xmltv-commit — XMLTV commit messages

[Xmltv-commit] CVS: xmltv/todo TASKS.epa,NONE,1.1

[Jerry V. <jve...@us...>]

Update of /cvsroot/xmltv/xmltv/todo
In directory usw-pr-cvs1:/tmp/cvs-serv16014/todo

Added Files:
	TASKS.epa 
Log Message:
xmltv-0.2.0

--- NEW FILE: TASKS.epa ---
This is my todo list, it relates mostly to UK listings-grabbing and
the other stuff I wrote.  -- Ed Avis, ep...@do...

4Learning - done
morning on 3 - done
(lw) (fm) - done
FILM 1978 DESC - done
trailing semicolon in desc - done
store yes/no, not maybe - done
richard e. grant stars in - done
news;weather (check mail) - done
As World Service
news; radio 3 lunchtime concert - contained in source HTML
pick.cgi actually output XML, or something
Pink Panther film vs cartoon
todos in DTD
Gareth Hale guests
blank lines in sort_listings output
FIXMEs
use the TVChannels.pm module

[Xmltv-commit] CVS: xmltv/todo/TVChannels_module - New directory

[Jerry V. <jve...@us...>]

Update of /cvsroot/xmltv/xmltv/todo/TVChannels_module
In directory usw-pr-cvs1:/tmp/cvs-serv15790/TVChannels_module

Log Message:
Directory /cvsroot/xmltv/xmltv/todo/TVChannels_module added to the repository

[Xmltv-commit] CVS: xmltv/todo - New directory

[Jerry V. <jve...@us...>]

Update of /cvsroot/xmltv/xmltv/todo
In directory usw-pr-cvs1:/tmp/cvs-serv15670/todo

Log Message:
Directory /cvsroot/xmltv/xmltv/todo added to the repository

[Xmltv-commit] CVS: xmltv tz.pl,NONE,1.1 ChangeLog,1.1,1.2 README,1.1,1.2 filter_shown,1.1,1.2 getlistings_pa,1.1,1.2 listings_to_latex,1.1,1.2 pick_cgi,1.1,1.2 xmltv.dtd,1.1,1.2

[Jerry V. <jve...@us...>]

Update of /cvsroot/xmltv/xmltv
In directory usw-pr-cvs1:/tmp/cvs-serv14737

Modified Files:
	ChangeLog README filter_shown getlistings_pa listings_to_latex 
	pick_cgi xmltv.dtd 
Added Files:
	tz.pl 
Log Message:
xmltv-0.1.1

--- NEW FILE: tz.pl ---
# tz.pl
# 
# Timezone stuff.  Include this _after_ importing Date::Manip.  Should
# really put this in a proper module.
# 
# -- Ed Avis, ep...@do..., 2000-07-28
# 


# parse_uk_date()
# 
# Wrapper for ParseDate() that tries to guess what timezone a date is
# in (UT or BST).  If the date already has either of these it is left
# alone.
# 
# This will probably fail horribly if you use it for dates which
# aren't in either of these timezones.  But since summer time starts
# and ends at the same time throughout the EU it wouldn't be hard to
# make it work for some other countries.
# 
# Parameters: unparsed date from the UK (or other places using UT/BST)
# Returns:    parsed date
# 
# There's a one hour window where dates are ambigous; we assume UT
# for these and print a warning.  Similarly there's a one hour window
# where dates without a timezone are impossible; we die on those.
# 
sub parse_uk_date($) {
    die 'usage: parse_uk_date(unparsed date)' if @_ != 1;
    my $date = shift;

    if (defined gettz($date)) {
	# An explicit timezone, no need for any funny business
	return ParseDate($date);
    }
    # FIXME: what if it has a timezone and it's not UT or UT+1?

    my $dp = ParseDate($date);
    die "bad date $date" if not defined $dp or $dp eq '';

    # Start and end of summer time in that year
    my ($start_bst, $end_bst) = @{bst_dates(UnixDate($dp, '%Y'))};

    # The clocks shift backwards and forwards by one hour.
    my $clock_shift = "1 hour";

    # The times that the clocks go forward to / back to.
    my $start_bst_skipto = DateCalc($start_bst, "+ $clock_shift");
    my $end_bst_backto = DateCalc($end_bst, "- $clock_shift");

    if (Date_Cmp($dp, $start_bst) < 0) {
	# Before the start of summer time.
	return $dp;
    }
    elsif (Date_Cmp($dp, $start_bst) == 0) {
	# Exactly _at_ the start of summer time.  Really such a date
	# should not exist since the clocks skip forward an hour at
	# that point.
	# 
	return $dp;
    }
    elsif (Date_Cmp($dp, $start_bst_skipto) < 0) {
	die("$date is impossible "
	    . "(summer time skips from $start_bst to $start_bst_skipto)" );
    }
    elsif (Date_Cmp($dp, $end_bst_backto) < 0) {
	# During summer time.  FIXME: store timezone names as
	# constants.
	# 
	return Date_ConvTZ($dp, 'BST', 'UT');
    }
    elsif (Date_Cmp($dp, $end_bst) <= 0) {
	warn("$date is ambiguous "
	     . "(clocks go back from $end_bst to $end_bst_backto), "
	     . "assuming not summer time" );

	return $dp;
    }
    else {
	# Definitely after the end of summer time.
	return $dp;
    }
}


