)).
+		 */
+		char szFmtStr[64];
+		/* be careful: there is ONE actual %d in the format string below ;) */
+		snprintf(szFmtStr, sizeof(szFmtStr),
+		         "%%04d-%%02d-%%02dT%%02d:%%02d:%%02d.%%0%dd%%s",
+			ts->secfracPrecision);
+		iRet = snprintf(pBuf, iLenBuf, szFmtStr, ts->year, ts->month, ts->day,
+			        ts->hour, ts->minute, ts->second, ts->secfrac, szTZ);
+	}
+	else
+		iRet = snprintf(pBuf, iLenBuf,
+		 		"%4.4d-%2.2d-%2.2dT%2.2d:%2.2d:%2.2d%s",
+				ts->year, ts->month, ts->day,
+			        ts->hour, ts->minute, ts->second, szTZ);
+	return(iRet);
+}
+
+/**
+ * Format a syslogTimestamp to a RFC3164 timestamp sring.
+ * The caller must provide the timestamp as well as a character
+ * buffer that will receive the resulting string. The function
+ * returns the size of the timestamp written in bytes (without
+ * the string termnator). If 0 is returend, an error occured.
+ */
+int formatTimestamp3164(struct syslogTime *ts, char* pBuf, size_t iLenBuf)
+{
+	static char* monthNames[13] = {"ERR", "Jan", "Feb", "Mar",
+	                               "Apr", "May", "Jun", "Jul",
+				       "Aug", "Sep", "Oct", "Nov", "Dec"};
+	assert(ts != NULL);
+	assert(pBuf != NULL);
+	
+	if(iLenBuf < 16)
+		return(0); /* we NEED 16 bytes */
+	return(snprintf(pBuf, iLenBuf, "%s %2d %2.2d:%2.2d:%2.2d",
+		monthNames[ts->month], ts->day, ts->hour,
+		ts->minute, ts->second
+		));
+}
+
+/**
+ * Format a syslogTimestamp to a text format.
+ * The caller must provide the timestamp as well as a character
+ * buffer that will receive the resulting string. The function
+ * returns the size of the timestamp written in bytes (without
+ * the string termnator). If 0 is returend, an error occured.
+ */
+#if 0 /* This method is currently not called, be we like to preserve it */
+static int formatTimestamp(struct syslogTime *ts, char* pBuf, size_t iLenBuf)
+{
+	assert(ts != NULL);
+	assert(pBuf != NULL);
+	
+	if(ts->timeType == 1) {
+		return(formatTimestamp3164(ts, pBuf, iLenBuf));
+	}
+
+	if(ts->timeType == 2) {
+		return(formatTimestamp3339(ts, pBuf, iLenBuf));
+	}
+
+	return(0);
+}
+#endif
+/* queryInterface function
+ * rgerhards, 2008-03-05
+ */
+BEGINobjQueryInterface(datetime)
+CODESTARTobjQueryInterface(datetime)
+	if(pIf->ifVersion != datetimeCURR_IF_VERSION) { /* check for current version, increment on each change */
+		ABORT_FINALIZE(RS_RET_INTERFACE_NOT_SUPPORTED);
+	}
+
+	/* ok, we have the right interface, so let's fill it
+	 * Please note that we may also do some backwards-compatibility
+	 * work here (if we can support an older interface version - that,
+	 * of course, also affects the "if" above).
+	 */
+	pIf->getCurrTime = getCurrTime;
+	pIf->ParseTIMESTAMP3339 = ParseTIMESTAMP3339;
+	pIf->ParseTIMESTAMP3164 = ParseTIMESTAMP3164;
+	pIf->formatTimestampToMySQL = formatTimestampToMySQL;
+	pIf->formatTimestampToPgSQL = formatTimestampToPgSQL;
+	pIf->formatTimestamp3339 = formatTimestamp3339;
+	pIf->formatTimestamp3164 = formatTimestamp3164;
+finalize_it:
+ENDobjQueryInterface(datetime)
+
+
+/* Initialize the datetime class. Must be called as the very first method
+ * before anything else is called inside this class.
+ * rgerhards, 2008-02-19
+ */
+BEGINAbstractObjClassInit(datetime, 1, OBJ_IS_CORE_MODULE) /* class, version */
+	/* request objects we use */
+	CHKiRet(objUse(errmsg, CORE_COMPONENT));
+
+ENDObjClassInit(datetime)
+
+/* vi:set ai:
+ */