use URI::Escape qw/uri_escape/;
use Encode qw/encode decode/;
use Encode::MIME::Header;
-use DateTime::Format::Mail;
use CGI qw(escapeHTML);
use POSIX qw(strftime);
+use Date::Parse qw(strptime);
use constant DATEFMT => '%Y-%m-%dT%H:%M:%SZ';
-our $dt_parser = DateTime::Format::Mail->new(loose => 1);
+use PublicInbox::View;
+use Mail::Thread;
# main function
sub generate {
- my ($class, $git_dir, $max) = @_;
- $max ||= 25;
+ my ($class, $args) = @_;
+ my $max = $args->{max} || 25;
+ my $top = $args->{top}; # bool
- local $ENV{GIT_DIR} = $git_dir;
- my $feed_opts = get_feedopts();
+ local $ENV{GIT_DIR} = $args->{git_dir};
+ my $feed_opts = get_feedopts($args);
my $feed = XML::Atom::SimpleFeed->new(
- title => $feed_opts->{title},
+ title => $feed_opts->{description} || "unnamed feed",
link => $feed_opts->{url} || "http://example.com/",
link => {
rel => 'self',
- href => $feed_opts->{atomUrl} ||
+ href => $feed_opts->{atomurl} ||
"http://example.com/atom",
},
- id => $feed_opts->{email} || 'public-inbox@example.com',
+ id => $feed_opts->{address} || 'public-inbox@example.com',
updated => strftime(DATEFMT, gmtime),
);
- my @entries;
+ each_recent_blob($max, sub {
+ my ($add) = @_;
+ add_to_feed($feed_opts, $feed, $add, $top);
+ });
+ $feed->as_string;
+}
+
+sub generate_html_index {
+ my ($class, $args) = @_;
+ my $max = $args->{max} || 50;
+ my $top = $args->{top}; # bool
+ local $ENV{GIT_DIR} = $args->{git_dir};
+ my $feed_opts = get_feedopts($args);
+ my $title = escapeHTML($feed_opts->{description} || "");
+ my @messages;
+ each_recent_blob($max, sub {
+ my $str = `git cat-file blob $_[0]`;
+ return 0 if $? != 0;
+ my $simple = Email::Simple->new($str);
+ $simple->body_set(""); # save some memory
+ push @messages, $simple;
+ 1;
+ });
+
+ my $th = Mail::Thread->new(@messages);
+ $th->thread;
+ my @args = (
+ "<html><head><title>$title</title>" .
+ '<link rel=alternate title=Atom.feed href=' .
+ $feed_opts->{atomurl} . ' type="application/atom+xml"/>' .
+ '</head><body><pre>');
+ push @args, $feed_opts->{midurl};
+ dump_html_line($_, 0, \@args) for $th->rootset;
+ $args[0] . '</pre></html>';
+}
+
+# private subs
+
+sub each_recent_blob {
+ my ($max, $cb) = @_;
# get recent messages
# we could use git log -z, but, we already know ssoma will not
# leave us with filenames with spaces in them..
- my $cmd = "git log --no-color --raw -r --no-abbrev HEAD |";
+ my $cmd = "git log --no-notes --no-color --raw -r --no-abbrev HEAD |";
my $pid = open my $log, $cmd or die "open `$cmd' pipe failed: $!\n";
my %deleted;
my $nr = 0;
if ($line =~ /^:000000 100644 0{40} ([a-f0-9]{40})/) {
my $add = $1;
next if $deleted{$add};
- $nr += add_to_feed($feed_opts, $feed, $add);
+ $nr += $cb->($add);
last if $nr >= $max;
} elsif ($line =~ /^:100644 000000 ([a-f0-9]{40}) 0{40}/) {
$deleted{$1} = 1;
}
}
- close $log;
- $feed->as_string;
+ close $log;
}
# private functions below
sub get_feedopts {
+ my ($args) = @_;
+ my $pi_config = $args->{pi_config};
+ my $listname = $args->{listname};
+ my $cgi = $args->{cgi};
my %rv;
- foreach my $key (qw(title url atomUrl email)) {
- my $tmp = `git config publicInboxFeed.$key`;
- chomp $tmp;
- $rv{$key} = $tmp;
+
+ if ($pi_config && defined $listname && length $listname) {
+ foreach my $key (qw(description address)) {
+ $rv{$key} = $pi_config->get($listname, $key) || "";
+ }
+ }
+ my $url_base;
+ if ($cgi) {
+ my $cgi_url = $cgi->self_url;
+ $url_base = $cgi_url;
+ if ($url_base =~ s!/(?:|index\.html)?\z!!) {
+ $rv{atomurl} = "$url_base/index.atom.xml";
+ } else {
+ $url_base =~ s!/?(?:index|all)\.atom\.xml\z!!;
+ $rv{atomurl} = $cgi_url;
+ }
+ } else {
+ $url_base = "http://example.com";
+ $rv{atomurl} = "$url_base/index.atom.xml";
}
+ $rv{url} ||= "$url_base/";
+ $rv{midurl} = "$url_base/m/";
+ $rv{fullurl} = "$url_base/f/";
+
\%rv;
}
sub utf8_header {
- my ($mime, $name) = @_;
- encode('utf8', decode('MIME-Header', $mime->header($name)));
+ my ($simple, $name) = @_;
+ my $val = $simple->header($name);
+ return "" unless defined $val;
+ $val =~ tr/\t\r\n / /s;
+ encode('utf8', decode('MIME-Header', $val));
}
sub feed_date {
my ($date) = @_;
- my $dt = $dt_parser->parse_datetime($date);
- $dt ? $dt->strftime(DATEFMT) : 0;
+ my @t = eval { strptime($date) };
+
+ scalar(@t) ? strftime(DATEFMT, @t) : 0;
}
# returns 0 (skipped) or 1 (added)
sub add_to_feed {
- my ($feed_opts, $feed, $add) = @_;
+ my ($feed_opts, $feed, $add, $top) = @_;
