X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FLeiInput.pm;h=8708356491965cbb1214c76d2d2249408b927430;hb=cf0c7ce3ce81b5c33cb149a7edc3efd99631a57d;hp=2a4968d416e6b1f00dcc7e957e1b749263293cc4;hpb=1bc8c2c264b47079d424ac766095efee3bbc5da0;p=public-inbox.git
diff --git a/lib/PublicInbox/LeiInput.pm b/lib/PublicInbox/LeiInput.pm
index 2a4968d4..87083564 100644
--- a/lib/PublicInbox/LeiInput.pm
+++ b/lib/PublicInbox/LeiInput.pm
@@ -1,10 +1,44 @@
# Copyright (C) 2021 all contributors
# License: AGPL-3.0+
-# parent class for LeiImport, LeiConvert
+# parent class for LeiImport, LeiConvert, LeiIndex
package PublicInbox::LeiInput;
use strict;
use v5.10.1;
+use PublicInbox::DS;
+use PublicInbox::Spawn qw(which popen_rd);
+
+# JMAP RFC 8621 4.1.1
+# https://www.iana.org/assignments/imap-jmap-keywords/imap-jmap-keywords.xhtml
+our @KW = (qw(seen answered flagged draft), # widely-compatible
+ qw(forwarded), # IMAP + Maildir
+ qw(phishing junk notjunk)); # rarely supported
+
+# note: RFC 8621 states "Users may add arbitrary keywords to an Email",
+# but is it good idea? Stick to the system and reserved ones, for now.
+# The widely-compatible ones map to IMAP system flags, Maildir flags
+# and mbox Status/X-Status headers.
+my %KW = map { $_ => 1 } @KW;
+my $L_MAX = 244; # Xapian term limit - length('L')
+
+# RFC 8621, sec 2 (Mailboxes) a "label" for us is a JMAP Mailbox "name"
+# "Servers MAY reject names that violate server policy"
+my %ERR = (
+ L => sub {
+ my ($label) = @_;
+ length($label) >= $L_MAX and
+ return "`$label' too long (must be <= $L_MAX)";
+ $label =~ m{\A[a-z0-9_](?:[a-z0-9_\-\./\@,]*[a-z0-9])?\z}i ?
+ undef : "`$label' is invalid";
+ },
+ kw => sub {
+ my ($kw) = @_;
+ $KW{$kw} ? undef : <fail("--$opt_key unset for $err");
}
+ return 1 if $fmt eq 'eml';
require PublicInbox::MboxLock if $files;
require PublicInbox::MboxReader;
- return 1 if $fmt eq 'eml';
- # XXX: should this handle {gz,bz2,xz}? that's currently in LeiToMail
PublicInbox::MboxReader->reads($fmt) or
return $lei->fail("--$opt_key=$fmt unrecognized");
1;
}
# import a single file handle of $name
-# Subclass must define ->eml_cb and ->mbox_cb
+# Subclass must define ->input_eml_cb and ->input_mbox_cb
sub input_fh {
my ($self, $ifmt, $fh, $name, @args) = @_;
if ($ifmt eq 'eml') {
@@ -36,26 +69,152 @@ error reading $name: $!