# date_to_uk()
# 
# Take a date in UT and convert it to a BST date if needed.
# 
# Parameters: date in UT (from ParseDate())
# 
# Returns ref to list of
#   new date (maybe shifted by one hour),
#   timezone of new date ('UT' or 'BST')
# 
# For example, date_to_uk of 13:00 on June 10th 2000 would be 14:00
# BST on the same day.  The input and output date are both in
# Date::Manip internal format.
# 
sub date_to_uk($) {
    die 'usage: date_to_uk(date in Date::Manip format)' if @_ != 1;
    my $d = shift;

    my $year = UnixDate($d, '%Y');
    if ((not defined $year) or ($year !~ tr/0-9//)) {
	die "cannot get year from '$d'";
    }

    # Find the start and end dates in March and October
    my ($start_bst, $end_bst) = @{bst_dates($year)};

    # The clocks shift backwards and forwards by one hour.
    my $clock_shift = "1 hour";

    if (Date_Cmp($d, $start_bst) < 0) {
	# Before the start of summer time.
	return [ $d, 'UT' ];
    }
    elsif (Date_Cmp($d, $end_bst) < 0) {
	# During summer time.
	return [ DateCalc($d, "+ $clock_shift"), 'BST' ];
    }
    else {
	# After summer time.
	return [ $d, 'UT' ];
    }
}


# bst_dates()
# 
# Return the dates (in UT) when British Summer Time starts and ends in
# a given year.
# 
# According to
# <http://www.rog.nmm.ac.uk/leaflets/summer/summer.html>, summer
# time starts at 01:00 on the last Sunday in March, and ends at
# 01:00 on the last Sunday in October.  This has been the case
# since 1998 - earlier dates are not handled.
# 
# Parameters: year (only 1998 or later works)
# 
# Returns: ref to list of
#   start time and date of summer time (in UT)
#   end time and date of summer time (in UT)
# 
sub bst_dates($) {
    die 'usage: bst_dates(year)' if @_ != 1;
    my $year = shift;
    die "don't know about BST before 1998" if $year < 1998;

    my ($start_bst, $end_bst);
    foreach (1 .. 31) {
	my $mar = "$year-03-$_" . ' 01:00';
	my $mar_d = ParseDate($mar) or die "cannot parse $mar";
	$start_bst = $mar_d if UnixDate($mar_d, "%A") =~ /Sunday/;

	# N.B. 01:00 UT == 00:00 BST.  A time between '23:00' and
	# '00:00' just before the last Sunday in October is ambiguous.
	# 
	my $oct = "$year-10-$_" . ' 00:00';
	my $oct_d = ParseDate($oct) or die "cannot parse $oct";
	$end_bst = $oct_d if UnixDate($oct_d, "%A") =~ /Sunday/;
    }
    die if not defined $start_bst or not defined $end_bst;

    return [ $start_bst, $end_bst ];
}


# gettz()
# 
# Parameters: unparsed date string
# Returns: timezone (a substring), or undef
# 
# The only timezones supported are UT (aka GMT) and BST.
# 
sub gettz($) {
    die 'usage: gettz(unparsed date string)' if @_ != 1;
    local $_ = shift;

    /(UT|GMT|BST)/    && return $1;
    /(\+0000|\+0100)/ && return $1;
    return undef;
}

1;

