X-Git-Url: http://git.hungrycats.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=driver%2Fxscreensaver-text;h=1d0170d9fbfaea30c2099a07807eace017082cc2;hb=d6b0217f2417bd19187f0ebc389d6c5c2233b11c;hp=e6a00226863743cca1e138b8051957ebafa38de7;hpb=07faf451b99879183ed7e909e43a0e065be1ee7f;p=xscreensaver diff --git a/driver/xscreensaver-text b/driver/xscreensaver-text index e6a00226..1d0170d9 100755 --- a/driver/xscreensaver-text +++ b/driver/xscreensaver-text @@ -1,5 +1,5 @@ #!/usr/bin/perl -w -# Copyright © 2005, 2006 Jamie Zawinski +# Copyright © 2005-2016 Jamie Zawinski # # Permission to use, copy, modify, distribute, and sell this software and its # documentation for any purpose is hereby granted without fee, provided that @@ -21,15 +21,23 @@ # Created: 19-Mar-2005. require 5; -use diagnostics; +#use diagnostics; # Fails on some MacOS 10.5 systems use strict; + +# Some Linux systems don't install LWP by default! +# Only error out if we're actually loading a URL instead of local data. +BEGIN { eval 'use LWP::UserAgent;' } + +# Not sure how prevalent this is. Hope it's part of the default install. +BEGIN { eval 'use HTML::Entities;' } + use Socket; use POSIX qw(strftime); use Text::Wrap qw(wrap); use bytes; my $progname = $0; $progname =~ s@.*/@@g; -my $version = q{ $Revision: 1.14 $ }; $version =~ s/^[^0-9]+([0-9.]+).*$/$1/; +my ($version) = ('$Revision: 1.44 $' =~ m/\s(\d[.\d]+)\s/s); my $verbose = 0; my $http_proxy = undef; @@ -39,109 +47,68 @@ my $text_mode = 'date'; my $text_literal = ''; my $text_file = ''; my $text_program = ''; -my $text_url = ''; +my $text_url = 'https://en.wikipedia.org/w/index.php?title=Special:NewPages&feed=rss'; +# Default URL needs to be set and match what's in OSX/XScreenSaverView.m -my $wrap_columns = undef; +my $wrap_columns = undef; +my $truncate_lines = undef; +my $latin1_p = 0; my $nyarlathotep_p = 0; -# Maps HTML character entities to the corresponding Latin1 characters. -# -my %entity_table = ( - "quot" => '"', "amp" => '&', "lt" => '<', "gt" => '>', - "nbsp" => ' ', "iexcl" => '¡', "cent" => '¢', "pound" => '£', - "curren" => '¤', "yen" => '¥', "brvbar" => '¦', "sect" => '§', - "uml" => '¨', "copy" => '©', "ordf" => 'ª', "laquo" => '«', - "not" => '¬', "shy" => '­', "reg" => '®', "macr" => '¯', - "deg" => '°', "plusmn" => '±', "sup2" => '²', "sup3" => '³', - "acute" => '´', "micro" => 'µ', "para" => '¶', "middot" => '·', - "cedil" => '¸', "sup1" => '¹', "ordm" => 'º', "raquo" => '»', - "frac14" => '¼', "frac12" => '½', "frac34" => '¾', "iquest" => '¿', - "Agrave" => 'À', "Aacute" => 'Á', "Acirc" => 'Â', "Atilde" => 'Ã', - "Auml" => 'Ä', "Aring" => 'Å', "AElig" => 'Æ', "Ccedil" => 'Ç', - "Egrave" => 'È', "Eacute" => 'É', "Ecirc" => 'Ê', "Euml" => 'Ë', - "Igrave" => 'Ì', "Iacute" => 'Í', "Icirc" => 'Î', "Iuml" => 'Ï', - "ETH" => 'Ð', "Ntilde" => 'Ñ', "Ograve" => 'Ò', "Oacute" => 'Ó', - "Ocirc" => 'Ô', "Otilde" => 'Õ', "Ouml" => 'Ö', "times" => '×', - "Oslash" => 'Ø', "Ugrave" => 'Ù', "Uacute" => 'Ú', "Ucirc" => 'Û', - "Uuml" => 'Ü', "Yacute" => 'Ý', "THORN" => 'Þ', "szlig" => 'ß', - "agrave" => 'à', "aacute" => 'á', "acirc" => 'â', "atilde" => 'ã', - "auml" => 'ä', "aring" => 'å', "aelig" => 'æ', "ccedil" => 'ç', - "egrave" => 'è', "eacute" => 'é', "ecirc" => 'ê', "euml" => 'ë', - "igrave" => 'ì', "iacute" => 'í', "icirc" => 'î', "iuml" => 