1 # Copyright (C) 2020-2021 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # *-external commands of lei
5 package PublicInbox::LeiExternal;
8 use PublicInbox::Config;
11 my ($self, $cb, @arg) = @_;
12 my $cfg = $self->_lei_cfg;
14 for my $sec (grep(/\Aexternal\./, @{$cfg->{-section_order}})) {
15 my $loc = substr($sec, length('external.'));
16 $boost{$loc} = $cfg->{"$sec.boost"};
18 return \%boost if !wantarray && !$cb;
20 # highest boost first, but stable for alphabetic tie break
22 my @order = sort { $boost{$b} <=> $boost{$a} } sort keys %boost;
23 if (ref($cb) eq 'CODE') {
24 for my $loc (@order) {
25 $cb->(@arg, $loc, $boost{$loc});
27 } elsif (ref($cb) eq 'HASH') {
30 @order; # scalar or array
33 sub ext_canonicalize {
35 if ($location !~ m!\Ahttps?://!) {
36 PublicInbox::Config::rel2abs_collapsed($location);
39 my $uri = URI->new($location)->canonical;
40 my $path = $uri->path . '/';
41 $path =~ tr!/!/!s; # squeeze redundant '/'
47 # TODO: we will probably extract glob2re into a separate module for
48 # PublicInbox::Filter::Base and maybe other places
49 my %re_map = ( '*' => '[^/]*?', '?' => '[^/]',
50 '[' => '[', ']' => ']', ',' => ',' );
57 my $schema_host_port = '';
59 # don't glob URL-looking things that look like IPv6
60 if ($re =~ s!\A([a-z0-9\+]+://\[[a-f0-9\:]+\](?::[0-9]+)?/)!!i) {
61 $schema_host_port = quotemeta $1; # "http://[::1]:1234"
63 my $changes = ($re =~ s!(.)!
64 $re_map{$p eq '\\' ? '' : do {
65 if ($1 eq '[') { ++$in_bracket }
66 elsif ($1 eq ']') { --$in_bracket }
67 elsif ($1 eq ',') { ++$qm } # no change
71 ($p eq '-' && $in_bracket) ? $p : (++$qm, "\Q$p")
73 # bashism (also supported by curl): {a,b,c} => (a|b|c)
74 $changes += ($re =~ s/([^\\]*)\\\{([^,]*,[^\\]*)\\\}/
75 (my $in_braces = $2) =~ tr!,!|!;
78 ($changes - $qm) ? $schema_host_port.$re : undef;
81 # get canonicalized externals list matching $loc
82 # $is_exclude denotes it's for --exclude
83 # otherwise it's for --only/--include is assumed
85 my ($self, $loc, $is_exclude) = @_;
86 return (ext_canonicalize($loc)) if -e $loc;
88 my @cur = externals_each($self);
89 my $do_glob = !$self->{opt}->{globoff}; # glob by default
90 if ($do_glob && (my $re = glob2re($loc))) {
91 @m = grep(m!$re!, @cur);
92 return @m if scalar(@m);
93 } elsif (index($loc, '/') < 0) { # exact basename match:
94 @m = grep(m!/\Q$loc\E/?\z!, @cur);
95 return @m if scalar(@m) == 1;
96 } elsif ($is_exclude) { # URL, maybe:
97 my $canon = ext_canonicalize($loc);
98 @m = grep(m!\A\Q$canon\E\z!, @cur);
99 return @m if scalar(@m) == 1;
101 return (ext_canonicalize($loc));
103 if (scalar(@m) == 0) {
104 $self->fail("`$loc' is unknown");
106 $self->fail("`$loc' is ambiguous:\n", map { "\t$_\n" } @m);
111 # TODO: does this need JSON output?
112 sub lei_ls_external {
113 my ($self, $filter) = @_;
114 my $opt = $self->{opt};
115 my $do_glob = !$opt->{globoff}; # glob by default
116 my ($OFS, $ORS) = $opt->{z} ? ("\0", "\0\0") : (" ", "\n");
118 my $re = $do_glob ? glob2re($filter) : undef;
119 $re //= index($filter, '/') < 0 ?