Index: ChangeLog
===================================================================
RCS file: /cvsroot/xmltv/xmltv/ChangeLog,v
retrieving revision 1.1
retrieving revision 1.2
diff -C2 -d -r1.1 -r1.2
*** ChangeLog	2001/11/02 17:58:43	1.1
--- ChangeLog	2001/11/02 18:07:18	1.2
***************
*** 1,2 ****
--- 1,31 ----
+ 2000-07-28  Edward Avis  <ep...@do...>
+ 
+ 	* xmltv.dtd: specify date format for all dates
+ 	
+ 	* getlistings_pa: print timezones when they change
+ 
+ 	* tz.pl: new file for British timezone routines
+ 
+ 	* listings_to_latex: persuaded XMLin() to read ordinary filenames
+ 
+ 	* getlistings_pa: use the Memoize module to speed things up
+ 
+ 	* getlistings_pa: dates converted to UT when reading, and back again
+ 			  when writing
+ 
+ 2000-07-24  Edward Avis  <ep...@do...>
+ 
+ 	* getlistings_pa: tell HTML::Parser not to break chunks of text
+ 
+ 2000-07-21  Edward Avis  <ep...@do...>
+ 
+ 	* filter_shown: add -d flag to specify date instead of 'now'
+ 
+ 2000-07-11  Edward Avis  <ep...@do...>
+ 
+ 	* getlistings_pa: changed a couple of dies to warnings
+ 	(extract_prog($)): remove '&' characters from description
+ 	(extract_prog($)): clean up all text, not just descriptions
+ 
  2000-07-03  Edward Avis  <ep...@do...>
  

Index: README
===================================================================
RCS file: /cvsroot/xmltv/xmltv/README,v
retrieving revision 1.1
retrieving revision 1.2
diff -C2 -d -r1.1 -r1.2
*** README	2001/11/02 17:58:43	1.1
--- README	2001/11/02 18:07:18	1.2
***************
*** 1,3 ****
! XMLTV 0.1
  
  Suite of programs for processing TV listings and organizing your
--- 1,3 ----
! XMLTV 0.1.1
  
  Suite of programs for processing TV listings and organizing your
***************
*** 9,13 ****
  pick_cgi as a CGI script.
  
! This release is not ready for end users.
  
  This is free software distributed under the GPL, see COPYING.
--- 9,15 ----
  pick_cgi as a CGI script.
  
! This release is not ready for end users.  There have been lots of
! changes since the last release, notably the introduction of timezones,
! and the fact that it might actually work.
  
  This is free software distributed under the GPL, see COPYING.
***************
*** 16,18 ****
  <http://www.doc.ic.ac.uk/~epa98/work/apps/xmltv/>.
  
! -- Ed Avis, ep...@do..., 2000-07-03
--- 18,20 ----
  <http://www.doc.ic.ac.uk/~epa98/work/apps/xmltv/>.
  
! -- Ed Avis, ep...@do..., 2000-07-28

Index: filter_shown
===================================================================
RCS file: /cvsroot/xmltv/xmltv/filter_shown,v
retrieving revision 1.1
retrieving revision 1.2
diff -C2 -d -r1.1 -r1.2
*** filter_shown	2001/11/02 17:58:43	1.1
--- filter_shown	2001/11/02 18:07:18	1.2
***************
*** 3,9 ****
  # filter_shown
  # 
! # Filter TV listings, removing programmes that have already finished.
  # 
! # Usage: filter_shown [filename]
  # 
  # Reads from filename, or stdin if no filename given, and writes
--- 3,10 ----
  # filter_shown
  # 
! # Filter TV listings, removing programmes that would have already
! # finished at a given date and time (defaults to now).
  # 
! # Usage: filter_shown [-d date] [filename]
  # 
  # Reads from filename, or stdin if no filename given, and writes
***************
*** 16,25 ****
  use XML::DOM;
  use Date::Manip;
  
! die "usage: $0 [filename]" if @ARGV >= 2;
! @ARGV = ('-') if not @ARGV;
  
! # TODO: Maybe a command line option to give the date?
! my $now = ParseDate('today');
  
  my $p = new XML::DOM::Parser;
--- 17,29 ----
  use XML::DOM;
  use Date::Manip;
+ use Getopt::Std;
  
! use vars '$opt_d';
! getopt('d');
! $opt_d = 'today' if not defined $opt_d;
! my $now = ParseDate($opt_d);
  
! die "usage: $0 [-d date] [filename]" if @ARGV >= 2;
! @ARGV = ('-') if not @ARGV;
  
  my $p = new XML::DOM::Parser;

Index: getlistings_pa
===================================================================
RCS file: /cvsroot/xmltv/xmltv/getlistings_pa,v
retrieving revision 1.1
retrieving revision 1.2
diff -C2 -d -r1.1 -r1.2
*** getlistings_pa	2001/11/02 17:58:43	1.1
--- getlistings_pa	2001/11/02 18:07:18	1.2
***************
*** 45,55 ****
  use strict;
  use HTML::TokeParser;
! use Date::Manip;
  use LWP::Simple;
- use Term::ReadKey;
  use XML::Writer;
- use Text::Wrap;
  
  
  ########
  # Configuration
--- 45,81 ----
  use strict;
  use HTML::TokeParser;
! 
! # Date::Manip is used for parsing dates; all dates are stored
! # internally in UT (aka GMT).
! # 
! use Date::Manip; Date_Init('TZ=UT');
! require 'tz.pl';
! 
  use LWP::Simple;
  use XML::Writer;
  
