X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FFeed.pm;h=d7df07ee9c4577a1d15fe5410f68e6bdd68ee3e2;hb=be4d92e74dd0b1c8de85bc08c3c522af2d8d8656;hp=0b7ef7f84290633efdaf28b5ce231e0e8258cebd;hpb=7245596edd0167791e7324c2d34b7fd340a4557c;p=public-inbox.git diff --git a/lib/PublicInbox/Feed.pm b/lib/PublicInbox/Feed.pm index 0b7ef7f8..d7df07ee 100644 --- a/lib/PublicInbox/Feed.pm +++ b/lib/PublicInbox/Feed.pm @@ -1,125 +1,111 @@ -# Copyright (C) 2013, Eric Wong and all contributors -# License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt) +# Copyright (C) 2013-2018 all contributors +# License: AGPL-3.0+ +# +# Used for generating Atom feeds for web-accessible mailing list archives. package PublicInbox::Feed; use strict; use warnings; -use Email::Address; -use Email::MIME; -use Date::Parse qw(strptime); -use PublicInbox::Hval; -use PublicInbox::GitCatFile; +use PublicInbox::MIME; use PublicInbox::View; -use constant { - DATEFMT => '%Y-%m-%dT%H:%M:%SZ', # atom standard - MAX_PER_PAGE => 25, # this needs to be tunable - PRE_WRAP => "", -}; +use PublicInbox::WwwAtomStream; +use PublicInbox::SearchMsg; # this loads w/o Search::Xapian # main function sub generate { - my ($class, $ctx) = @_; - require XML::Atom::SimpleFeed; - require POSIX; - my $max = $ctx->{max} || MAX_PER_PAGE; + my ($ctx) = @_; + my $msgs = recent_msgs($ctx); + return _no_thread() unless @$msgs; - my $feed_opts = get_feedopts($ctx); - my $addr = $feed_opts->{address}; - $addr = $addr->[0] if ref($addr); - my $feed = XML::Atom::SimpleFeed->new( - title => $feed_opts->{description} || "unnamed feed", - link => $feed_opts->{url} || "http://example.com/", - link => { - rel => 'self', - href => $feed_opts->{atomurl} || - "http://example.com/atom.xml", - }, - id => 'mailto:' . ($addr || 'public-inbox@example.com'), - updated => POSIX::strftime(DATEFMT, gmtime), - ); - $feed->no_generator; + my $ibx = $ctx->{-inbox}; + PublicInbox::WwwAtomStream->response($ctx, 200, sub { + while (my $smsg = shift @$msgs) { + $ibx->smsg_mime($smsg) and return $smsg; + } + }); +} - my $git = PublicInbox::GitCatFile->new($ctx->{git_dir}); - each_recent_blob($ctx, sub { - my ($add, undef) = @_; - add_to_feed($feed_opts, $feed, $add, $git); +sub generate_thread_atom { + my ($ctx) = @_; + my $mid = $ctx->{mid}; + my $msgs = $ctx->{srch}->get_thread($mid); + return _no_thread() unless @$msgs; + + my $ibx = $ctx->{-inbox}; + my $html_url = $ibx->base_url($ctx->{env}); + $html_url .= PublicInbox::Hval->new_msgid($mid)->{href}; + $ctx->{-html_url} = $html_url; + PublicInbox::WwwAtomStream->response($ctx, 200, sub { + while (my $smsg = shift @$msgs) { + $ibx->smsg_mime($smsg) and return $smsg; + } }); - $git = undef; # destroy pipes - Email::Address->purge_cache; - $feed->as_string; } sub generate_html_index { - my ($class, $ctx) = @_; - - my $max = $ctx->{max} || MAX_PER_PAGE; - my $feed_opts = get_feedopts($ctx); - - my $title = $feed_opts->{description} || ''; - $title = PublicInbox::Hval->new_oneline($title)->as_html; + my ($ctx) = @_; + # if the 'r' query parameter is given, it is a legacy permalink + # which we must continue supporting: + my $qp = $ctx->{qp}; + if ($qp && !$qp->{r} && $ctx->{srch}) { + return PublicInbox::View::index_topics($ctx); + } - my $html = "$title" . - '{atomurl} . "\"\ntype=\"application/atom+xml\"/>" . - '' . PRE_WRAP; + my $env = $ctx->{env}; + my $url = $ctx->{-inbox}->base_url($env) . 'new.html'; + my $qs = $env->{QUERY_STRING}; + $url .= "?$qs" if $qs ne ''; + [302, [ 'Location', $url, 'Content-Type', 'text/plain'], + [ "Redirecting to $url\n" ] ]; +} - my $state; - my $git = PublicInbox::GitCatFile->new($ctx->{git_dir}); - my (undef, $last) = each_recent_blob($ctx, sub { - my ($path, $commit) = @_; - unless (defined $state) { - $state = [ $ctx->{srch}, {}, $commit, 0 ]; +sub new_html { + my ($ctx) = @_; + my $msgs = recent_msgs($ctx); + if (!@$msgs) { + return [404, ['Content-Type', 'text/plain'], + ["No messages, yet\n"] ]; + } + $ctx->{-html_tip} = '
';
+	$ctx->{-upfx} = '';
+	$ctx->{-hr} = 1;
+	my $ibx = $ctx->{-inbox};
+	PublicInbox::WwwStream->response($ctx, 200, sub {
+		while (my $smsg = shift @$msgs) {
+			my $m = $ibx->smsg_mime($smsg) or next;
+			my $more = scalar @$msgs;
+			return PublicInbox::View::index_entry($m, $ctx, $more);
 		}
-		my $mime = do_cat_mail($git, $_[0]) or return 0;
-		$html .= PublicInbox::View->index_entry($mime, 0, $state);
-		1;
+		PublicInbox::View::pagination_footer($ctx, './new.html');
 	});
-	Email::Address->purge_cache;
-	$git = undef; # destroy pipes.