'ï', - "eth" => 'ð', "ntilde" => 'ñ', "ograve" => 'ò', "oacute" => 'ó', - "ocirc" => 'ô', "otilde" => 'õ', "ouml" => 'ö', "divide" => '÷', - "oslash" => 'ø', "ugrave" => 'ù', "uacute" => 'ú', "ucirc" => 'û', - "uuml" => 'ü', "yacute" => 'ý', "thorn" => 'þ', "yuml" => 'ÿ', - "apos" => '\'' -); - -# Maps certain UTF8 characters (2 or 3 bytes) to the corresponding -# Latin1 characters. -# -my %unicode_latin1_table = ( - "\xC2\xA1" => '¡', "\xC2\xA2" => '¢', "\xC2\xA3" => '£', "\xC2\xA4" => '¤', - "\xC2\xA5" => '¥', "\xC2\xA6" => '¦', "\xC2\xA7" => '§', "\xC2\xA8" => '¨', - "\xC2\xA9" => '©', "\xC2\xAA" => 'ª', "\xC2\xAB" => '«', "\xC2\xAC" => '¬', - "\xC2\xAD" => '­', "\xC2\xAE" => '®', "\xC2\xAF" => '¯', "\xC2\xB0" => '°', - "\xC2\xB1" => '±', "\xC2\xB2" => '²', "\xC2\xB3" => '³', "\xC2\xB4" => '´', - "\xC2\xB5" => 'µ', "\xC2\xB6" => '¶', "\xC2\xB7" => '·', "\xC2\xB8" => '¸', - "\xC2\xB9" => '¹', "\xC2\xBA" => 'º', "\xC2\xBB" => '»', "\xC2\xBC" => '¼', - "\xC2\xBD" => '½', "\xC2\xBE" => '¾', "\xC2\xBF" => '¿', "\xC3\x80" => 'À', - "\xC3\x81" => 'Á', "\xC3\x82" => 'Â', "\xC3\x83" => 'Ã', "\xC3\x84" => 'Ä', - "\xC3\x85" => 'Å', "\xC3\x86" => 'Æ', "\xC3\x87" => 'Ç', "\xC3\x88" => 'È', - "\xC3\x89" => 'É', "\xC3\x8A" => 'Ê', "\xC3\x8B" => 'Ë', "\xC3\x8C" => 'Ì', - "\xC3\x8D" => 'Í', "\xC3\x8E" => 'Î', "\xC3\x8F" => 'Ï', "\xC3\x90" => 'Ð', - "\xC3\x91" => 'Ñ', "\xC3\x92" => 'Ò', "\xC3\x93" => 'Ó', "\xC3\x94" => 'Ô', - "\xC3\x95" => 'Õ', "\xC3\x96" => 'Ö', "\xC3\x97" => '×', "\xC3\x98" => 'Ø', - "\xC3\x99" => 'Ù', "\xC3\x9A" => 'Ú', "\xC3\x9B" => 'Û', "\xC3\x9C" => 'Ü', - "\xC3\x9D" => 'Ý', "\xC3\x9E" => 'Þ', "\xC3\x9F" => 'ß', "\xC3\xA0" => 'à', - "\xC3\xA1" => 'á', "\xC3\xA2" => 'â', "\xC3\xA3" => 'ã', "\xC3\xA4" => 'ä', - "\xC3\xA5" => 'å', "\xC3\xA6" => 'æ', "\xC3\xA7" => 'ç', "\xC3\xA8" => 'è', - "\xC3\xA9" => 'é', "\xC3\xAA" => 'ê', "\xC3\xAB" => 'ë', "\xC3\xAC" => 'ì', - "\xC3\xAD" => 'í', "\xC3\xAE" => 'î', "\xC3\xAF" => 'ï', "\xC3\xB0" => 'ð', - "\xC3\xB1" => 'ñ', "\xC3\xB2" => 'ò', "\xC3\xB3" => 'ó', "\xC3\xB4" => 'ô', - "\xC3\xB5" => 'õ', "\xC3\xB6" => 'ö', "\xC3\xB7" => '÷', "\xC3\xB8" => 'ø', - "\xC3\xB9" => 'ù', "\xC3\xBA" => 'ú', "\xC3\xBB" => 'û', "\xC3\xBC" => 'ü', - "\xC3\xBD" => 'ý', "\xC3\xBE" => 'þ', "\xC3\xBF" => 'ÿ', - - "\xE2\x80\x93" => '--', "\xE2\x80\x94" => '--', - "\xE2\x80\x98" => '`', "\xE2\x80\x99" => '\'', - "\xE2\x80\x9C" => "``", "\xE2\x80\x9D" => "''", - "\xE2\x80\xA6" => '...', -); - - # Convert any HTML entities to Latin1 characters. # sub de_entify($) { my ($text) = @_; - $text =~ s/(&(\#)?([[:alpha:]\d]+);?)/ - { - my $c; - if ($2) { - $c = chr($3); # the &#number is always decimal, right? - } else { - $c = $entity_table{$3}; - } -# print STDERR "$progname: warning: unknown HTML character entity \"$1\"\n" -# unless $c; - ($c ? $c : "[$3]"); - } - /gexi; - return $text; + + return '' unless defined($text); + return $text unless ($text =~ m/&/s); + + # Convert any HTML entities to Unicode characters, + # if the HTML::Entities module is installed. + eval { + my $t2 = $text; + $text = undef; + $text = HTML::Entities::decode_entities ($t2); + }; + return $text if defined($text); + + # If it's not installed, just complain instead of trying to halfass it. + print STDOUT ("\n\tPerl is broken. Do this to repair it:\n" . + "\n\tsudo cpan HTML::Entities\n\n"); + exit (1); } # Convert any Unicode characters to Latin1 if possible. # Unconvertable bytes are left alone. # -sub de_unicoddle($) { +sub utf8_to_latin1($) { my ($text) = @_; - foreach my $key (keys (%unicode_latin1_table)) { - my $val = $unicode_latin1_table{$key}; - $text =~ s/$key/$val/gs; - } + + utf8::encode ($text); # Unpack Unicode back to multi-byte UTF-8. + + # Maybe it would be better to handle this in the Unicode domain + # by doing things like s/\x{2018}/\"/g, but without decoding the + # string back to UTF-8 first, I'm at a loss as to how to have + # "á" print as "\340" instead of as "\303\240". + + $text =~ s/ \xC2 ( [\xA0-\xFF] ) / $1 /gsex; + $text =~ s/ \xC3 ( [\x80-\xFF] ) / chr (ord($1) | 0x40) /gsex; + + # Handles a few 3-byte sequences too. + $text =~ s/\xE2\x80\x93/--/gs; + $text =~ s/\xE2\x80\x94/--/gs; + $text =~ s/\xE2\x80\x98/`/gs; + $text =~ s/\xE2\x80\x99/'/gs; + $text =~ s/\xE2\x80\x9C/``/gs; + $text =~ s/\xE2\x80\x9D/'/gs; + $text =~ s/\xE2\x80\xA2/•/gs; + $text =~ s/\xE2\x80\xA6/.../gs; + $text =~ s/\xE2\x80\xB2/'/gs; + $text =~ s/\xE2\x84\xA2/™/gs; + $text =~ s/\xE2\x86\x90/ ← /gs; + return $text; } @@ -150,13 +117,12 @@ sub de_unicoddle($) { # sub get_x11_prefs() { my $got_any_p = 0; - local *IN; - if (open (IN, "<$config_file")) { + if (open (my $in, '<', $config_file)) { print STDERR "$progname: reading $config_file\n" if ($verbose > 1); - my $body = ''; - while () { $body .= $_; } - close IN; + local $/ = undef; # read entire file + my $body = <$in>; + close $in; $got_any_p = get_x11_prefs_1 ($body); } elsif ($verbose > 1) { @@ -174,15 +140,16 @@ sub get_x11_prefs() { } if ($verbose > 1) { - printf STDERR "$progname: mode: $text_mode\n"; - printf STDERR "$progname: literal: $text_literal\n"; - printf STDERR "$progname: file: $text_file\n"; - printf STDERR "$progname: program: $text_program\n"; - printf STDERR "$progname: url: $text_url\n"; + print STDERR "$progname: mode: $text_mode\n"; + print STDERR "$progname: literal: $text_literal\n"; + print STDERR "$progname: file: $text_file\n"; + print STDERR "$progname: program: $text_program\n"; + print STDERR "$progname: url: $text_url\n"; } $text_mode =~ tr/A-Z/a-z/; $text_literal =~ s@\\n@\n@gs; + $text_literal =~ s@\\\n@\n@gs; } @@ -191,6 +158,7 @@ sub get_x11_prefs_1($) { my $got_any_p = 0; $body =~ s@\\\n@@gs; + $body =~ s@^[ \t]*#[^\n]*$@@gm; if ($body =~ m/^[.*]*textMode:[ \t]*([^\s]+)\s*$/im) { $text_mode = $1; @@ -223,15 +191,21 @@ sub get_cocoa_prefs($) { $text_mode = $v if defined ($v); # The "textMode" pref is set to a number instead of a string because I - # can't figure out the black magic to make Cocoa bindings work right. + # couldn't figure out the black magic to make Cocoa bindings work right. + # + # Update: as of 5.33, Cocoa writes strings instead of numbers, but + # pre-existing saved preferences might still have numbers in them. # if ($text_mode eq '0') { $text_mode = 'date'; } elsif ($text_mode eq '1') { $text_mode = 'literal'; } elsif ($text_mode eq '2') { $text_mode = 'file'; } elsif ($text_mode eq '3') { $text_mode = 'url'; } + elsif ($text_mode eq '4') { $text_mode = 'program'; } $v = get_cocoa_pref_1 ($id, "textLiteral"); $text_literal = $v if defined ($v); + $text_literal =~ s@\\n@\n@gs; + $text_literal =~ s@\\\n@\n@gs; $v = get_cocoa_pref_1 ($id, "textFile"); $text_file = $v