Eric Wong <[email protected]> wrote:
> diff --git a/lib/PublicInbox/LeiRemote.pm b/lib/PublicInbox/LeiRemote.pm
> index 54750062..15013baa 100644
> --- a/lib/PublicInbox/LeiRemote.pm
> +++ b/lib/PublicInbox/LeiRemote.pm
> @@ -52,7 +52,7 @@ sub mset {
> $self->{smsg} = [];
> $fh = IO::Uncompress::Gunzip->new($fh, MultiStream => 1);
> PublicInbox::MboxReader->mboxrd($fh, \&_each_mboxrd_eml, $self);
> - my $wait = $self->{lei}->{sto}->wq_do('done');
> + $self->{lei}->sto_done_request;
That's usually not the normal lei/store, but the {tmp_sto} one
from LeiRediff. So it must be synchronous in that case because
we make multiple queries, there.
So I'll revert that hunk.
And ugh, the new lei-store-fail.t test is so nasty; I don't
think a 100ms delay is enough for some systems...
diff --git a/t/lei-store-fail.t b/t/lei-store-fail.t
index e9ad779f..fb0f2b75 100644
--- a/t/lei-store-fail.t
+++ b/t/lei-store-fail.t
@@ -31,7 +31,7 @@ Message-ID: <$_\@t>
will this save?
EOM
}
- tick 0.1; # XXX ugh, this is so hacky
+ tick 0.2; # XXX ugh, this is so hacky
# make sto_done_request fail:
remove_tree("$ENV{HOME}/.local/share/lei/store");