X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FMsgIter.pm;h=6c18d2bf3102c56e070d8a1ff8dbce44b47e847d;hb=95bdac7f09c69036efed537a4d03d5bdd2ae4eb6;hp=ef0d209f6b391efc45ab843c3c8154a5cbd66557;hpb=c1de840e23416eea63739b21995b07b2e8a1e9ac;p=public-inbox.git diff --git a/lib/PublicInbox/MsgIter.pm b/lib/PublicInbox/MsgIter.pm index ef0d209f..6c18d2bf 100644 --- a/lib/PublicInbox/MsgIter.pm +++ b/lib/PublicInbox/MsgIter.pm @@ -1,30 +1,22 @@ -# Copyright (C) 2016 all contributors +# Copyright (C) 2016-2020 all contributors # License: AGPL-3.0+ -# + +# read-only utilities for Email::MIME package PublicInbox::MsgIter; use strict; use warnings; use base qw(Exporter); -our @EXPORT = qw(msg_iter); -use Email::MIME; -use Scalar::Util qw(readonly); - -# Workaround Email::MIME versions without -# commit dcef9be66c49ae89c7a5027a789bbbac544499ce -# ("removing all trailing newlines was too much") -# This is necessary for Debian jessie -my $bad = 1.923; -my $good = 1.935; -my $ver = $Email::MIME::VERSION; -my $extra_nl = 1 if ($ver >= $bad && $ver < $good); +our @EXPORT = qw(msg_iter msg_part_text); +use PublicInbox::MIME; # Like Email::MIME::walk_parts, but this is: # * non-recursive # * passes depth and indices to the iterator callback -sub msg_iter ($$) { - my ($mime, $cb) = @_; +sub msg_iter ($$;$$) { + my ($mime, $cb, $cb_arg, $do_undef) = @_; my @parts = $mime->subparts; if (@parts) { + $mime = $_[0] = undef if $do_undef; # saves some memory my $i = 0; @parts = map { [ $_, 1, ++$i ] } @parts; while (my $p = shift @parts) { @@ -36,22 +28,47 @@ sub msg_iter ($$) { @sub = map { [ $_, $depth, @idx, ++$i ] } @sub; @parts = (@sub, @parts); } else { - if ($extra_nl) { - my $lf = $part->{mycrlf}; - my $bref = $part->{body}; - if (readonly($$bref)) { - my $s = $$bref . $lf; - $part->{body} = \$s; - } else { - $$bref .= $lf; - } - } - $cb->($p); + $cb->($p, $cb_arg); } } } else { - $cb->([$mime, 0, 0]); + $cb->([$mime, 0, 0], $cb_arg); + } +} + +sub msg_part_text ($$) { + my ($part, $ct) = @_; + + # TODO: we may offer a separate sub for people who need to index + # HTML-only mail, but the majority of HTML mail is multipart/alternative + # with a text part which we don't have to waste cycles decoding + return if $ct =~ m!\btext/x?html\b!; + + my $s = eval { $part->body_str }; + my $err = $@; + + # text/plain is the default, multipart/mixed happened a few + # times when it should not have been: + # <87llgalspt.fsf@free.fr> + # <200308111450.h7BEoOu20077@mail.osdl.org> + if ($err && ($ct =~ m!\btext/\b!i || + $ct =~ m!\bmultipart/mixed\b!i)) { + my $cte = $part->header_raw('Content-Transfer-Encoding'); + if (defined($cte) && $cte =~ /\b7bit\b/i) { + $s = $part->body; + $err = undef if $s =~ /\A[[:ascii:]]+\z/s; + } else { + # Try to assume UTF-8 because Alpine seems to + # do wacky things and set charset=X-UNKNOWN + $part->charset_set('UTF-8'); + $s = eval { $part->body_str }; + } + + # If forcing charset=UTF-8 failed, + # caller will warn further down... + $s = $part->body if $@; } + ($s, $err); } 1;