X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FQspawn.pm;h=53d0ad55ee84d47ccc10bd2d7516136fd38207e8;hb=23af251dd607c4e75ab1e68063f2c885c48cc035;hp=d395a10b3b156b1fac05facbad8a4eca04ef44c2;hpb=97e77af0af2db3ce89b100ad4e1cfef7055870eb;p=public-inbox.git diff --git a/lib/PublicInbox/Qspawn.pm b/lib/PublicInbox/Qspawn.pm index d395a10b..53d0ad55 100644 --- a/lib/PublicInbox/Qspawn.pm +++ b/lib/PublicInbox/Qspawn.pm @@ -1,4 +1,4 @@ -# Copyright (C) 2016-2020 all contributors +# Copyright (C) 2016-2021 all contributors # License: AGPL-3.0+ # Like most Perl modules in public-inbox, this is internal and @@ -12,12 +12,13 @@ # operate in. This can be useful to ensure smaller inboxes can # be cloned while cloning of large inboxes is maxed out. # -# This does not depend on PublicInbox::DS or any other external -# scheduling mechanism, you just need to call start() and finish() -# appropriately. However, public-inbox-httpd (which uses PublicInbox::DS) -# will be able to schedule this based on readability of stdout from -# the spawned process. See GitHTTPBackend.pm and SolverGit.pm for -# usage examples. It does not depend on any form of threading. +# This does not depend on the PublicInbox::DS::event_loop or any +# other external scheduling mechanism, you just need to call +# start() and finish() appropriately. However, public-inbox-httpd +# (which uses PublicInbox::DS) will be able to schedule this +# based on readability of stdout from the spawned process. +# See GitHTTPBackend.pm and SolverGit.pm for usage examples. +# It does not depend on any form of threading. # # This is useful for scheduling CGI execution of both long-lived # git-http-backend(1) process (for "git clone") as well as short-lived @@ -25,8 +26,8 @@ package PublicInbox::Qspawn; use strict; -use warnings; use PublicInbox::Spawn qw(popen_rd); +use PublicInbox::GzipFilter; # n.b.: we get EAGAIN with public-inbox-httpd, and EINTR on other PSGI servers use Errno qw(EAGAIN EINTR); @@ -56,9 +57,9 @@ sub _do_spawn { $self->{cmd} = $o{quiet} ? undef : $cmd; eval { # popen_rd may die on EMFILE, ENFILE - ($self->{rpipe}, $self->{pid}) = popen_rd($cmd, $cmd_env, \%o); + $self->{rpipe} = popen_rd($cmd, $cmd_env, \%o); - die "E: $!" unless defined($self->{pid}); + die "E: $!" unless defined($self->{rpipe}); $limiter->{running}++; $start_cb->($self); # EPOLL_CTL_ADD may ENOSPC/ENOMEM @@ -75,11 +76,6 @@ sub child_err ($) { $msg; } -sub log_err ($$) { - my ($env, $msg) = @_; - $env->{'psgi.errors'}->print($msg, "\n"); -} - sub finalize ($$) { my ($self, $err) = @_; @@ -103,7 +99,7 @@ sub finalize ($$) { $self->{err} = $err; } if ($env && $self->{cmd}) { - log_err($env, join(' ', @{$self->{cmd}}) . ": $err"); + warn join(' ', @{$self->{cmd}}) . ": $err"; } } if ($qx_cb) { @@ -115,41 +111,14 @@ sub finalize ($$) { } } -# callback for dwaitpid -sub waitpid_err ($$) { - my ($self, $pid) = @_; - my $xpid = delete $self->{pid}; - my $err; - if (defined $pid) { - if ($pid > 0) { # success! - $err = child_err($?); - } elsif ($pid < 0) { # ??? does this happen in our case? - $err = "W: waitpid($xpid, 0) => $pid: $!"; - } # else should not be called with pid == 0 - } - finalize($self, $err); -} - -sub do_waitpid ($) { - my ($self) = @_; - my $pid = $self->{pid}; - # PublicInbox::DS may not be loaded - eval { PublicInbox::DS::dwaitpid($pid, \&waitpid_err, $self) }; - # done if we're running in PublicInbox::DS::EventLoop - if ($@) { - # non public-inbox-{httpd,nntpd} callers may block: - my $ret = waitpid($pid, 0); - waitpid_err($self, $ret); - } -} +# callback for dwaitpid or ProcessPipe +sub waitpid_err { finalize($_[0], child_err($?)) } sub finish ($;$) { my ($self, $err) = @_; - if (delete $self->{rpipe}) { - do_waitpid($self); - } else { - finalize($self, $err); - } + my $tied_pp = delete($self->{rpipe}) or return finalize($self, $err); + my PublicInbox::ProcessPipe $pp = tied *$tied_pp; + @$pp{qw(cb arg)} = (\&waitpid_err, $self); # for ->DESTROY } sub start ($$$) { @@ -214,7 +183,7 @@ sub psgi_qx { # PSGI servers. sub event_step { my ($self, $err) = @_; # $err: $! - log_err($self->{psgi_env}, "psgi_{return,qx} $err") if defined($err); + warn "psgi_{return,qx} $err" if defined($err); finish($self); my ($fh, $qx_fh) = delete(@$self{qw(fh qx_fh)}); $fh->close if $fh; # async-only (psgi_return) @@ -223,7 +192,6 @@ sub event_step { sub rd_hdr ($) { my ($self) = @_; # typically used for reading CGI headers - # we must loop until EAGAIN for EPOLLET in HTTPD/Async.pm # We also need to check EINTR for generic PSGI servers. my $ret; my $total_rd = 0; @@ -236,14 +204,14 @@ sub rd_hdr ($) { $total_rd += $r; eval { $ret = $ph_cb->($total_rd, $hdr_buf, $ph_arg) }; if ($@) { - log_err($self->{psgi_env}, "parse_hdr: $@"); + warn "parse_hdr: $@"; $ret = [ 500, [], [ "Internal error\n" ] ]; } } else { # caller should notify us when it's ready: return if $! == EAGAIN; next if $! == EINTR; # immediate retry - log_err($self->{psgi_env}, "error reading header: $!"); + warn "error reading header: $!"; $ret = [ 500, [], [ "Internal error\n" ] ]; } } until (defined $ret); @@ -255,7 +223,9 @@ sub psgi_return_init_cb { my ($self) = @_; my $r = rd_hdr($self) or return; my $env = $self->{psgi_env}; - my $filter = delete $env->{'qspawn.filter'}; + my $filter = delete $env->{'qspawn.filter'} // + PublicInbox::GzipFilter::qsp_maybe($r->[1], $env); + my $wcb = delete $env->{'qspawn.wcb'}; my $async = delete $self->{async}; if (scalar(@$r) == 3) { # error @@ -357,12 +327,12 @@ sub new { } sub setup_rlimit { - my ($self, $name, $config) = @_; + my ($self, $name, $cfg) = @_; foreach my $rlim (@PublicInbox::Spawn::RLIMITS) { my $k = lc($rlim); $k =~ tr/_//d; $k = "publicinboxlimiter.$name.$k"; - defined(my $v = $config->{$k}) or next; + defined(my $v = $cfg->{$k}) or next; my @rlimit = split(/\s*,\s*/, $v); if (scalar(@rlimit) == 1) { push @rlimit, $rlimit[0];