# but no Content-Length or "From " escaping.
# "git format-patch" also generates such files by default.
$buf =~ s/\A[\r\n]*From [^\r\n]*\r?\n//s;
- $self->eml_cb(PublicInbox::Eml->new(\$buf), @args);
+
+ # a user may feed just a body: git diff | lei rediff -U9
+ if ($self->{-force_eml}) {
+ my $eml = PublicInbox::Eml->new($buf);
+ substr($buf, 0, 0) = "\n\n" if !$eml->{bdy};
+ }
+ $self->input_eml_cb(PublicInbox::Eml->new(\$buf), @args);
} else {
# prepare_inputs already validated $ifmt
my $cb = PublicInbox::MboxReader->reads($ifmt) //
die "BUG: bad fmt=$ifmt";
- $cb->(undef, $fh, $self->can('mbox_cb'), $self, @args);
+ $cb->(undef, $fh, $self->can('input_mbox_cb'), $self, @args);
+ }
+}
+
+# handles mboxrd endpoints described in Documentation/design_notes.txt
+sub handle_http_input ($$@) {
+ my ($self, $url, @args) = @_;
+ my $lei = $self->{lei} or die 'BUG: {lei} missing';
+ my $curl_opt = delete $self->{"-curl-$url"} or
+ die("BUG: $url curl options not prepared");
+ my $uri = pop @$curl_opt;
+ my $curl = PublicInbox::LeiCurl->new($lei, $self->{curl}) or return;
+ push @$curl, '-s', @$curl_opt;
+ my $cmd = $curl->for_uri($lei, $uri);
+ $lei->qerr("# $cmd");
+ my $rdr = { 2 => $lei->{2}, pgid => 0 };
+ my ($fh, $pid) = popen_rd($cmd, undef, $rdr);
+ grep(/\A--compressed\z/, @$curl) or
+ $fh = IO::Uncompress::Gunzip->new($fh, MultiStream => 1);
+ eval { $self->input_fh('mboxrd', $fh, $url, @args) };
+ my $err = $@;
+ waitpid($pid, 0);
+ $? || $err and
+ $lei->child_error($? || 1, "@$cmd failed".$err ? " $err" : '');
+}
+
+sub input_path_url {
+ my ($self, $input, @args) = @_;
+ my $lei = $self->{lei};
+ my $ifmt = lc($lei->{opt}->{'in-format'} // '');
+ # TODO auto-detect?
+ if ($input =~ m!\Aimaps?://!i) {
+ $lei->{net}->imap_each($input, $self->can('input_net_cb'),
+ $self, @args);
+ return;
+ } elsif ($input =~ m!\A(?:nntps?|s?news)://!i) {
+ $lei->{net}->nntp_each($input, $self->can('input_net_cb'),
+ $self, @args);
+ return;
+ } elsif ($input =~ m!\Ahttps?://!i) {
+ handle_http_input($self, $input, @args);
+ return;
+ }
+ if ($input =~ s!\A([a-z0-9]+):!!i) {
+ $ifmt = lc($1);
+ } elsif ($input =~ /\.(?:patch|eml)\z/i) {
+ $ifmt = 'eml';
+ }
+ my $devfd = $lei->path_to_fd($input) // return;
+ if ($devfd >= 0) {
+ $self->input_fh($ifmt, $lei->{$devfd}, $input, @args);
+ } elsif (-f $input && $ifmt eq 'eml') {
+ open my $fh, '<', $input or
+ return $lei->fail("open($input): $!");
+ $self->input_fh($ifmt, $fh, $input, @args);
+ } elsif (-f _) {
+ my $m = $lei->{opt}->{'lock'} //
+ PublicInbox::MboxLock->defaults;
+ my $mbl = PublicInbox::MboxLock->acq($input, 0, $m);
+ my $zsfx = PublicInbox::MboxReader::zsfx($input);
+ if ($zsfx) {
+ my $in = delete $mbl->{fh};
+ $mbl->{fh} =
