X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FInbox.pm;h=b76d4e5abb017383f831af4a97c73903e7cfa80a;hb=227a1d886672767e37cc86a3432952c14eb8a143;hp=286555f65cb3b4c8f4a7820b0489e4f6f6005e72;hpb=7a3946ef122e8218c6ce3355d7f968562212d53b;p=public-inbox.git
diff --git a/lib/PublicInbox/Inbox.pm b/lib/PublicInbox/Inbox.pm
index 286555f6..b76d4e5a 100644
--- a/lib/PublicInbox/Inbox.pm
+++ b/lib/PublicInbox/Inbox.pm
@@ -1,4 +1,4 @@
-# Copyright (C) 2016-2018 all contributors
+# Copyright (C) 2016-2019 all contributors
# License: AGPL-3.0+
#
# Represents a public-inbox (which may have multiple mailing addresses)
@@ -7,47 +7,65 @@ use strict;
use warnings;
use PublicInbox::Git;
use PublicInbox::MID qw(mid2path);
-use Devel::Peek qw(SvREFCNT);
use PublicInbox::MIME;
-use POSIX qw(strftime);
+# Long-running "git-cat-file --batch" processes won't notice
+# unlinked packs, so we need to restart those processes occasionally.
+# Xapian and SQLite file handles are mostly stable, but sometimes an
+# admin will attempt to replace them atomically after compact/vacuum
+# and we need to be prepared for that.
my $cleanup_timer;
-eval {
- $cleanup_timer = 'disabled';
- require PublicInbox::EvCleanup;
- $cleanup_timer = undef; # OK if we get here
-};
-my $cleanup_broken = $@;
-
+my $cleanup_avail = -1; # 0, or 1
+my $have_devel_peek;
my $CLEANUP = {}; # string(inbox) -> inbox
sub cleanup_task () {
$cleanup_timer = undef;
my $next = {};
for my $ibx (values %$CLEANUP) {
my $again;
- foreach my $f (qw(mm search)) {
- delete $ibx->{$f} if SvREFCNT($ibx->{$f}) == 1;
+ if ($have_devel_peek) {
+ foreach my $f (qw(mm search over)) {
+ # we bump refcnt by assigning tmp, here:
+ my $tmp = $ibx->{$f} or next;
+ next if Devel::Peek::SvREFCNT($tmp) > 2;
+ delete $ibx->{$f};
+ # refcnt is zero when tmp is out-of-scope
+ }
}
- my $expire = time - 60;
if (my $git = $ibx->{git}) {
- $again = $git->cleanup($expire);
+ $again = $git->cleanup;
}
if (my $gits = $ibx->{-repo_objs}) {
foreach my $git (@$gits) {
- $again = 1 if $git->cleanup($expire);
+ $again = 1 if $git->cleanup;
}
}
- $again ||= !!($ibx->{mm} || $ibx->{search});
+ if ($have_devel_peek) {
+ $again ||= !!($ibx->{over} || $ibx->{mm} ||
+ $ibx->{search});
+ }
$next->{"$ibx"} = $ibx if $again;
}
$CLEANUP = $next;
}
+sub cleanup_possible () {
+ # no need to require DS, here, if it were enabled another
+ # module would've require'd it, already
+ eval { PublicInbox::DS::in_loop() } or return 0;
+
+ eval {
+ require Devel::Peek; # needs separate package in Fedora
+ $have_devel_peek = 1;
+ };
+ 1;
+}
+
sub _cleanup_later ($) {
my ($self) = @_;
- return if $cleanup_broken;
- return unless PublicInbox::EvCleanup::enabled();
- $cleanup_timer ||= PublicInbox::EvCleanup::later(*cleanup_task);
+ $cleanup_avail = cleanup_possible() if $cleanup_avail < 0;
+ return if $cleanup_avail != 1;
+ $cleanup_timer ||= PublicInbox::DS::later(*cleanup_task);
$CLEANUP->{"$self"} = $self;
}
@@ -56,7 +74,7 @@ sub _set_uint ($$$) {
my $val = $opts->{$field};
if (defined $val) {
$val = $val->[-1] if ref($val) eq 'ARRAY';
- $val = undef if $val !~ /\A\d+\z/;
+ $val = undef if $val !~ /\A[0-9]+\z/;
}
$opts->{$field} = $val || $default;
}
@@ -69,7 +87,7 @@ sub _set_limiter ($$$) {
my $mkey = $pfx.'max';
my $val = $self->{$mkey} or return;
my $lim;
- if ($val =~ /\A\d+\z/) {
+ if ($val =~ /\A[0-9]+\z/) {