+ # I have a module I use for trace messages; you can get it at
+ # <http://www.doc.ic.ac.uk/~epa98/lib/perl5/site_perl/Dbg.pm>.  That's
+ # what all the t('hello') type statements are.  But you can probably
+ # just leave them commented out.
+ # 
+ #use Dbg; $Dbg::On = 0;
  
+ # Memoize some subroutines if possible (25x speedup)
+ eval { require Memoize };
+ if ($@) {
+     warn "could not load Memoize module, things will run slowly";
+ }
+ else {
+     foreach (qw/istime time12to24 parse_uk_date date_to_uk bst_dates
+ 	        gettz ParseDate UnixDate DateCalc Date_Cmp
+                 ParseDateDelta/)
+     {
+ 	Memoize::memoize($_) or die "cannot memoize $_: $!";
+     }
+ }
+ 
+ 
+ 
  ########
  # Configuration
***************
*** 62,66 ****
  
  # Quicker version for testing
! #my @CHANNELS = qw[bbc1];
  #my $NUMDAYS = 1;
  
--- 88,92 ----
  
  # Quicker version for testing
! #my @CHANNELS = qw[ch5];
  #my $NUMDAYS = 1;
  
***************
*** 235,243 ****
  # Given a list of programme details (from extract_prog()) and the date
  # of this listing, remove the silly 12 hour times and add proper time
! # and date for each programme.
  # 
  # Parameters:
  #   reference to list of programme details
! #   date (in Date::Manip format)
  # 
  # Modifies the list passed in, so that each programme no longer has
--- 261,269 ----
  # Given a list of programme details (from extract_prog()) and the date
  # of this listing, remove the silly 12 hour times and add proper time
! # and date for each programme.  In UT.
  # 
  # Parameters:
  #   reference to list of programme details
! #   date in format YYYYMMDD
  # 
  # Modifies the list passed in, so that each programme no longer has
***************
*** 247,265 ****
      die 'usage: date_programmes(ref to list of programmes, date)'
        if @_ != 2;
!     my ($progs, $date) = @_;
! 
      my $prev_date;
-     my $base_day = $date;
      my $day_later = ParseDateDelta('+1 day') or die;
    
      foreach (@$progs) {
! 	my $t = timecvt($_->{'time'});
! 	
! 	# Calculate the time and day of the programme
! 	my $time_delta = ParseDateDelta("$t:00") or die;
! 	my $err;
! 	my $prog_date = DateCalc($base_day, $time_delta, \$err);
! 	die "error from DateCalc(): $err" if defined $err;
  	
  	if (defined $prev_date
  	    and (Date_Cmp($prev_date, $prog_date) > 0))
--- 273,296 ----
      die 'usage: date_programmes(ref to list of programmes, date)'
        if @_ != 2;
!     my ($progs, $base_day) = @_;
!     die 'bad base day, format should be YYYYMMDD'
!       if $base_day !~ /^\d{8}$/;
!     
      my $prev_date;
      my $day_later = ParseDateDelta('+1 day') or die;
    
      foreach (@$progs) {
! 	my $t = time12to24($_->{'time'});
! #	t('converted 12 hour time ' . d($_->{'time'}) . ' to ' . d($t));
  	
+ 	# $t should be hours and minutes, add seconds
+ 	$t =~ s/\b(\d\d):(\d\d)\b/$1:$2:00/ or die "bad hh:mm $t";
+ #	t('added seconds, time now ' . d($t));
+ 
+ 	my $prog_date = parse_uk_date("$base_day $t");
+ 	die "bad base day + time '$t'"
+ 	  if not defined $prog_date or $prog_date eq '';
+ #	t('canonical form is ' . d($prog_date));
+ 
  	if (defined $prev_date
  	    and (Date_Cmp($prev_date, $prog_date) > 0))
***************
*** 269,276 ****
  	    # 
  	    my $err;
! 	    $base_day = DateCalc($base_day, $day_later, \$err);
! 	    die "error from DateCalc(): $err" if defined $err;
! 	    $prog_date = DateCalc($base_day, $time_delta, \$err);
  	    die "error from DateCalc(): $err" if defined $err;
  	}
      
--- 300,308 ----
  	    # 
  	    my $err;
! 	    $base_day = UnixDate DateCalc($base_day, $day_later, \$err), '%Q';
  	    die "error from DateCalc(): $err" if defined $err;
+ 	    $prog_date = parse_uk_date("$base_day $t");
+ 	    die "bad base day + time '$t'"
+ 	      if not defined $prog_date or $prog_date eq '';
  	}
      