+ PublicInbox::MboxReader::zsfxcat($in, $zsfx, $lei);
+ }
+ local $PublicInbox::DS::in_loop = 0 if $zsfx; # dwaitpid
+ $self->input_fh($ifmt, $mbl->{fh}, $input, @args);
+ } elsif (-d _ && (-d "$input/cur" || -d "$input/new")) {
+ return $lei->fail(<new->maildir_each_eml($input,
+ $self->can('input_maildir_cb'),
+ $self, @args);
+ } else {
+ $lei->fail("$input unsupported (TODO)");
+ }
+}
+
+sub bad_http ($$;$) {
+ my ($lei, $url, $alt) = @_;
+ my $x = $alt ? "did you mean <$alt>?" : 'download and import manually';
+ $lei->fail("E: <$url> not recognized, $x");
+}
+
+sub prepare_http_input ($$$) {
+ my ($self, $lei, $url) = @_;
+ require URI;
+ require PublicInbox::MboxReader;
+ require PublicInbox::LeiCurl;
+ require IO::Uncompress::Gunzip;
+ $self->{curl} //= which('curl') or
+ return $lei->fail("curl missing for <$url>");
+ my $uri = URI->new($url);
+ my $path = $uri->path;
+ my %qf = $uri->query_form;
+ my @curl_opt;
+ if ($path =~ m!/(?:t\.mbox\.gz|all\.mbox\.gz)\z!) {
+ # OK
+ } elsif ($path =~ m!/raw\z!) {
+ push @curl_opt, '--compressed';
+ # convert search query to mboxrd request since they require POST
+ # this is only intended for PublicInbox::WWW, and will false-positive
+ # on many other search engines... oh well
+ } elsif (defined $qf{'q'}) {
+ $qf{x} = 'm';
+ $uri->query_form(\%qf);
+ push @curl_opt, '-d', '';
+ $$uri ne $url and $lei->qerr(<<"");
+# <$url> rewritten to <$$uri> with HTTP POST
+
+ # try to provide hints for /$INBOX/$MSGID/T/ and /$INBOX/
+ } elsif ($path =~ s!/[tT]/\z!/t.mbox.gz! ||
+ $path =~ s!/t\.atom\z!/t.mbox.gz! ||
+ $path =~ s!/([^/]+\@[^/]+)/\z!/$1/raw!) {
+ $uri->path($path);
+ return bad_http($lei, $url, $$uri);
+ } else {
+ return bad_http($lei, $url);
}
+ $self->{"-curl-$url"} = [ @curl_opt, $uri ]; # for handle_http_input
}
sub prepare_inputs { # returns undef on error
my ($self, $lei, $inputs) = @_;
my $in_fmt = $lei->{opt}->{'in-format'};
+ my $sync = $lei->{opt}->{'mail-sync'} ? {} : undef; # using LeiMailSync
if ($lei->{opt}->{stdin}) {
@$inputs and return
$lei->fail("--stdin and @$inputs do not mix");
check_input_format($lei) or return;
- $self->{0} = $lei->{0};
+ push @$inputs, '/dev/stdin';
+ push @{$sync->{no}}, '/dev/stdin' if $sync;
}
my $net = $lei->{net}; # NetWriter may be created by l2m
- my $fmt = $lei->{opt}->{'in-format'};
my (@f, @d);
# e.g. Maildir:/home/user/Mail/ or imaps://example.com/INBOX
for my $input (@$inputs) {
@@ -64,6 +223,11 @@ sub prepare_inputs { # returns undef on error
require PublicInbox::NetReader;
$net //= PublicInbox::NetReader->new;
$net->add_url($input);
+ push @{$sync->{ok}}, $input if $sync;
+ } elsif ($input_path =~ m!\Ahttps?://!i) { # mboxrd.gz
+ # TODO: how would we detect r/w JMAP?