require PublicInbox::Qspawn;
$lim = PublicInbox::Qspawn::Limiter->new($val);
} elsif ($val =~ /\A[a-z][a-z0-9]*\z/) {
@@ -91,7 +109,7 @@ sub new {
_set_limiter($opts, $pi_config, 'httpbackend');
_set_uint($opts, 'feedmax', 25);
$opts->{nntpserver} ||= $pi_config->{'publicinbox.nntpserver'};
- my $dir = $opts->{mainrepo};
+ my $dir = $opts->{inboxdir};
if (defined $dir && -f "$dir/inbox.lock") {
$opts->{version} = 2;
}
@@ -107,11 +125,13 @@ sub new {
bless $opts, $class;
}
-sub git_part {
- my ($self, $part) = @_;
- ($self->{version} || 1) == 2 or return;
- $self->{"$part.git"} ||= eval {
- my $git_dir = "$self->{mainrepo}/git/$part.git";
+sub version { $_[0]->{version} // 1 }
+
+sub git_epoch {
+ my ($self, $epoch) = @_;
+ $self->version == 2 or return;
+ $self->{"$epoch.git"} ||= eval {
+ my $git_dir = "$self->{inboxdir}/git/$epoch.git";
my $g = PublicInbox::Git->new($git_dir);
$g->{-httpbackend_limiter} = $self->{-httpbackend_limiter};
# no cleanup needed, we never cat-file off this, only clone
@@ -122,8 +142,8 @@ sub git_part {
sub git {
my ($self) = @_;
$self->{git} ||= eval {
- my $git_dir = $self->{mainrepo};
- $git_dir .= '/all.git' if (($self->{version} || 1) == 2);
+ my $git_dir = $self->{inboxdir};
+ $git_dir .= '/all.git' if $self->version == 2;
my $g = PublicInbox::Git->new($git_dir);
$g->{-httpbackend_limiter} = $self->{-httpbackend_limiter};
_cleanup_later($self);
@@ -131,27 +151,26 @@ sub git {
};
}
-sub max_git_part {
+sub max_git_epoch {
my ($self) = @_;
- my $v = $self->{version};
- return unless defined($v) && $v == 2;
- my $part = $self->{-max_git_part};
+ return if $self->version < 2;
+ my $cur = $self->{-max_git_epoch};
my $changed = git($self)->alternates_changed;
- if (!defined($part) || $changed) {
+ if (!defined($cur) || $changed) {
$self->git->cleanup if $changed;
- my $gits = "$self->{mainrepo}/git";
+ my $gits = "$self->{inboxdir}/git";
if (opendir my $dh, $gits) {
my $max = -1;
while (defined(my $git_dir = readdir($dh))) {
- $git_dir =~ m!\A(\d+)\.git\z! or next;
+ $git_dir =~ m!\A([0-9]+)\.git\z! or next;
$max = $1 if $1 > $max;
}
- $part = $self->{-max_git_part} = $max if $max >= 0;
+ $cur = $self->{-max_git_epoch} = $max if $max >= 0;
} else {
warn "opendir $gits failed: $!\n";
}
}
- $part;
+ $cur;
}
sub mm {
@@ -159,8 +178,8 @@ sub mm {
$self->{mm} ||= eval {
require PublicInbox::Msgmap;
_cleanup_later($self);
- my $dir = $self->{mainrepo};
- if (($self->{version} || 1) >= 2) {
+ my $dir = $self->{inboxdir};
+ if ($self->version >= 2) {
PublicInbox::Msgmap->new_file("$dir/msgmap.sqlite3");
} else {
PublicInbox::Msgmap->new($dir);
@@ -168,12 +187,24 @@ sub mm {
};
}
-sub search {
- my ($self) = @_;
- $self->{search} ||= eval {
+sub search ($;$) {
+ my ($self, $over_only) = @_;
+ my $srch = $self->{search} ||= eval {
_cleanup_later($self);
- PublicInbox::Search->new($self, $self->{altid});
+ require PublicInbox::Search;
+ PublicInbox::Search->new($self);
};
+ ($over_only || eval { $srch->xdb }) ? $srch : undef;
+}
+
+sub over ($) {
+ my ($self) = @_;
+ my $srch = search($self, 1) or return;
+ $self->{over} ||= eval {
+ my $over = $srch->{over_ro};
+ $over->dbh_new; # may fail
+ $over;
+ }
}
sub try_cat {
@@ -190,11 +221,11 @@ sub description {
my ($self) = @_;
my $desc = $self->{description};
return $desc if defined $desc;
- $desc = try_cat("$self->{mainrepo}/description");