120 qr!/\Q$filter\E/?\z! : # exact basename match
121 qr/\Q$filter\E/; # grep -F semantics
122 my @ext = externals_each($self, my $boost = {});
123 @ext = $opt->{'invert-match'} ? grep(!/$re/, @ext)
125 if ($opt->{'local'} && !$opt->{remote}) {
126 @ext = grep(!m!\A[a-z\+]+://!, @ext);
127 } elsif ($opt->{remote} && !$opt->{'local'}) {
128 @ext = grep(m!\A[a-z\+]+://!, @ext);
131 $self->out($loc, $OFS, 'boost=', $boost->{$loc}, $ORS);
135 sub add_external_finish {
136 my ($self, $location) = @_;
137 my $cfg = $self->_lei_cfg(1);
138 my $new_boost = $self->{opt}->{boost} // 0;
139 my $key = "external.$location.boost";
140 my $cur_boost = $cfg->{$key};
141 return if defined($cur_boost) && $cur_boost == $new_boost; # idempotent
142 $self->lei_config($key, $new_boost);
145 sub lei_add_external {
146 my ($self, $location) = @_;
147 my $opt = $self->{opt};
148 my $mirror = $opt->{mirror} // do {
150 for my $sw ($self->index_opt, $self->curl_opt,
151 qw(no-torsocks torsocks inbox-version)) {
152 my ($f) = (split(/|/, $sw, 2))[0];
153 next unless defined $opt->{$f};
154 $f = length($f) == 1 ? "-$f" : "--$f";
157 if (scalar(@fail) == 1) {
158 return $self->("@fail requires --mirror");
160 my $last = pop @fail;
161 my $fail = join(', ', @fail);
162 return $self->("@fail and $last require --mirror");
166 my $new_boost = $opt->{boost} // 0;
167 $location = ext_canonicalize($location);
168 if (defined($mirror) && -d $location) {
169 $self->fail(<<""); # TODO: did you mean "update-external?"
170 --mirror destination `$location' already exists
172 } elsif (-d $location) {
173 index($location, "\n") >= 0 and
174 return $self->fail("`\\n' not allowed in `$location'");
176 if ($location !~ m!\Ahttps?://! && !-d $location) {
177 $mirror // return $self->fail("$location not a directory");
178 index($location, "\n") >= 0 and
179 return $self->fail("`\\n' not allowed in `$location'");
180 $mirror = ext_canonicalize($mirror);
181 require PublicInbox::LeiMirror;
182 PublicInbox::LeiMirror->start($self, $mirror => $location);
184 add_external_finish($self, $location);
188 sub lei_forget_external {
189 my ($self, @locations) = @_;
190 my $cfg = $self->_lei_cfg(1);
191 my $quiet = $self->{opt}->{quiet};
193 for my $loc (@locations) {
194 my (@unset, @not_found);
195 for my $l ($loc, ext_canonicalize($loc)) {
197 my $key = "external.$l.boost";
198 delete($cfg->{$key});
199 $self->_config('--unset', $key);
202 } elsif (($? >> 8) == 5) {
205 $self->err("# --unset $key error");
206 return $self->x_it($?);
211 $self->err("# $_ gone") for @unset;
212 } elsif (@not_found) {
213 $self->err("# $_ not found") for @not_found;
214 } # else { already exited
218 # returns an anonymous sub which returns an array of potential results
219 sub complete_url_prepare {
221 # Workaround bash word-splitting URLs to ['https', ':', '//' ...]
222 # Maybe there's a better way to go about this in
223 # contrib/completion/lei-completion.bash
225 my $cur = pop(@$argv) // '';
228 if ($cur eq ':' && @x) {
232 while (@x > 2 && $x[0] !~ /\A(?:http|nntp|imap)s?\z/i &&
236 if (@x >= 2) { # qw(https : hostname : 443) or qw(http :)
238 } else { # just filter out the flags and hope for the best
239 $re = join('', grep(!/^-/, @$argv));
241 $re = quotemeta($re);
244 # only return the part specified on the CLI
245 # don't duplicate if already 100% completed
246 $_[0] =~ /\A$re(\Q$cur\E.*)/ ? ($cur eq $1 ? () : $1) : ()
248 wantarray ? ($re, $cur, $match_cb) : $match_cb;
251 # shell completion helper called by lei__complete
252 sub _complete_forget_external {
253 my ($self, @argv) = @_;
254 my $cfg = $self->_lei_cfg;
255 my ($cur, $re, $match_cb) = complete_url_prepare(\@argv);
256 # FIXME: bash completion off "http:" or "https:" when the last
257 # character is a colon doesn't work properly even if we're
258 # returning "//$HTTP_HOST/$PATH_INFO/", not sure why, could
261 $match_cb->(substr($_, length('external.')));
262 } grep(/\Aexternal\.$re\Q$cur/, @{$cfg->{-section_order}});
265 sub _complete_add_external { # for bash, this relies on "compopt -o nospace"
266 my ($self, @argv) = @_;
267 my $cfg = $self->_lei_cfg;
268 my $match_cb = complete_url_prepare(\@argv);
271 my $u = URI->new(substr($_, length('external.')));
272 my ($base) = ($u->path =~ m!((?:/?.*)?/)[^/]+/?\z!);
274 $match_cb->($u->as_string);
275 } grep(m!\Aexternal\.https?://!, @{$cfg->{-section_order}});