***************
*** 282,286 ****
  
      
! # extract_prog
  # 
  # Given a programme as a ref to a list of bits of text, each with
--- 314,318 ----
  
      
! # extract_prog()
  # 
  # Given a programme as a ref to a list of bits of text, each with
***************
*** 298,301 ****
--- 330,345 ----
  	$t =~ s/^\s+//; $t =~ s/\s+$//; next if $t eq '';
  
+ 	# Clean up some of the mess from the HTML
+ 	$t =~ s/\xa3((?:\d|,)*)/($1 ne '') ? "$1 pounds" : ''/ge;
+ 	$t =~ s/\Q&nbsp;\E/ /g;
+ 
+ 	# XML::Simple turns '&' into '&amp;', but nsgmls -wxml doesn't
+ 	# like this.  Kludge round the problem by getting rid of '&'
+ 	# altogether.
+ 	# 
+ 	$t =~ s/\Q\&amp\;\E/ and /g;
+ 	$t =~ s/\Q&\E/ and /g;
+ 
+ #	t("got bit of text '$t', bold=$b, italic=$i");
  	if ($b and not $i and istime($t)) {
  	    die "time already set to $time but $t found"
***************
*** 329,333 ****
  	    }
  	    elsif ($t =~ /^Director: (.+)\.$/) {
! 		die "italic text $t found, but director already defined"
  		  if defined $director;
  		$director = $t;
--- 373,377 ----
  	    }
  	    elsif ($t =~ /^Director: (.+)\.$/) {
! 		warn "italic text $t found, but director already defined - ignoring", next
  		  if defined $director;
  		$director = $t;
***************
*** 335,339 ****
  	    else {
  		# Must be sub-title.
! 		die "italic text $t found, but sub-title already defined"
  		  if defined $sub_title;
  		$sub_title = $t;
--- 379,383 ----
  	    else {
  		# Must be sub-title.
! 		warn "italic text $t found, but sub-title already defined - ignoring", next
  		  if defined $sub_title;
  		$sub_title = $t;
***************
*** 359,369 ****
      $title = '(unknown title)' if not defined $title;
  	
-     if (defined $desc) {
- 	# Clean up some of the mess from the HTML
- 	$desc =~ s/\xa3((?:\d|,)*)/($1 ne '') ? "$1 pounds" : ''/ge;
- 	$desc =~ s/&nbsp;/ /g;
- 	$desc =~ s/&amp;/ & /g;
-     }
- 
      return { time => $time, title => $title,
  	     sub_title => $sub_title, desc => $desc };
--- 403,406 ----
***************
*** 387,392 ****
      use HTML::TokeParser;
      my $p = new HTML::TokeParser(\$html) or die;
-     my @r = ();
  
      my ($bold, $it) = (0, 0);
      while (my $t = $p->get_token()) {
--- 424,435 ----
      use HTML::TokeParser;
      my $p = new HTML::TokeParser(\$html) or die;
  
+     # HTML::TokeParser is based on HTML::Parser, and has the same
+     # tendency to break chunks of text into multiple tokens.  Turn
+     # that off. 
+     # 
+     $p->unbroken_text(1);
+ 
+     my @r = ();
      my ($bold, $it) = (0, 0);
      while (my $t = $p->get_token()) {
***************
*** 454,468 ****
      die 'usage: istime(possible time string)' if @_ != 1;
      local $_ = shift;
      return /^(\d\d?)[.:]?(\d\d)(am|pm)$/;
  }
  
  
! # timecvt()
  # 
  # Convert a 12 hour time to 24 hour.
  # 
! sub timecvt($) {
!     die 'usage: timecvt(12 hour time string)' if @_ != 1;
      local $_ = shift;
      s/\s+//;
      my ($hours, $mins, $ampm) = /^(\d\d?)[.:]?(\d\d)(am|pm)$/
--- 497,521 ----
      die 'usage: istime(possible time string)' if @_ != 1;
      local $_ = shift;
+ 
+     my $tz = gettz($_);
+     s/\Q$tz\E// if defined $tz;
+ 
+     s/\s+//g;
      return /^(\d\d?)[.:]?(\d\d)(am|pm)$/;
  }
  
  
! # time12to24()
  # 
  # Convert a 12 hour time to 24 hour.
  # 
! sub time12to24($) {
!     die 'usage: time12to24(12 hour time string)' if @_ != 1;
      local $_ = shift;
+     
+     # Remove the timezone and stick it back on afterwards.
+     my $tz = gettz($_);
+     s/\Q$tz\E// if defined $tz;
+ 
      s/\s+//;
      my ($hours, $mins, $ampm) = /^(\d\d?)[.:]?(\d\d)(am|pm)$/
***************
*** 487,492 ****
      }
      else { die }
!     
!     return sprintf('%02d:%02d', $hours, $mins);
  }
  
