X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FFeed.pm;h=e5d57550f414140fd7c1fc4f182bb8657b3e0169;hb=3fc59df0d633a17e0c5e43d633d12e8772c06ec3;hp=d34978c1e35a08445f5fc38445a491983bb3d004;hpb=f85f1781634dc03e093a0790a41bb4ae90ea907b;p=public-inbox.git
diff --git a/lib/PublicInbox/Feed.pm b/lib/PublicInbox/Feed.pm
index d34978c1..e5d57550 100644
--- a/lib/PublicInbox/Feed.pm
+++ b/lib/PublicInbox/Feed.pm
@@ -1,156 +1,126 @@
-# Copyright (C) 2013-2015, all contributors
+# Copyright (C) 2013-2015 all contributors
# License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt)
+#
+# 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 PublicInbox::MID qw/mid_clean mid_compress/;
-use constant {
- DATEFMT => '%Y-%m-%dT%H:%M:%SZ', # atom standard
- MAX_PER_PAGE => 25, # this needs to be tunable
-};
-
-use Encode qw/find_encoding/;
-my $enc_utf8 = find_encoding('UTF-8');
+use PublicInbox::WwwAtomStream;
# main function
sub generate {
my ($ctx) = @_;
- sub { emit_atom($_[0], $ctx) };
+ my @paths;
+ each_recent_blob($ctx, sub { push @paths, $_[0] });
+ return _no_thread() unless @paths;
+
+ my $ibx = $ctx->{-inbox};
+ PublicInbox::WwwAtomStream->response($ctx, 200, sub {
+ while (my $path = shift @paths) {
+ my $mime = do_cat_mail($ibx, $path) or next;
+ return $mime;
+ }
+ });
}
-sub generate_html_index {
+sub generate_thread_atom {
my ($ctx) = @_;
- sub { emit_html_index($_[0], $ctx) };
-}
-
-# private subs
-
-sub emit_atom {
- my ($cb, $ctx) = @_;
- require POSIX;
- my $fh = $cb->([ 200, ['Content-Type' => 'application/xml']]);
- my $max = $ctx->{max} || MAX_PER_PAGE;
- my $feed_opts = get_feedopts($ctx);
- my $addr = $feed_opts->{address};
- $addr = $addr->[0] if ref($addr);
- $addr ||= 'public-inbox@example.com';
- my $title = $feed_opts->{description} || "unnamed feed";
- $title = PublicInbox::Hval->new_oneline($title)->as_html;
- my $type = index($title, '&') >= 0 ? "\ntype=\"html\"" : '';
- my $url = $feed_opts->{url} || "http://example.com/";
- my $atomurl = $feed_opts->{atomurl};
- $fh->write(qq(\n) .
- qq{
'; + $ctx->{-upfx} = ''; + $ctx->{-hr} = 1; + PublicInbox::WwwStream->response($ctx, 200, sub { + while (my $path = shift @paths) { + my $m = do_cat_mail($ctx->{-inbox}, $path) or next; + my $more = scalar @paths; + my $s = PublicInbox::View::index_entry($m, $ctx, $more); + return $s; } + new_html_footer($ctx, $last); }); - 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; - $footer = "' + my ($ibx, $path) = @_; + my $mime = eval { $ibx->msg_by_path($path) } or return; + PublicInbox::MIME->new($mime); } 1;$footer"; - } - $fh->write(dump_topics($topics)) if $topics; - $fh->write("$footer"); - $fh->close; +# private subs + +sub _no_thread () { + [404, ['Content-Type', 'text/plain'], ["No feed found for thread\n"]]; } -sub nav_footer { - my ($cgi, $last, $feed_opts, $state) = @_; - $cgi or return ''; - my $old_r = $cgi->param('r'); - my $head = ' '; +sub new_html_footer { + my ($ctx, $last) = @_; + my $qp = delete $ctx->{qp} or return; + my $old_r = $qp->{r}; + my $latest = ''; my $next = ' '; - my $first = $state->{first_commit}; - my $anchor = $state->{anchor_idx}; if ($last) { - $next = qq!next!; + $next = qq!next!; } if ($old_r) { - $head = $cgi->path_info; - $head = qq!head!; + $latest = qq! latest!; } - my $atom = "{atomurl}\">atom"; - my $permalink = "permalink"; - "page: $next $head $atom $permalink"; + "page: $next$latest"; } sub each_recent_blob { my ($ctx, $cb) = @_; - my $max = $ctx->{max} || MAX_PER_PAGE; + my $max = $ctx->{-inbox}->{feedmax}; 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 $refhex = qr/(?:HEAD|${hex}{4,40})(?:~\d+)?/; - my $cgi = $ctx->{cgi}; + my $qp = $ctx->{qp}; # 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 = $r; } @@ -158,24 +128,22 @@ sub each_recent_blob { # get recent messages # we could use git log -z, but, we already know ssoma will not # leave us with filenames with spaces in them.. - my @cmd = ('git', "--git-dir=$ctx->{git_dir}", - qw/log --no-notes --no-color --raw -r - --abbrev=16 --abbrev-commit/, - "--format=%h%x00%ct%x00%an%x00%s%x00"); - push @cmd, $range; - - my $pid = open(my $log, '-|', @cmd) or - die('open `'.join(' ', @cmd) . " pipe failed: $!