Index: src/test/java/org/apache/pdfbox/util/TestDateUtil.java =================================================================== --- src/test/java/org/apache/pdfbox/util/TestDateUtil.java (revision 1566603) +++ src/test/java/org/apache/pdfbox/util/TestDateUtil.java (working copy) @@ -65,35 +65,30 @@ * * @throws Exception when there is an exception */ - public void testExtract() throws Exception + public void testToCalendar() throws Exception { - TimeZone timezone = TimeZone.getDefault(); - TimeZone.setDefault(TimeZone.getTimeZone("UTC")); - try - { - assertCalendarEquals( new GregorianCalendar( 2005, 4, 12 ), - DateConverter.toCalendar( "D:05/12/2005" ) ); - assertCalendarEquals( new GregorianCalendar( 2005, 4,12,15,57,16 ), - DateConverter.toCalendar( "5/12/2005 15:57:16" ) ); - } - catch (IOException ex) - { - ex.printStackTrace(); - } - finally - { - TimeZone.setDefault(timezone); - } - // check that new toCalendar gives NullPointer for a null arg - try - { - DateConverter.toCalendar(null, null); - assertNotNull(null); // failed to have expected exception - } - catch (NullPointerException ex) - { - // expected outcome - } + TimeZone zuluTime = TimeZone.getTimeZone("UTC"); + + GregorianCalendar date = new GregorianCalendar( 2005, 4, 12 ); + DateConverter.adjustTimeZoneNicely(date, zuluTime); + assertCalendarEquals(date, + DateConverter.toCalendar( "D:05/12/2005" ) ); + + date = new GregorianCalendar( 2005, 4,12,15,57,16 ); + DateConverter.adjustTimeZoneNicely(date, zuluTime); + assertCalendarEquals( date, + DateConverter.toCalendar( "5/12/2005 15:57:16" ) ); + + date = new GregorianCalendar(1902, 4, 7,0,0,0); + DateConverter.adjustTimeZoneNicely(date, zuluTime); + assertCalendarEquals(date, DateConverter.toCalendar("7 of May in 1902", + new String[] {"d 'of' MMM 'in' yyyy"})); + + assertNull(DateConverter.toCalendar((String)null)); + assertNull(DateConverter.toCalendar((COSString)null)); + assertNull(DateConverter.toCalendar("")); + assertNull(DateConverter.toCalendar(new COSString(""))); + assertNull(DateConverter.toCalendar(" ")); } /** @@ -213,6 +208,11 @@ checkParse(2020, 5,26,11,25,10, 0, "26 May 2020 11:25:10"); checkParse(2021, 5,26,11,23, 0, 0, "26 May 2021 11:23"); + // try null or empty args + checkParse(BAD, 0, 0, 0, 0, 0, 0, null); + checkParse(BAD, 0, 0, 0, 0, 0, 0, ""); + checkParse(BAD, 0, 0, 0, 0, 0, 0, " "); + // try dates invalid due to out of limit values checkParse(BAD, 0, 0, 0, 0, 0, 0, "Tuesday, May 32 2000 11:27 UCT"); checkParse(BAD, 0, 0, 0, 0, 0, 0, "32 May 2000 11:25"); @@ -410,11 +410,11 @@ } /** - * Timezone offset testcase. + * Timezone offset test case. * * @throws Exception */ - public void testFormatTZoffset() + public void testFormatTZoffset() throws Exception { checkFormatOffset(-12.1, "+11:54"); checkFormatOffset(12.1, "-11:54"); Index: src/main/java/org/apache/pdfbox/util/DateConverter.java =================================================================== --- src/main/java/org/apache/pdfbox/util/DateConverter.java (revision 1566603) +++ src/main/java/org/apache/pdfbox/util/DateConverter.java (working copy) @@ -77,7 +77,7 @@ * must be within bounds. If an attempt is made to parse an invalid date * field, toCalendar(String, String[]) returns Jan 1 in year INVALID_YEAR. */ - public static final int INVALID_YEAR = 999; + public static final int INVALID_YEAR = 999; // ("nein, nein, nein") /** @@ -628,7 +628,7 @@ * - The formats tried are alphaStartFormats or digitStartFormat and * any listed in the value of moreFmts. * - * @param text The String that may begin with a date. Must not be null. + * @param text The String that may begin with a date. * Initial spaces and "D:" are skipped over. * @param moreFmts Additional formats to be tried after trying the * built-in formats. @@ -645,7 +645,10 @@ public static Calendar parseDate(String text, String[] moreFmts, ParsePosition initialWhere) { - // place to remember longestr date string + if (text == null || text.equals("")) + return null; + + // remember