+ push @{$sync->{no}}, $input if $sync;
+ prepare_http_input($self, $lei, $input_path) or return;
} elsif ($input_path =~ s/\A([a-z0-9]+)://is) {
my $ifmt = lc $1;
if (($in_fmt // $ifmt) ne $ifmt) {
@@ -71,28 +235,62 @@ sub prepare_inputs { # returns undef on error
--in-format=$in_fmt and `$ifmt:' conflict
}
- if (-f $input_path) {
+ if ($ifmt =~ /\A(?:maildir|mh)\z/i) {
+ push @{$sync->{ok}}, $input if $sync;
+ } else {
+ push @{$sync->{no}}, $input if $sync;
+ }
+ my $devfd = $lei->path_to_fd($input_path) // return;
+ if ($devfd >= 0 || (-f $input_path || -p _)) {
require PublicInbox::MboxLock;
require PublicInbox::MboxReader;
PublicInbox::MboxReader->reads($ifmt) or return
$lei->fail("$ifmt not supported");
- } elsif (-d _) {
+ } elsif (-d $input_path) {
require PublicInbox::MdirReader;
$ifmt eq 'maildir' or return
$lei->fail("$ifmt not supported");
+ $input = $lei->abs_path($input) if $sync;
} else {
return $lei->fail("Unable to handle $input");
}
- } elsif (-f $input) { push @f, $input }
- elsif (-d _) { push @d, $input }
- else { return $lei->fail("Unable to handle $input") }
+ } elsif ($input =~ /\.(?:eml|patch)\z/i && -f $input) {
+ lc($in_fmt//'eml') eq 'eml' or return $lei->fail(<<"");
+$input is `eml', not --in-format=$in_fmt
+
+ require PublicInbox::Eml;
+ push @{$sync->{no}}, $input if $sync;
+ } else {
+ my $devfd = $lei->path_to_fd($input) // return;
+ if ($devfd >= 0 || -f $input || -p _) {
+ push @{$sync->{no}}, $input if $sync;
+ push @f, $input;
+ } elsif (-d $input) {
+ if ($sync) {
+ $input = $lei->abs_path($input);
+ push @{$sync->{ok}}, $input;
+ }
+ push @d, $input;
+ } else {
+ return $lei->fail("Unable to handle $input")
+ }
+ }
}
if (@f) { check_input_format($lei, \@f) or return }
if (@d) { # TODO: check for MH vs Maildir, here
require PublicInbox::MdirReader;
}
+ if ($sync && $sync->{no}) {
+ return $lei->fail(<<"") if !$sync->{ok};
+--mail-sync specified but no inputs support it
+
+ # non-fatal if some inputs support support sync
+ $lei->err("# --mail-sync will only be used for @{$sync->{ok}}");
+ $lei->err("# --mail-sync is not supported for: @{$sync->{no}}");
+ }
if ($net) {
- if (my $err = $net->errors) {
+ $net->{-can_die} = 1;
+ if (my $err = $net->errors($lei)) {
return $lei->fail($err);
}
$net->{quiet} = $lei->{opt}->{quiet};
@@ -103,6 +301,21 @@ sub prepare_inputs { # returns undef on error
$self->{inputs} = $inputs;
}
+sub process_inputs {
+ my ($self) = @_;
+ my $err;
+ for my $input (@{$self->{inputs}}) {
+ eval { $self->input_path_url($input) };
+ next unless $@;
+ $err = "$input: $@";
+ last;
+ }
+ # always commit first, even on error partial work is acceptable for
+ # lei
+ my $wait = $self->{lei}->{sto}->ipc_do('done') if $self->{lei}->{sto};
+ $self->{lei}->fail($err) if $err;
+}
+
sub input_only_atfork_child {
my ($self) = @_;
my $lei = $self->{lei};
@@ -112,4 +325,33 @@ sub input_only_atfork_child {
undef;
}
+# alias this as "net_merge_all_done" to use as an LeiAuth callback
+sub input_only_net_merge_all_done {
+ my ($self) = @_;
+ $self->wq_io_do('process_inputs');
+ $self->wq_close(1);
+}
+
+# like Getopt::Long, but for +kw:FOO and -kw:FOO to prepare
+# for update_xvmd -> update_vmd
+sub vmd_mod_extract {
+ my $argv = $_[-1];
+ my $vmd_mod = {};
+ my @new_argv;
+ for my $x (@$argv) {
+ if ($x =~ /\A(\+|\-)(kw|L):(.+)\z/) {
+ my ($op, $pfx, $val) = ($1, $2, $3);
+ if (my $err = $ERR{$pfx}->($val)) {
+ push @{$vmd_mod->{err}}, $err;
+ } else { # set "+kw", "+L", "-L", "-kw"
+ push @{$vmd_mod->{$op.$pfx}}, $val;
+ }
+ } else {
+ push @new_argv, $x;
+ }
+ }
+ @$argv = @new_argv;
+ $vmd_mod;
+}
+
1;