\n"); + my $log = $ctx->{-inbox}->git->popen(qw/log + --no-notes --no-color --raw -r + --abbrev=16 --abbrev-commit/, + "--format=%h%x00%ct%x00%an%x00%s%x00", + $range); my %deleted; # only an optimization at this point my $last; my $nr = 0; my ($cur_commit, $first_commit, $last_commit); my ($ts, $subj, $u); + local $/ = "\n"; while (defined(my $line = <$log>)) { if ($line =~ /$addmsg/o) { my $add = $1; next if $deleted{$add}; # optimization-only - $nr += $cb->($add, $cur_commit, $ts, $u, $subj); + $cb->($add, $cur_commit, $ts, $u, $subj) and $nr++; if ($nr >= $max) { $last = 1; last; @@ -191,6 +159,7 @@ sub each_recent_blob { } if ($last) { + local $/ = "\n"; while (my $line = <$log>) { if ($line =~ /^(${hex}{7,40})/o) { $last_commit = $1; @@ -199,181 +168,14 @@ sub each_recent_blob { } } - 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, $fh, $add, $git) = @_; - - my $mime = do_cat_mail($git, $add) or return 0; - my $fullurl = $feed_opts->{fullurl} || 'http://example.com/f/'; - - my $header_obj = $mime->header_obj; - my $mid = $header_obj->header('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 $title = mime_header($header_obj, 'Subject') or return 0; - $title = PublicInbox::Hval->new_oneline($title)->as_html; - my $type = index($title, '&') >= 0 ? "\ntype=\"html\"" : ''; - - my $from = mime_header($header_obj, 'From') or return 0; - my @from = Email::Address->parse($from) or return 0; - my $name = PublicInbox::Hval->new_oneline($from[0]->name)->as_html; - my $email = $from[0]->address; - $email = PublicInbox::Hval->new_oneline($email)->as_html; - - my $date = $header_obj->header('Date'); - $date = PublicInbox::Hval->new_oneline($date); - $date = feed_date($date->raw) or return 0; - - $fh->write(""); - 1; -} - sub do_cat_mail { - my ($git, $path) = @_; - my $mime = eval { - my $str = $git->cat_file("HEAD:$path"); - Email::MIME->new($str); - }; - $@ ? undef : $mime; -} - -# accumulate recent topics if search is supported -sub add_topic { - my ($git, $srch, $topics, $path, $ts, $u, $subj) = @_; - my ($order, $subjs) = @$topics; - my $header_obj; - - # legacy ssoma did not set commit titles based on Subject - $subj = $enc_utf8->decode($subj); - if ($subj eq 'mda') { - my $mime = do_cat_mail($git, $path) or return 0; - $header_obj = $mime->header_obj; - $subj = mime_header($header_obj, 'Subject'); - } - - my $topic = $subj = $srch->subject_normalized($subj); - - # kill "[PATCH v2]" etc. for summarization - $topic =~ s/\A\s*\[[^\]]+\]\s*//g; - - if (++$subjs->{$topic} == 1) { - unless ($header_obj) { - my $mime = do_cat_mail($git, $path) or return 0; - $header_obj = $mime->header_obj; - } - my $mid = $header_obj->header('Message-ID'); - $mid = mid_compress(mid_clean($mid)); - $u = $enc_utf8->decode($u); - push @$order, [ $mid, $ts, $u, $subj, $topic ]; - return 1; - } - 0; # old topic, continue going -} - -sub dump_topics { - my ($topics) = @_; - my ($order, $subjs) = @$topics; - my $dst = ''; - $dst .= "\n[No recent topics]" unless (scalar @$order); - while (defined(my $info = shift @$order)) { - my ($mid, $ts, $u, $subj, $topic) = @$info; - my $n = delete $subjs->{$topic}; - $mid = PublicInbox::Hval->new($mid)->as_href; - $subj = PublicInbox::Hval->new($subj)->as_html; - $u = PublicInbox::Hval->new($u)->as_html; - $dst .= "\n$subj\n- "; - $ts = POSIX::strftime('%Y-%m-%d %H:%M', gmtime($ts)); - if ($n == 1) { - $dst .= "created by $u @ $ts UTC\n" - } else { - # $n isn't the total number of posts on the topic, - # just the number of posts in the current "git log" - # window, so leave it unlabeled - $dst .= "updated by $u @ $ts UTC ($n)\n" - } - } - $dst .= ' $name $title " . - "$date " . - qq{} . - qq{ }. - "}); - $fh->write($content); - - $add =~ tr!/!!d; - my $h = '[a-f0-9]'; - my (@uuid5) = ($add =~ m!\A($h{8})($h{4})($h{4})($h{4})($h{12})!o); - my $id = 'urn:uuid:' . join('-', @uuid5); - my $midurl = $feed_opts->{midurl} || 'http://example.com/m/'; - $fh->write(qq{$id