if defined ($v); @@ -304,6 +278,8 @@ sub which($) { sub output() { + binmode (STDOUT, ($latin1_p ? ':raw' : ':utf8')); + # Do some basic sanity checking (null text, null file names, etc.) # if (($text_mode eq 'literal' && $text_literal =~ m/^\s*$/i) || @@ -316,6 +292,7 @@ sub output() { if ($text_mode eq 'literal') { $text_literal = strftime ($text_literal, localtime); + $text_literal = utf8_to_latin1($text_literal) if ($latin1_p); $text_literal =~ y/A-Za-z/N-ZA-Mn-za-m/ if ($nyarlathotep_p); print STDOUT $text_literal; print STDOUT "\n" unless ($text_literal =~ m/\n$/s); @@ -324,23 +301,25 @@ sub output() { $text_file =~ s@^~/@$ENV{HOME}/@s; # allow literal "~/" - local *IN; - if (open (IN, "<$text_file")) { + if (open (my $in, '<:raw', $text_file)) { print STDERR "$progname: reading $text_file\n" if ($verbose); + binmode (STDOUT, ':raw'); - if ($wrap_columns && $wrap_columns > 0) { + if (($wrap_columns && $wrap_columns > 0) || $truncate_lines) { # read it, then reformat it. - my $body = ''; - while () { $body .= $_; } - reformat_text ($body); + local $/ = undef; # read entire file + my $body = <$in>; + $body = reformat_text ($body); + print STDOUT $body; } else { - # stream it - while () { + # stream it by lines + while (<$in>) { + $_ = utf8_to_latin1($_) if ($latin1_p); y/A-Za-z/N-ZA-Mn-za-m/ if ($nyarlathotep_p); - print $_; + print STDOUT $_; } } - close IN; + close $in; } else { error ("$text_file: $!"); } @@ -351,10 +330,21 @@ sub output() { $text_program = which ($prog) . $args; print STDERR "$progname: running $text_program\n" if ($verbose); - if ($wrap_columns && $wrap_columns > 0) { + if (($wrap_columns && $wrap_columns > 0) || $truncate_lines) { # read it, then reformat it. - my $body = `( $text_program ) 2>&1`; - reformat_text ($body); + my $lines = 0; + my $body = ""; + my $cmd = "( $text_program ) 2>&1"; + # $cmd .= " | sed -l"; # line buffer instead of 4k pipe buffer + open (my $pipe, '-|:unix', $cmd); + while (my $line = <$pipe>) { + $body .= $line; + $lines++; + last if ($truncate_lines && $lines > $truncate_lines); + } + close $pipe; + $body = reformat_text ($body); + print STDOUT $body; } else { # stream it safe_system ("$text_program"); @@ -366,12 +356,14 @@ sub output() { } else { # $text_mode eq 'date' - safe_system ("uname", "-n"); + my $n = `uname -n`; + $n =~ s/\.local\n/\n/s; + print $n; my $unamep = 1; if (-f "/etc/redhat-release") { # "Fedora Core release 4 (Stentz)" - system ("cat", "/etc/redhat-release"); + safe_system ("cat", "/etc/redhat-release"); } if (-f "/etc/release") { # "Solaris 10 3/05 s10_74L2a X86" @@ -380,10 +372,11 @@ sub output() { if (-f "/usr/sbin/system_profiler") { # "Mac OS X 10.4.5 (8H14)" my $sp = # "iMac G5" - `/usr/sbin/system_profiler SPSoftwareDataType SPHardwareDataType`; + `/usr/sbin/system_profiler SPSoftwareDataType SPHardwareDataType 2>/dev/null`; + # system_profiler on OS X 10.10 generates spurious error messages. my ($v) = ($sp =~ m/^\s*System Version:\s*(.*)$/mi); - my ($s) = ($sp =~ m/^\s*CPU Speed:\s*(.*)$/mi); - my ($t) = ($sp =~ m/^\s*Machine Name:\s*(.*)$/mi); + my ($s) = ($sp =~ m/^\s*(?:CPU|Processor) Speed:\s*(.*)$/mi); + my ($t) = ($sp =~ m/^\s*(?:Machine|Model) Name:\s*(.