# we can use git cat-file --batch if performance becomes a
# problem, but I doubt it...
return 0 if $? != 0;
my $mime = Email::MIME->new($str);
- my $content = msg_content($mime);
+ if ($top && $mime->header("In-Reply-To")) {
+ return 0;
+ }
+
+ my $midurl = $feed_opts->{midurl} || 'http://example.com/m/';
+ my $fullurl = $feed_opts->{fullurl} || 'http://example.com/f/';
+
+ my $content = PublicInbox::View->as_feed_entry($mime, $fullurl);
defined($content) or return 0;
- my $mid_url = $feed_opts->{mid_url} || "http://example.com/mid/%s";
my $mid = utf8_header($mime, "Message-ID") or return 0;
- $mid =~ s/\A<//;
- $mid =~ s/>\z//;
+ $mid =~ s/\A<//; $mid =~ s/>\z//;
my $subject = utf8_header($mime, "Subject") || "";
- defined($subject) && length($subject) or return 0;
+ length($subject) or return 0;
my $from = utf8_header($mime, "From") or return 0;
my $email = $from[0]->address;
defined $email or $email = "";
- my $url = sprintf($mid_url, uri_escape($mid));
+ my $url = $midurl . uri_escape($mid);
my $date = utf8_header($mime, "Date");
$date or return 0;
$date = feed_date($date) or return 0;
1;
}
-# returns a plain-text message body without quoted text
-# returns undef if there was nothing
-sub msg_content {
- my ($mime) = @_;
- my $rv;
-
- # scan through all parts, looking for displayable text
- $mime->walk_parts(sub {
- return if $rv;
- my ($part) = @_;
- return if $part->subparts; # walk_parts already recurses
- my $ct = $part->content_type || 'text/plain';
- return if $ct !~ m!\btext/[a-z0-9\+\._-]+\b!i;
- my @body;
- my $killed_wrote; # omit "So-and-so wrote:" line
-
- # no quoted text in Atom feed summary
- # $part->body should already be decoded for us (no QP)
-
- my $state = 0; # 0: beginning, 1: keep, 2: quoted
- foreach my $l (split(/\r?\n/, $part->body)) {
- if ($state == 0) {
- # drop leading blank lines
- next if $l =~ /\A\s*\z/;
-
- $state = ($l =~ /\A>/) ? 2 : 1; # fall-through
- }
- if ($state == 2) { # quoted text, drop it
- if ($l !~ /\A>/) {
- push @body, "<quoted text snipped>";
- if ($l =~ /\S/) {
- push @body, $l;
- }
- $state = 1;
- }
- }
- if ($state == 1) { # stuff we may keep
- if ($l =~ /\A>/) {
- # drop "So-and-so wrote:" line
- if (@body && !$killed_wrote &&
- $body[-1] =~ /:\z/) {
- $killed_wrote = 1;
- pop @body;
- }
- $state = 2;
- } else {
- push @body, $l;
- }
- }
- }
- $rv = "<pre>" .
- join("\n", map { escapeHTML($_) } @body) .
- "</pre>";
- });
-
- $rv;
+sub dump_html_line {
+ my ($self, $level, $args) = @_; # args => [ $html, $midurl ]
+ $args->[0] .= (' ' x $level);
+ if ($self->message) {
+ my $simple = $self->message;
+ my $subj = utf8_header($simple, "Subject");
+ my $mid = utf8_header($simple, "Message-ID");
+ $mid =~ s/\A<//;
+ $mid =~ s/>\z//;
+ my $url = $args->[1] . uri_escape($mid);
+ my $from = utf8_header($simple, "From");
+ my @from = Email::Address->parse($from);
+ $from = $from[0]->name;
+ (defined($from) && length($from)) or $from = $from[0]->address;
+ $from = escapeHTML($from);
+ $subj = escapeHTML($subj);
+ $args->[0] .= "<a href=$url>`-> $subj</a> $from\n";
+ } else {
+ $args->[0] .= "[ Message not available ]\n";
+ }
+ dump_html_line($self->child, $level+1, $args) if $self->child;
+ dump_html_line($self->next, $level, $args) if $self->next;
}
1;