#!/usr/bin/perl -w
use strict;
# Print People I Know, public release 4 or so, 1999-10-26
# minor change, 2006-12-20: add style and favicon links
# Take a list of people, formatted roughly as follows:
# Here are the people I know.
#
# Here is another paragraph about them.
# --
# sttng: Star Trek: The Next Generation
# James Kirk, Captain Kirk
# Captain of the original Enterprise; predecessor to Jean-Luc Picard.
#
# Jean-Luc Picard, Captain Picard
# Captain of the Enterprise in sttng; successor to Captain Kirk.
# and transform it into a set of web pages, one per person, with
# hyperlinks where the pages mention one another's names.
#
# sttng is a macro.
# The directory to create the HTML files in.
my $dir = shift @ARGV;
if ((not defined $dir) or not -d $dir) {
die "Usage: $0 dir [inputfiles]\n";
}
my @data = ();
my %macros = ();
my %names = ();
my @names = ();
my @para = ();
# the real data is preceded by an initial page, followed by a line
# containing just -- at the beginning of the line.
my @initpage = ();
my $doing_initpage = 1;
# when we reach a blank line or EOF, we put the currently-being-read
# name away and prepare to read a new one.
sub namedone {
if (@names or @para) {
if (not (@names and @para)) {
warn "Oops, boo-boo, line $.\n";
next;
}
my $data =
[
[
# split on commas and chop off white space
map { /^\s*(.*?)\s*$/ && $1 }
split /,/,
join " ", @names
],
# chop off white space and join with spaces
join " ", map { /^\s*(.*?)\s*$/ && $1 } @para
]
;
@names = ();
@para = ();
for (@{$data->[0]}) {
$names{$_} = $#data + 1;
}
push @data, $data;
} # else do nothing
}
# main input loop. We make one pass over the file, reading it all into
# memory; when we're done, we spit it all out (see below) with changes.
while (<>) {
chomp;
next if /^#/; # ignore comment lines
if ($doing_initpage) {
if (/^--\s*$/) {
$doing_initpage = 0;
} else {
push @initpage, ($_ || "
");
}
} else {
if (/^(\S.*?)\s*:\s*(.*?)\s*$/) {
$macros{$1} = $2;
} elsif (/^\s*$/) { # blank line
namedone;
} elsif (/^\s+/) {
push @para, $_;
} else {
push @names, $_;
}
}
}
# in case the last name didn't have a blank line after it
namedone;
# convert the lines of the initpage into one string
my $initpage = join " ", map { /^\s*(.*?)\s*$/ && $1 } @initpage;
# pass this guy the title of your page as an argument.
sub htmlbegin {
qq(
$_[0]
$_[0]
);
}
sub htmlend {"\n" }
# Figure out what filenames to assign to names.
my @filenames = ();
my %takenfnames = ("index" => 1);
for (0..$#data) {
my $fname = $data[$_]->[0]->[0];
$fname =~ tr/-A-Za-z0-9/-/c;
$fname =~ tr(A-Z)(a-z);
while (exists $takenfnames{$fname}) {
# this is not the ideal strategy if you have many people with the
# same primary name, but it works for me.
$fname .= "-x";
}
push @filenames, $fname;
}
# This expands macros and makes hyperlinks to names, returning the
# munged string.
sub munge {
my ($text) = @_;
# these commented-out statements were to try to debug a problem I was
# having with URLs.
# my $http = 0;
# do { warn "text is $text\n"; $http=1;} if $text =~ /http/;
$text =~ s#(http://[^ ]*[^ ,.])#$1#g;
local $_;
# sort in order of descending length to keep from replacing
# "Carolyn Hogg" before "Carolyn Hogg Sittler". (This doesn't
# work. Carolyn Hogg gets replaced anyway, presumably later.)
for (sort {(length $b) <=> (length $a)} keys %names) {
$text =~ s#\b\Q$_\E\b#$_#g;
}
for (keys %macros) {
$text =~ s/\Q$_\E/$macros{$_}/g;
}
$text =~ s#([-A-Za-z0-9.]*?@[-A-Za-z0-9.]*[-A-Za-z0-9])#$1#g;
# warn "text became $text\n" if $http;
return $text . "\n";
}
my $trailer = qq(
Index of people | Search people | email Kragen | Kragen's home page\n);
# Here we start to spit things out. We've read in all the names, so we
# can put a list of them here, and we can expand macros in the initpage.
open OUTFILE, ">$dir/index.html" or die "can't open $dir/index.html: $!\n";
print OUTFILE htmlbegin ("Index of people Kragen Sitaker knows");
print OUTFILE "
", munge ($initpage), "
\n";
for (sort keys %names) {
print OUTFILE "
", (munge $_), "
";
}
print OUTFILE "
\n", $trailer, htmlend;
close OUTFILE;
# create a text file suitable for searching
open OUTFILE, ">$dir/dir.txt" or die "can't open $dir/dir.txt: $!\n";
for (sort keys %names) {
print OUTFILE "$_,$filenames[$names{$_}].html,$data[$names{$_}]->[0]->[0]\n";
}
close OUTFILE;
# here we spit out the data from the actual names, in the order it was
# in the input file.
for (0..$#data) {
my $fname = $filenames[$_];
open OUTFILE, ">$dir/$fname.html" or die "can't open $dir/$fname.html: $!\n";
print OUTFILE htmlbegin "$data[$_]->[0]->[0]: known by Kragen Sitaker";
print OUTFILE "
", join (", ", @{$data[$_]->[0]}),
"
\n
", munge ($data[$_]->[1]), $trailer, htmlend;
close OUTFILE;
}
# BUGS:
#
# Once names, macros, URLs, and email addresses are expanded, they are
# vulnerable to further expansion in an unpredictable manner. This
# often produces syntactically invalid HTML.
#
# If you have several people with the same primary name (that being the
# first name in the list of their names), then adding or deleting one of
# them may cause the others to get new URLs.
#
# If you have several people with the same primary name, you will have
# long filenames:
# john-doe-x-x-x-x-x-x-x-x-x-x-x-x-x-x-x-x-x-x-x-x-x.html, etc. If you
# have a thousand people with the same primary name, one of them will
# have a filename that is at least 2000 characters long. This could be
# a problem under some circumstances.
#
# It has my name (Kragen Sitaker) hardcoded into it.
#
# It won't recognize all valid email addresses as email addresses.
#
# Recognizing URLs, it only recognizes http: URLs, and it requires them
# to be terminated by whitespace.
#
# You might wish you could put a blank line in the entry for a particular
# person. It won't let you do that, although you can put a
in.
# There's a cheat that lets you put blank lines in the initpage.
#
# It produces extremely long lines, which may not be ideal.
#
# It won't notice filesystem-full errors.
#
# It will accept funky syntax; in particular, you can interleave lines
# of names with lines of description about the person, and it will
# happily put all the names and all the description together, when it was
# probably actually meant to be two or more separate people.
#
# You have to create the directory for it, and if the directory already
# exists, you should clean it out.
#
# It won't bother to warn you if it is overwriting existing files.
#
# Searching through a list sorted by alphabetized first names is not the
# ideal user interface. I should add a grep.