*)$/mi); print "$v\n" if ($v); print "$s $t\n" if ($s && $t); $unamep = !defined ($v); @@ -405,166 +398,6 @@ sub output() { } -# Loads the given URL, returns: $http, $head, $body. -# -sub get_url_1($;$) { - my ($url, $referer) = @_; - - if (! ($url =~ m@^(http|feed)://@i)) { - error ("not an HTTP URL: $url"); - } - - my ($url_proto, $dummy, $serverstring, $path) = split(/\//, $url, 4); - $path = "" unless $path; - - my ($them,$port) = split(/:/, $serverstring); - $port = 80 unless $port; - - my $them2 = $them; - my $port2 = $port; - if ($http_proxy) { - $serverstring = $http_proxy if $http_proxy; - $serverstring =~ s@^[a-z]+://@@; - ($them2,$port2) = split(/:/, $serverstring); - $port2 = 80 unless $port2; - } - - my ($remote, $iaddr, $paddr, $proto, $line); - $remote = $them2; - if ($port2 =~ /\D/) { $port2 = getservbyname($port2, 'tcp') } - if (!$port2) { - error ("unrecognised port in $url"); - } - - $iaddr = inet_aton($remote); - error ("host not found: $remote") unless ($iaddr); - - $paddr = sockaddr_in($port2, $iaddr); - - - my $head = ""; - my $body = ""; - - $proto = getprotobyname('tcp'); - if (!socket(S, PF_INET, SOCK_STREAM, $proto)) { - error ("socket: $!"); - } - if (!connect(S, $paddr)) { - error ("connect($serverstring): $!"); - } - - select(S); $| = 1; select(STDOUT); - - my $user_agent = "$progname/$version"; - - my $hdrs = ("GET " . ($http_proxy ? $url : "/$path") . " HTTP/1.0\r\n" . - "Host: $them\r\n" . - "User-Agent: $user_agent\r\n"); - if ($referer) { - $hdrs .= "Referer: $referer\r\n"; - } - $hdrs .= "\r\n"; - - if ($verbose > 3) { - foreach (split('\r?\n', $hdrs)) { - print STDERR " ==> $_\n"; - } - } - print S $hdrs; - my $http = || ""; - - $_ = $http; - s/[\r\n]+$//s; - print STDERR " <== $_\n" if ($verbose > 3); - - while () { - $head .= $_; - s/[\r\n]+$//s; - last if m@^$@; - print STDERR " <== $_\n" if ($verbose > 3); - } - - print STDERR " <== \n" if ($verbose > 4); - my $lines = 0; - while () { - s/\r\n/\n/gs; - print STDERR " <== $_" if ($verbose > 4); - $body .= $_; - $lines++; - } - - print STDERR " <== [ body ]: $lines lines, " . length($body) . " bytes\n" - if ($verbose == 4); - - close S; - - if (!$http) { - error ("null response: $url"); - } - - return ( $http, $head, $body ); -} - - -# Loads the given URL, processes redirects, returns (content-type, body). -# -sub get_url($;$) { - my ($url, $referer) = @_; - - print STDERR "$progname: loading $url\n" if ($verbose > 2); - - my $orig_url = $url; - my $loop_count = 0; - my $max_loop_count = 10; - - do { - my ( $http, $head, $body ) = get_url_1 ($url, $referer); - - $http =~ s/[\r\n]+$//s; - - if ( $http =~ m@^HTTP/[0-9.]+ 30[123]@ ) { - $_ = $head; - - my ( $location ) = m@^location:[ \t]*(.*)$@im; - if ( $location ) { - $location =~ s/[\r\n]$//; - - print STDERR "$progname: redirect from $url to $location\n" - if ($verbose > 3); - - $referer = $url; - $url = $location; - - if ($url =~ m@^/@) { - $referer =~ m@^(http://[^/]+)@i; - $url = $1 . $url; - } elsif (! ($url =~ m@^[a-z]+:@i)) { - $_ = $referer; - s@[^/]+$@@g if m@^http://[^/]+/@i; - $_ .= "/" if m@^http://[^/]+$@i; - $url = $_ . $url; - } - - } else { - error ("no Location with \"$http\""); - } - - if ($loop_count++ > $max_loop_count) { - error ("too many redirects ($max_loop_count) from $orig_url"); - } - - } elsif ( $http =~ m@^HTTP/[0-9.]+ ([4-9][0-9][0-9].