--- 540,547 ----
      }
      else { die }
! 
!     my $r = sprintf('%02d:%02d', $hours, $mins);
!     $r .= " $tz" if defined $tz;
!     return $r;
  }
  
***************
*** 494,497 ****
--- 549,555 ----
  # get_date_from_page()
  # 
+ # Returns date in 'YYYYMMDD' format from Ananova page, or undef if
+ # none found.
+ # 
  sub get_date_from_page($) {
      die 'usage: get_date_from_page(html string)' if @_ != 1;
***************
*** 505,510 ****
  	    die "date already set to $date, but headline $text found"
  	      if defined $date;
! 	    $date = ParseDate($text)
  	      or die "bad date string $text\n";
  	}
      }
--- 563,569 ----
  	    die "date already set to $date, but headline $text found"
  	      if defined $date;
! 	    my $pd = ParseDate($text)
  	      or die "bad date string $text\n";
+ 	    $date = UnixDate($pd, '%Q');
  	}
      }
***************
*** 528,543 ****
        = @$p{qw(channel date end title sub_title desc director)};
      
      if (defined $end) {
! 	$w->startTag('programme', start   => $start,
! 		                  stop    => $end,
  		                  channel => $channel );
      }
      else {
! 	$w->startTag('programme', start   => $start,
  			          channel => $channel );
      }
  
      $w->dataElement('title', $title, lang => $LANG);
!     $w->dataElement('sub_title', $sub_title, lang => $LANG)
        if defined $sub_title;
  
--- 587,603 ----
        = @$p{qw(channel date end title sub_title desc director)};
      
+     # The dates are converted to the appropriate timezone for output.
      if (defined $end) {
! 	$w->startTag('programme', start   => join(' ', @{date_to_uk($start)}),
! 		                  stop    => join(' ', @{date_to_uk($end)}),
  		                  channel => $channel );
      }
      else {
! 	$w->startTag('programme', start   => join(' ', @{date_to_uk($start)}),
  			          channel => $channel );
      }
  
      $w->dataElement('title', $title, lang => $LANG);
!     $w->dataElement('sub-title', $sub_title, lang => $LANG)
        if defined $sub_title;
  

Index: listings_to_latex
===================================================================
RCS file: /cvsroot/xmltv/xmltv/listings_to_latex,v
retrieving revision 1.1
retrieving revision 1.2
diff -C2 -d -r1.1 -r1.2
*** listings_to_latex	2001/11/02 17:58:43	1.1
--- listings_to_latex	2001/11/02 18:07:18	1.2
***************
*** 25,28 ****
--- 25,29 ----
  use IO::File;
  use Date::Manip;
+ require 'tz.pl';
  
  die "usage: $0 [filename]" if @ARGV >= 2;
***************
*** 42,46 ****
  my $xml;
  if (@ARGV == 1) {
!     $xml = XMLin($ARGV[0], cache => [ 'storable' ]);
  }
  elsif (@ARGV == 0) {
--- 43,47 ----
  my $xml;
  if (@ARGV == 1) {
!     $xml = XMLin($ARGV[0], cache => [ 'storable' ], searchpath => '.');
  }
  elsif (@ARGV == 0) {
***************
*** 52,56 ****
  my @chunks = chunk(30, @programmes);
  
! my $curr_date;
  foreach (@chunks) {
      my @programmes = @$_;
--- 53,57 ----
  my @chunks = chunk(30, @programmes);
  
! my ($curr_date, $curr_tz);
  foreach (@chunks) {
      my @programmes = @$_;
***************
*** 68,72 ****
  	my ($start, $stop, $channel) = @h{qw(start stop channel)};
  	my $title = $h{title}->{content};
! 	my $sub_title = $h{sub_title}->{content};
  	
  	die 'no start'   if not defined $start;
--- 69,73 ----
  	my ($start, $stop, $channel) = @h{qw(start stop channel)};
  	my $title = $h{title}->{content};
! 	my $sub_title = $h{'sub-title'}->{content};
  	
  	die 'no start'   if not defined $start;
***************
*** 75,80 ****
  	die 'no channel' if not defined $channel;
  
! 	# Convert date-and-time to just hhmm
! 	foreach ($start, $stop) { s/^\d{8}(\d\d):(\d\d):00$/$1$2/ }
  