longest date string int longestLen = -999999; // theorem: this value will never be used // proof: longestLen is only used if longestDate is not null GregorianCalendar longestDate = null; // null says no date found yet @@ -740,7 +743,9 @@ * @return The Calendar that the text string represents. * Or null if text was null. * @throws IOException If the date string is not in the correct format. - * @deprecated This method throws an IOException for failure. Replace + * @deprecated This method creates an unnecessary dependency + * of DateConverter on pdfbox. The method should be removed as part of + * moving DateConverter to a support package. Replace * calls to it with {@link #toCalendar(String, String[])} * and test for failure with * (value == null || value.get(Calendar.YEAR) == INVALID_YEAR) @@ -763,7 +768,8 @@ * * @param text The string representation of the calendar. * @return The Calendar that this string represents - * or null if the incoming text is null. + * or null if the incoming text is null, the empty string or + * a string with only spaces. * @throws IOException If the date string is non-null * and not a parseable date. * @deprecated This method throws an IOException for failure. Replace @@ -774,12 +780,12 @@ */ public static Calendar toCalendar(String text) throws IOException { - if (text == null) + if (text == null || text.equals("") || text.trim().equals("")) { return null; } Calendar val = toCalendar(text, null); - if (val != null && val.get(Calendar.YEAR) == INVALID_YEAR) + if (val != null && isBad(val)) { throw new IOException("Error converting date: " + text); } @@ -795,18 +801,31 @@ * * @param text The text to parse. Initial spaces and "D:" are skipped over. * @param moreFmts An Array of formats (as Strings) to try - * in addition to the standard list. - * @return the Calendar value corresponding to the date text. - * If text does not represent a valid date, - * the value is January 1 on year INVALID_YEAR at 0:0:0 GMT. + * in addition to the standard list. Example: + * Calendar date = toCalendar("7 of May in 1902", + * new String[] {"d of MMM in yyyy"}); + * @return the Calendar value corresponding to the date text. + * The design is such that DateConverter ALWAYS returns a Calendar, + * even for a null or empty argument. Client code need not test + * for validity if it does not care what the date is. + * If 'text' does not represent a valid date, + * the value returned is January 1 of year INVALID_YEAR at 0:0:0 GMT. * */ public static Calendar toCalendar(String text, String[] moreFmts) { + Calendar retCal; ParsePosition where = new ParsePosition(0); - skipOptionals(text, where, " "); - skipString(text, "D:", where); - Calendar retCal = parseDate(text, moreFmts, where); // PARSE THE TEXT + if (text == null || "".equals(text) || "".equals(text.trim())) + { + retCal = null; + } + else + { + skipOptionals(text, where, " "); + skipString(text, "D:", where); + retCal = parseDate(text, moreFmts, where); // PARSE THE TEXT + } if (retCal == null || where.getIndex() != text.length()) { // the date string is invalid for all formats we tried, @@ -815,4 +834,38 @@ } return retCal; } + + /** + * Returns whether or not the Calendar created by parse + * is a bad date. This is syntactic sugar for the transition + * from returning null or a sentinel Calendar value to + * returning only a sentinel Calendar value + * + * @param cal Calendar to be checked + * @return was this a null or the sentinel bad calendar + */ + public static boolean isBad(Calendar cal) + { + if (cal == null) + { + return true; + } + + //Sentinel bad is Gregorian + if (! (cal instanceof GregorianCalendar)) + { + return false; + } + if (cal.get(Calendar.YEAR) == INVALID_YEAR && + cal.get(Calendar.MONTH) == 0 && + cal.get(Calendar.DAY_OF_WEEK_IN_MONTH) == 1 && + cal.get(Calendar.HOUR_OF_DAY) == 0 && + cal.get(Calendar.MINUTE) == 0 && + cal.get(Calendar.SECOND) == 0 && + cal.get(Calendar.MILLISECOND) == 0) + { + return true; + } + return false; + } }