*)$@ ) { - error ("failed: $1 ($url)"); - - } else { - my $ct = 'text/plain'; - $ct = $1 if ($head =~ m/^content-type:\s*([^\s]+)/mi); - return ($ct, $body); - } - } while (1); -} - - # Make an educated guess as to what's in this document. # We don't necessarily take the Content-Type header at face value. # Returns 'html', 'rss', or 'text'; @@ -592,16 +425,31 @@ sub guess_content_type($$) { return 'text'; } + sub reformat_html($$) { my ($body, $rss_p) = @_; $_ = $body; + # In HTML, try to preserve newlines inside of PRE. + # + if (! $rss_p) { + s@(]*>\s*)(.*?)(/gs; + $a . $b . $c; + }@gsexi; + } + if (! $rss_p) { - # In HTML, unfold lines (this breaks PRE. Sue me.) + # In HTML, unfold lines. # In RSS, assume \n means literal line break. s@[\r\n]@ @gsi; } + # This right here is the part where I doom us all to inhuman + # toil for the One whose Name cannot be expressed in the + # Basic Multilingual Plane. http://jwz.org/b/yhAT He comes. + s@@@gsi; # lose comments s@<(STYLE|SCRIPT)\b[^<>]*>.*?@@gsi; # lose css and js @@ -615,9 +463,57 @@ sub reformat_html($$) { s@<[^<>]*>?@@gs; # lose all other HTML tags $_ = de_entify ($_); # convert HTML entities - # elide any remaining non-Latin1 binary data... - s/([\177-\377]+(\s*[\177-\377]+)[^a-z\d]*)/«...» /g; - #s/([\177-\377]+(\s*[\177-\377]+)[^a-z\d]*)/«$1» /g; + # For Wikipedia: delete anything inside {{ }} and unwrap [[tags]], + # among other things. + # + if ($rss_p eq 'wiki') { + + s@@@gsi; # lose HTML comments again + + # Creation line is often truncated: screws up parsing with unbalanced {{. + s@(: +[^a-zA-Z ]* *Created page) with [^\n]+@$1@s; + + s@/\*.*?\*/@@si; # /* ... */ + + # Try to omit all tables, since they're impossible to read as text. + # + 1 while (s/\{\{[^{}]*}}/ /gs); # {{ ... }} + 1 while (s/\{\|.*?\|\}/\n\n/gs); # {| ... |} + 1 while (s/\|-.*?\|/ /gs); # |- ... | (table cell) + + # Convert anchors to something more readable. + # + s/\[\[([^\[\]\|]+)\|([^\[\]]+)\]\]/$2/gs; # [[link|anchor]] + s/\[\[([^:\[\]\|]+)\]\]/$1/gs; # [[anchor]] + s/\[https?:[^\[\]\s]+\s+([^\[\]]+)\]/$1/gs; # [url anchor] + + # Convert all references to asterisks. + s@\s*\s*.*?@*@gs; # ... -> "*" + s@\n[ \t]*\d+\s*\^\s*http[^\s]+[ \t]*\n@\n@gs; # 1 ^ URL (a Reflist) + + s@\[\[File:([^\|\]]+).*?\]\]@\n$1\n@gs; # [[File: X | ... ]] + s@\[\[Category:.*?\]\]@@gs; # omit categories + + s/<[^<>]*>//gs; # Omit all remaining tags + s/\'{3,}//gs; # Omit ''' and '''' + s/\'\'/\"/gs; # '' -> " + s/\`\`/\"/gs; # `` -> " + s/\"\"+/\"/gs; # "" -> " + + s/^[ \t]*[*#]+[ \t]*$//gm; # Omit lines with just * or # on them + + # Omit trailing headlines with no text after them (e.g. == Notes ==) + 1 while (s/\n==+[ \t]*[^\n=]+[ \t]*==+\s*$/\n/s); + + $_ = de_entify ($_); # convert HTML entities, again + } + + + # elide any remaining non-Latin1 binary data. + if ($latin1_p) { + utf8::encode ($_); # Unpack Unicode back to multi-byte UTF-8. + s/([^\000-\176]+(\s*[^\000-\176]+)[^a-z\d]*)/\xAB...\xBB /g; + } $_ .= "\n"; @@ -626,18 +522,29 @@ sub reformat_html($$) { if (!defined($wrap_columns) || $wrap_columns > 0) { $Text::Wrap::columns = ($wrap_columns || 72); - $_ = wrap ("", " ", $_); # wrap the lines as a paragraph - s/[ \t]+$//gm; # lose whitespace at end of line again + $Text::Wrap::break = '[\s/|]'; # wrap on slashes for URLs + $_ = wrap ("", " ", $_); # wrap the lines as a paragraph + s/[ \t]+$//gm; # lose whitespace at end of line again + } + + s/^\n+//gs; + + if ($truncate_lines) { + s/^(([^\n]*\n){$truncate_lines}).