  	# Hardcoded channel lookup
--- 76,102 ----
  	die 'no channel' if not defined $channel;
  
! 	# Convert date-and-time to just hhmm, adding explicit timezone
! 	# whenever it changes.
! 	# 
! 	foreach (\$start, \$stop) {
! 	    my $p = ParseDate($$_) or die "bad date $$_";
! 	    my $tz = gettz($$_) || 'UT';
! 	    $$_ = UnixDate($p, '%R');
! 	    
! 	    if (not defined $curr_tz) {
! 		# First date on this page.  Just assume it doesn't
! 		# need printing (FIXME).
! 		# 
! 		$curr_tz = $tz;
! 	    }
! 	    elsif ($curr_tz ne $tz) {
! 		# Add timezone to printed hh:mm
! 		$$_ .= " $tz";
! 		$curr_tz = $tz;
! 	    }
! 	    else {
! 		# Timezone hasn't changed, don't print it again
! 	    }
! 	}
  
  	# Hardcoded channel lookup
***************
*** 125,135 ****
  
  
  sub start_date($) {
      my $prog = shift;
!     die "bad programme: " . d($prog) if not defined $prog->{title};
      die "no start date for $prog->{title}"
        if not defined $prog->{start};
!     local $_ = $prog->{start};
!     /^(\d{8})/ or die "bad date $_";
!     return $1;
  }
--- 147,162 ----
  
  
+ # start_date()
+ # 
+ # Returns the day on which a programme starts, in the form YYYYMMDD.
+ # 
  sub start_date($) {
      my $prog = shift;
!     die "bad programme" if not defined $prog->{title};
      die "no start date for $prog->{title}"
        if not defined $prog->{start};
! 
!     my $p = ParseDate($prog->{start});
!     die "bad start date for programme" if not defined $p or $p eq '';
!     return UnixDate($p, '%Q');
  }

Index: pick_cgi
===================================================================
RCS file: /cvsroot/xmltv/xmltv/pick_cgi,v
retrieving revision 1.1
retrieving revision 1.2
diff -C2 -d -r1.1 -r1.2
*** pick_cgi	2001/11/02 17:58:43	1.1
--- pick_cgi	2001/11/02 18:07:18	1.2
***************
*** 81,87 ****
  use XML::Simple;
  use Fcntl ':flock';
- use Dbg; $Dbg::On = 0; $Dbg::As_HTML = 1;
  use Date::Manip;
! 
  
  ########
--- 81,86 ----
  use XML::Simple;
  use Fcntl ':flock';
  use Date::Manip;
! #use Dbg; $Dbg::On = 0; $Dbg::As_HTML = 1;
  
  ########
***************
*** 136,140 ****
  seek PREFS, 0, 0;
  while (<PREFS>) {
!     t("got line from $PREFS_FILE: " . d($_));
      s/^\s+//; s/\s+$//;
      s/\#.*//;
--- 135,139 ----
  seek PREFS, 0, 0;
  while (<PREFS>) {
! #    t("got line from $PREFS_FILE: " . d($_));
      s/^\s+//; s/\s+$//;
      s/\#.*//;
***************
*** 145,149 ****
      else { die "$PREFS_FILE:$.: bad line (remnant is $_)\n" }
  }
! t('\%wanted=' . d(\%wanted));
  
  my ($skip, $next) = (url_param('skip'), url_param('next'));
--- 144,148 ----
      else { die "$PREFS_FILE:$.: bad line (remnant is $_)\n" }
  }
! #t('\%wanted=' . d(\%wanted));
  
  my ($skip, $next) = (url_param('skip'), url_param('next'));
***************
*** 151,155 ****
      die "bad URL parameter $_" if defined and tr/0-9//c;
  }
! t('$skip=' . d($skip) . ', $next=',  d($next));
  
  if (defined $skip and defined $next) {
--- 150,154 ----
      die "bad URL parameter $_" if defined and tr/0-9//c;
  }
! #t('$skip=' . d($skip) . ', $next=',  d($next));
  
  if (defined $skip and defined $next) {
***************
*** 278,282 ****
  	my $val = param("prog$i");
  	my $title = $programmes[$i]->{title}->{content};
! 	if ($wanted{$title} eq 'always' or $val eq 'yes') {
  	    print TOWATCH "$LISTINGS/$i\n";
  	    print br(), "Planning to watch $title\n";
--- 277,283 ----
  	my $val = param("prog$i");
  	my $title = $programmes[$i]->{title}->{content};
! 	if ((defined $wanted{$title} and $wanted{$title} eq 'always')
!             or (defined $val and $val eq 'yes') )
!         {
  	    print TOWATCH "$LISTINGS/$i\n";
  	    print br(), "Planning to watch $title\n";
***************
*** 403,412 ****
  # 
  sub print_date_for($$;) {
!     local $Dbg::On = 0;
      die 'usage: print_date_for(programme, [prev programme])'
        unless 1 <= @_ and @_ < 3;
      my ($prog, $prev) = @_;
!     t('$prog=' . d($prog));
!     t('$prev=' . d($prev));
  