-
-	my $footer = nav_footer($ctx->{cgi}, $last, $feed_opts, $state);
-	if ($footer) {
-		my $list_footer = $ctx->{footer};
-		$footer .= "\n" . $list_footer if $list_footer;
-		$footer = "
" . PRE_WRAP . "$footer
"; - } - $html . "$footer"; } # private subs -sub nav_footer { - my ($cgi, $last, $feed_opts, $state) = @_; - $cgi or return ''; - my $old_r = $cgi->param('r'); - my $head = ' '; - my $next = ' '; - my $first = $state->[2]; - my $anchor = $state->[3]; +sub _no_thread () { + [404, ['Content-Type', 'text/plain'], ["No feed found for thread\n"]]; +} - if ($last) { - $next = qq!next!; +sub recent_msgs { + my ($ctx) = @_; + my $ibx = $ctx->{-inbox}; + my $max = $ibx->{feedmax}; + my $qp = $ctx->{qp}; + my $v = $ibx->{version} || 1; + if ($v > 2) { + die "BUG: unsupported inbox version: $v\n"; } - if ($old_r) { - $head = $cgi->path_info; - $head = qq!head!; + if (my $srch = $ibx->search) { + return PublicInbox::View::paginate_recent($ctx, $max); } - my $atom = "{atomurl}\">atom"; - my $permalink = "permalink"; - "page: $next $head $atom $permalink"; -} -sub each_recent_blob { - my ($ctx, $cb) = @_; - my $max = $ctx->{max} || MAX_PER_PAGE; my $hex = '[a-f0-9]'; - my $addmsg = qr!^:000000 100644 \S+ \S+ A\t(${hex}{2}/${hex}{38})$!; - my $delmsg = qr!^:100644 000000 \S+ \S+ D\t(${hex}{2}/${hex}{38})$!; - my $refhex = qr/${hex}{4,40}(?:~\d+)?/; - my $cgi = $ctx->{cgi}; + my $addmsg = qr!^:000000 100644 \S+ (\S+) A\t${hex}{2}/${hex}{38}$!; + my $delmsg = qr!^:100644 000000 (\S+) \S+ D\t(${hex}{2}/${hex}{38})$!; + my $refhex = qr/(?:HEAD|${hex}{4,40})(?:~\d+)?/; # revision ranges may be specified my $range = 'HEAD'; - my $r = $cgi->param('r') if $cgi; + my $r = $qp->{r} if $qp; if ($r && ($r =~ /\A(?:$refhex\.\.)?$refhex\z/o)) { $range = $r; } @@ -127,157 +113,41 @@ sub each_recent_blob { # get recent messages # we could use git log -z, but, we already know ssoma will not # leave us with filenames with spaces in them.. - my @cmd = ('git', "--git-dir=$ctx->{git_dir}", - qw/log --no-notes --no-color --raw -r - --abbrev=16 --abbrev-commit/); - push @cmd, $range; - - my $pid = open(my $log, '-|', @cmd) or - die('open `'.join(' ', @cmd) . " pipe failed: $!\n"); + my $log = $ibx->git->popen(qw/log + --no-notes --no-color --raw -r + --no-abbrev --abbrev-commit/, + "--format=%h", $range); my %deleted; # only an optimization at this point my $last; - my $nr = 0; - my ($cur_commit, $first_commit, $last_commit); - while (my $line = <$log>) { + my $last_commit; + local $/ = "\n"; + my @oids; + while (defined(my $line = <$log>)) { if ($line =~ /$addmsg/o) { my $add = $1; next if $deleted{$add}; # optimization-only - $nr += $cb->($add, $cur_commit); - if ($nr >= $max) { + push @oids, $add; + if (scalar(@oids) >= $max) { $last = 1; last; } } elsif ($line =~ /$delmsg/o) { $deleted{$1} = 1; - } elsif ($line =~ /^commit (${hex}{7,40})/o) { - $cur_commit = $1; - $first_commit = $1 unless defined $first_commit; } } if ($last) { + local $/ = "\n"; while (my $line = <$log>) { - if ($line =~ /^commit (${hex}{7,40})/o) { + if ($line =~ /^(${hex}{7,40})/) { $last_commit = $1; last; } } } - close $log; # we may EPIPE here - # for pagination - ($first_commit, $last_commit); -} - -# private functions below -sub get_feedopts { - my ($ctx) = @_; - my $pi_config = $ctx->{pi_config}; - my $listname = $ctx->{listname}; - my $cgi = $ctx->{cgi}; - my %rv; - if (open my $fh, '<', "$ctx->{git_dir}/description") { - chomp($rv{description} = <$fh>); - close $fh; - } - - if ($pi_config && defined $listname && length $listname) { - foreach my $key (qw(address)) { - $rv{$key} = $pi_config->get($listname, $key) || ""; - } - } - my $url_base; - if ($cgi) { - my $path_info = $cgi->path_info; - my $base; - if (ref($cgi) eq 'CGI') { - $base = $cgi->url(-base); - } else { - $base = $cgi->base->as_string; - $base =~ s!/\z!!; - } - $url_base = $path_info; - if ($url_base =~ s!/(?:|index\.html)?\z!!) { - $rv{atomurl} = "$base$url_base/atom.xml"; - } else { - $url_base =~ s!/atom\.xml\z!!; - $rv{atomurl} = $base . $path_info; - $url_base = $base . $url_base; # XXX is this needed? - } - } else { - $url_base = "http://example.com"; - $rv{atomurl} = "$url_base/atom.xml"; - } - $rv{url} ||= "$url_base/"; - $rv{midurl} = "$url_base/m/"; - $rv{fullurl} = "$url_base/f/"; - - \%rv; -} - -sub mime_header { - my ($mime, $name) = @_; - PublicInbox::Hval->new_oneline($mime->header($name))->raw; -} - -sub feed_date { - my ($date) = @_; - my @t = eval { strptime($date) }; - - scalar(@t) ? POSIX::strftime(DATEFMT, @t) : 0; -} - -# returns 0 (skipped) or 1 (added) -sub add_to_feed { - my ($feed_opts, $feed, $add, $git) = @_; - - my $mime = do_cat_mail($git, $add) or return 0; - my $midurl = $feed_opts->{midurl} || 'http://example.com/m/'; - my $fullurl = $feed_opts->{fullurl} || 'http://example.com/f/'; - - my $header_obj = $mime->header_obj; - my $mid = $header_obj->header_raw('Message-ID'); - defined $mid or return 0; - $mid = PublicInbox::Hval->new_msgid($mid); - my $href = $mid->as_href . '.html'; - my $content = PublicInbox::View->feed_entry($mime, $fullurl . $href); - defined($content) or return 0; - $mime = undef; - - my $subject = mime_header($header_obj, 'Subject') or return 0; - - my $from = mime_header($header_obj, 'From') or return 0; - my @from = Email::Address->parse($from); - my $name = $from[0]->name; - defined $name or $name = ""; - my $email = $from[0]->address; - defined $email or $email = ""; - - my $date = $header_obj->header('Date'); - $date = PublicInbox::Hval->new_oneline($date); - $date = feed_date($date->raw) or return 0; - $add =~ tr!/!!d; - my $h = '[a-f0-9]'; - my (@uuid5) = ($add =~ m!\A($h{8})($h{4})($h{4})($h{4})($h{12})!o); - - $feed->add_entry( - author => { name => $name, email => $email }, - title => $subject, - updated => $date, - content => { type => 'xhtml', content => $content }, - link => $midurl . $href, - id => 'urn:uuid:' . join('-', @uuid5), - ); - 1; -} - -sub do_cat_mail { - my ($git, $path) = @_; - my $mime = eval { - my $str = $git->cat_file("HEAD:$path"); - Email::MIME->new($str); - }; - $@ ? undef : $mime; + $ctx->{next_page} = "r=$last_commit" if $last_commit; + [ map { bless {blob => $_ }, 'PublicInbox::SearchMsg' } @oids ]; } 1;