*$/$1/s; } + $_ = utf8_to_latin1($_) if ($latin1_p); y/A-Za-z/N-ZA-Mn-za-m/ if ($nyarlathotep_p); - print STDOUT $_; + + return $_; } sub reformat_rss($) { my ($body) = @_; + my $wiki_p = ($body =~ m@[^<>]*Wiki@si); + $body =~ s/(<(ITEM|ENTRY)\b)/\001\001$1/gsi; my @items = split (/\001\001/, $body); @@ -659,6 +566,8 @@ sub reformat_rss($) { } } + my $out = ''; + my $i = -1; foreach (@items) { $i++; @@ -694,16 +603,25 @@ sub reformat_rss($) { $title = rss_field_to_html ($title || ''); $body1 = rss_field_to_html ($body1 || ''); - reformat_html ("$title

$body1", 1); - print "\n"; + $title = '' if ($body1 eq $title); # Identical in Twitter's atom feed. + + $out .= reformat_html ("$title

$body1", $wiki_p ? 'wiki' : 'rss'); + $out .= "\n"; } + + if ($truncate_lines) { + $out =~ s/^(([^\n]*\n){$truncate_lines}).*$/$1/s; + } + + return $out; } sub rss_field_to_html($) { my ($body) = @_; - # Assume that if is present, everything inside that. + # If is present, everything inside that is HTML, + # and not double-encoded. # if ($body =~ m/^\s* 0) { print STDERR "$progname: wrapping at $wrap_columns...\n" if ($verbose > 2); $Text::Wrap::columns = $wrap_columns; + $Text::Wrap::break = '[\s/]'; # wrap on slashes for URLs $body = wrap ("", "", $body); $body =~ s/[ \t]+$//gm; } + if ($truncate_lines) { + $body =~ s/^(([^\n]*\n){$truncate_lines}).*$/$1/s; + } + + $body = utf8_to_latin1($body) if ($latin1_p); $body =~ y/A-Za-z/N-ZA-Mn-za-m/ if ($nyarlathotep_p); - print STDOUT $body; + return $body; +} + + +# Figure out what the proxy server should be, either from environment +# variables or by parsing the output of the (MacOS) program "scutil", +# which tells us what the system-wide proxy settings are. +# +sub set_proxy($) { + my ($ua) = @_; + + my $proxy_data = `scutil --proxy 2>/dev/null`; + foreach my $proto ('http', 'https') { + my ($server) = ($proxy_data =~ m/\b${proto}Proxy\s*:\s*([^\s]+)/si); + my ($port) = ($proxy_data =~ m/\b${proto}Port\s*:\s*([^\s]+)/si); + my ($enable) = ($proxy_data =~ m/\b${proto}Enable\s*:\s*([^\s]+)/si); + + if ($server && $enable) { + # Note: this ignores the "ExceptionsList". + my $proto2 = 'http'; + $ENV{"${proto}_proxy"} = ("${proto2}://" . $server . + ($port ? ":$port" : "") . "/"); + print STDERR "$progname: MacOS $proto proxy: " . + $ENV{"${proto}_proxy"} . "\n" + if ($verbose > 2); + } + } + + $ua->env_proxy(); } sub get_url_text($) { my ($url) = @_; - # historical suckage: the environment variable name is lower case. - $http_proxy = $ENV{http_proxy} || $ENV{HTTP_PROXY}; + my $ua = eval 'LWP::UserAgent->new'; + + if (! $ua) { + print STDOUT ("\n\tPerl is broken. Do this to repair it:\n" . + "\n\tsudo cpan LWP::UserAgent\n\n"); + return; + } + + set_proxy ($ua); + $ua->agent ("$progname/$version"); + my $res = $ua->get ($url); + my $body; + my $ct; + + if ($res && $res->is_success) { + $body = $res->decoded_content || ''; + $ct = $res->header ('Content-Type') || 'text/plain'; - if ($http_proxy && $http_proxy =~ m@^http://([^/]*)/?$@ ) { - # historical suckage: allow "http://host:port" as well as "host:port". - $http_proxy = $1; + } else { + my $err = ($res ? $res->status_line : '') || ''; + $err = 'unknown error' unless $err; + $err = "$url: $err"; + # error ($err); + $body = "Error loading URL $err\n\n"; + $ct = 'text/plain'; } - my ($ct, $body) = get_url ($url); + utf8::decode ($body); # Pack multi-byte UTF-8 back into wide chars. $ct = guess_content_type ($ct, $body); if ($ct eq 'html') { print STDERR "$progname: converting HTML...