      my $DAY_FMT = '%A'; # roughly as for date(1)
--- 404,413 ----
  # 
  sub print_date_for($$;) {
! #    local $Dbg::On = 0;
      die 'usage: print_date_for(programme, [prev programme])'
        unless 1 <= @_ and @_ < 3;
      my ($prog, $prev) = @_;
! #    t('$prog=' . d($prog));
! #    t('$prev=' . d($prev));
  
      my $DAY_FMT = '%A'; # roughly as for date(1)
***************
*** 414,419 ****
      my $day = UnixDate($prog->{start}, $DAY_FMT);
      my $prev_day = defined $prev ? UnixDate($prev->{start}, $DAY_FMT) : undef;
!     t('$day=' . d($day));
!     t('$prev_day=' . d($prev_day));
  
      if ((not defined $prev_day) or ($day ne $prev_day)) {
--- 415,420 ----
      my $day = UnixDate($prog->{start}, $DAY_FMT);
      my $prev_day = defined $prev ? UnixDate($prev->{start}, $DAY_FMT) : undef;
! #    t('$day=' . d($day));
! #    t('$prev_day=' . d($prev_day));
  
      if ((not defined $prev_day) or ($day ne $prev_day)) {

Index: xmltv.dtd
===================================================================
RCS file: /cvsroot/xmltv/xmltv/xmltv.dtd,v
retrieving revision 1.1
retrieving revision 1.2
diff -C2 -d -r1.1 -r1.2
*** xmltv.dtd	2001/11/02 17:58:43	1.1
--- xmltv.dtd	2001/11/02 18:07:18	1.2
***************
*** 1,9 ****
  <!-- DTD for TV listings
  
! This is a first draft of a DTD to represent a TV listing.  Rather than
! breaking the listing up by channel, it mixes all the channels
! together.  Data about a TV programme are stored in the subelements of
! element 'programme', but metadata such as when it will be broadcast
! are stored as attributes.
  
  Many of the programme details have a 'lang' attribute so that you can
--- 1,9 ----
  <!-- DTD for TV listings
  
! This is a DTD to represent a TV listing.  Rather than breaking the
! listing up by channel, it mixes all the channels together.  Data about
! a TV programme are stored in the subelements of element 'programme',
! but metadata such as when it will be broadcast are stored as
! attributes.
  
  Many of the programme details have a 'lang' attribute so that you can
***************
*** 37,41 ****
  With thanks to Gottfried Szing and others for suggestions.
  
! Ed Avis, ep...@do..., 2000-06-23
  -->
  
--- 37,48 ----
  With thanks to Gottfried Szing and others for suggestions.
  
! All dates and times in this DTD follow the same format, loosely based
! on ISO 8601.  They can be 'YYYYMMDDhhmmss' or some initial
! substring, for example if you only know the year and month you can
! have 'YYYYMM'.  You can also append a timezone to the end; if no
! explicit timezone is given, UT is assumed.  Examples:
! '200007281733 BST', '200209', '19880523083000 +0300'.  (BST == +0100.)
! 
! Ed Avis, ep...@do..., 2000-07-28
  -->
  
***************
*** 74,78 ****
  programmes on the same channel.
  -->
! <!ELEMENT programme (title+, sub_title*, desc*, credits?, date?,
                       category*, language?, orig-language?, length?,
                       url*, country*, episode-num? )>
--- 81,85 ----
  programmes on the same channel.
  -->
! <!ELEMENT programme (title+, sub-title*, desc*, credits?, date?,
                       category*, language?, orig-language?, length?,
                       url*, country*, episode-num? )>
***************
*** 91,96 ****
  To do: part number for multi-part episodes.
  -->
! <!ELEMENT sub_title (#PCDATA)>
! <!ATTLIST sub_title lang CDATA #IMPLIED>
  
  <!-- Description of the programme or episode.
--- 98,103 ----
  To do: part number for multi-part episodes.
  -->
! <!ELEMENT sub-title (#PCDATA)>
! <!ATTLIST sub-title lang CDATA #IMPLIED>
  
  <!-- Description of the programme or episode.
***************
*** 118,124 ****
  
  <!-- The date the programme or film was finished.  This will probably
! be the same as the copyright date.  The format is YYYYMMDDhhmmss, or
! some initial substring of that (eg YYYYMM if you only know to the
! nearest month).
  -->
  <!ELEMENT date (#PCDATA)>
--- 125,129 ----
  
  <!-- The date the programme or film was finished.  This will probably
! be the same as the copyright date.
  -->
  <!ELEMENT date (#PCDATA)>