+ $desc = try_cat("$self->{inboxdir}/description");
local $/ = "\n";
chomp $desc;
$desc =~ s/\s+/ /smg;
- $desc = '($REPO_DIR/description missing)' if $desc eq '';
+ $desc = '($INBOX_DIR/description missing)' if $desc eq '';
$self->{description} = $desc;
}
@@ -202,7 +233,7 @@ sub cloneurl {
my ($self) = @_;
my $url = $self->{cloneurl};
return $url if $url;
- $url = try_cat("$self->{mainrepo}/cloneurl");
+ $url = try_cat("$self->{inboxdir}/cloneurl");
my @url = split(/\s+/s, $url);
local $/ = "\n";
chomp @url;
@@ -210,26 +241,22 @@ sub cloneurl {
}
sub base_url {
- my ($self, $env) = @_;
- my $scheme;
- if ($env && ($scheme = $env->{'psgi.url_scheme'})) { # PSGI env
- my $host_port = $env->{HTTP_HOST} ||
- "$env->{SERVER_NAME}:$env->{SERVER_PORT}";
- my $url = "$scheme://$host_port". ($env->{SCRIPT_NAME} || '/');
+ my ($self, $env) = @_; # env - PSGI env
+ if ($env) {
+ my $url = PublicInbox::Git::host_prefix_url($env, '');
# for mount in Plack::Builder
$url .= '/' if $url !~ m!/\z!;
- $url .= $self->{name} . '/';
- } else {
- # either called from a non-PSGI environment (e.g. NNTP/POP3)
- $self->{-base_url} ||= do {
- my $url = $self->{url} or return undef;
- # expand protocol-relative URLs to HTTPS if we're
- # not inside a web server
- $url = "https:$url" if $url =~ m!\A//!;
- $url .= '/' if $url !~ m!/\z!;
- $url;
- };
+ return $url .= $self->{name} . '/';
}
+ # called from a non-PSGI environment (e.g. NNTP/POP3):
+ $self->{-base_url} ||= do {
+ my $url = $self->{url}->[0] or return undef;
+ # expand protocol-relative URLs to HTTPS if we're
+ # not inside a web server
+ $url = "https:$url" if $url =~ m!\A//!;
+ $url .= '/' if $url !~ m!/\z!;
+ $url;
+ };
}
sub nntp_url {
@@ -267,12 +294,11 @@ sub nntp_url {
# nntp://news.example.com/alt.example
push @m, $u;
}
- my %seen = map { $_ => 1 } @urls;
- foreach my $u (@m) {
- next if $seen{$u};
- $seen{$u} = 1;
- push @urls, $u;
- }
+
+ # List::Util::uniq requires Perl 5.26+, maybe we
+ # can use it by 2030 or so
+ my %seen;
+ @urls = grep { !$seen{$_}++ } (@urls, @m);
}
\@urls;
};
@@ -280,8 +306,8 @@ sub nntp_url {
sub nntp_usable {
my ($self) = @_;
- my $ret = $self->mm && $self->search;
- $self->{mm} = $self->{search} = undef;
+ my $ret = mm($self) && over($self);
+ $self->{mm} = $self->{over} = $self->{search} = undef;
$ret;
}
@@ -322,30 +348,32 @@ sub mid2num($$) {
sub smsg_by_mid ($$) {
my ($self, $mid) = @_;
- my $srch = search($self) or return;
+ my $over = over($self) or return;
# favor the Message-ID we used for the NNTP article number:
defined(my $num = mid2num($self, $mid)) or return;
- my $smsg = $srch->lookup_article($num) or return;
+ my $smsg = $over->get_art($num) or return;
PublicInbox::SearchMsg::psgi_cull($smsg);
}
sub msg_by_mid ($$;$) {
my ($self, $mid, $ref) = @_;
- my $srch = search($self) or
+
+ over($self) or
return msg_by_path($self, mid2path($mid), $ref);
+
my $smsg = smsg_by_mid($self, $mid);
$smsg ? msg_by_smsg($self, $smsg, $ref) : undef;
}
sub recent {
my ($self, $opts, $after, $before) = @_;
- search($self)->{over_ro}->recent($opts, $after, $before);
+ over($self)->recent($opts, $after, $before);
}
sub modified {
my ($self) = @_;
- if (my $srch = search($self)) {
- my $msgs = $srch->{over_ro}->recent({limit => 1});
+ if (my $over = over($self)) {
+ my $msgs = $over->recent({limit => 1});
if (my $smsg = $msgs->[0]) {
return $smsg->{ts};
}