\n" if ($verbose > 2); - reformat_html ($body, 0); + $body = reformat_html ($body, 0); } elsif ($ct eq 'rss') { - reformat_rss ($body); + $body = reformat_rss ($body); } else { print STDERR "$progname: plain text...\n" if ($verbose > 2); - reformat_text ($body); + $body = reformat_text ($body); } + print STDOUT $body; } @@ -793,6 +764,10 @@ sub usage() { " it will be converted to plain-text.\n" . "\n" . " --cols N Wrap lines at this column. Default 72.\n" . + "\n" . + " --lines N No more than N lines of output.\n" . + "\n" . + " --latin1 Emit Latin1 instead of UTF-8.\n" . "\n"); exit 1; } @@ -809,19 +784,23 @@ sub main() { elsif (m/^--?date$/) { $text_mode = 'date'; $load_p = 0; } elsif (m/^--?text$/) { $text_mode = 'literal'; - $text_literal = shift @ARGV; + $text_literal = shift @ARGV || ''; + $text_literal =~ s@\\n@\n@gs; + $text_literal =~ s@\\\n@\n@gs; $load_p = 0; } elsif (m/^--?file$/) { $text_mode = 'file'; - $text_file = shift @ARGV; + $text_file = shift @ARGV || ''; $load_p = 0; } elsif (m/^--?program$/) { $text_mode = 'program'; - $text_program = shift @ARGV; + $text_program = shift @ARGV || ''; $load_p = 0; } elsif (m/^--?url$/) { $text_mode = 'url'; - $text_url = shift @ARGV; + $text_url = shift @ARGV || ''; $load_p = 0; } elsif (m/^--?col(umn)?s?$/) { $wrap_columns = 0 + shift @ARGV; } + elsif (m/^--?lines?$/) { $truncate_lines = 0 + shift @ARGV; } elsif (m/^--?cocoa$/) { $cocoa_id = shift @ARGV; } + elsif (m/^--?latin1$/) { $latin1_p++; } elsif (m/^--?nyarlathotep$/) { $nyarlathotep_p++; } elsif (m/^-./) { usage; } else { usage; } @@ -842,6 +821,42 @@ sub main() { } output(); + + + if (defined ($cocoa_id)) { + # + # On MacOS, sleep for 10 seconds between when the last output is + # printed, and when this process exits. This is because MacOS + # 10.5.0 and later broke ptys in a new and exciting way: basically, + # once the process at the end of the pty exits, you have exactly + # 1 second to read all the queued data off the pipe before it is + # summarily flushed. + # + # Many of the screen savers were written to depend on being able + # to read a small number of bytes, and continue reading until they + # reached EOF. This is no longer possible. + # + # Note that the current MacOS behavior has all four of these + # awesome properties: 1) Inconvenient; 2) Has no sane workaround; + # 3) Different behavior than MacOS 10.1 through 10.4; and 4) + # Different behavior than every other Unix in the world. + # + # See http://jwz.org/b/DHke, and for those of you inside Apple, + # "Problem ID 5606018". + # + # One workaround would be to rewrite the savers to have an + # internal buffer, and always read as much data as possible as + # soon as a pipe has input available. However, that's a lot more + # work, so instead, let's just not exit right away, and hope that + # 10 seconds is enough. + # + # This will solve the problem for invocations of xscreensaver-text + # that produce little output (e.g., date-mode); and won't solve it + # in cases where a large amount of text is generated in a short + # amount of time (e.g., url-mode.) + # + sleep (10); + } } main();