4 files changed, 74 insertions, 30 deletions
diff --git a/Documentation/dc-dlvr-spam-flow.txt b/Documentation/dc-dlvr-spam-flow.txt
index 81aba766..d151d272 100644
--- a/Documentation/dc-dlvr-spam-flow.txt
+++ b/Documentation/dc-dlvr-spam-flow.txt
@@ -45,4 +45,4 @@ delivery path as well as removing the message from the git tree.
  * spamc / spamd - SpamAssassin: http://spamassassin.apache.org/
  
  * report-spam / dc-dlvr - distributed with public-inbox in the scripts/
-  directory: git clone https://public-inbox.org/ public-inbox
+  directory: git clone https://public-inbox.org/public-inbox.git
diff --git a/Documentation/include.mk b/Documentation/include.mk
index 4af8d455..1460604d 100644
--- a/Documentation/include.mk
+++ b/Documentation/include.mk
@@ -13,7 +13,6 @@ MAN = man
  
  # this is "xml" on FreeBSD and maybe some other distros:
  XMLSTARLET = xmlstarlet
-AWK = awk
  
  # same as pod2text
  COLUMNS = 76
@@ -89,11 +88,9 @@ doc_install :: install-man
  %.1 %.5 %.7 %.8 : Documentation/%.pod
          $(podman) -s $(subst .,,$(suffix $@)) $< $@+ && mv $@+ $@
  
-# n.b. not sure if our usage of man(1) is portable or not, so not
-# enabled by default "check" target:
-# check :: check-man
-check_man = @echo CHECK80 $<;COLS=80 $(MAN) ./$^ | \
-        $(AWK) 'length>80{print;err=1}END{exit(err)}' >&2
+check :: check-man
+check_man = @echo CHECK80 $<;COLUMNS=80 $(MAN) ./$^ | \
+        $(AWK) '{gsub(/\b./,"")}length>80{print;err=1}END{exit(err)}' >&2
  
  %.1.cols : %.1; $(check_man)
  %.5.cols : %.5; $(check_man)
diff --git a/Documentation/public-inbox-mda.pod b/Documentation/public-inbox-mda.pod
index 921b7a15..01bb8476 100644
--- a/Documentation/public-inbox-mda.pod
+++ b/Documentation/public-inbox-mda.pod
@@ -16,6 +16,31 @@ By default, it relies on L<spamc(1)> for filtering mail,
  but may be disabled via
  L<public-inbox-config(5)/publicinboxmda.spamcheck>
  
+=head1 OPTIONS
+
+=over 8
+
+=item --no-precheck
+
+By default, public-inbox-mda does some simple checks before
+invoking L<spamc(1)> since it is intended to receive mail before
+it goes to a mailing list.
+
+However, some users prefer to use public-inbox-mda to mirror
+mailing lists.  This option exists to support those users.
+
+Using this option, the following prechecks are disabled:
+
+  * multiple Message-IDs
+  * non-existent Message-IDs
+  * Message-IDs longer than 244 characters long
+  * From: header shorter than 3 characters
+  * Subject: header shorter than 2 characters
+  * unusable Date: headers
+  * inbox address specified in To: or Cc: header
+
+=back
+
  =head1 ENVIRONMENT
  
  =over 8
diff --git a/Documentation/public-inbox-overview.pod b/Documentation/public-inbox-overview.pod
index 46060d0c..91d4a88e 100644
--- a/Documentation/public-inbox-overview.pod
+++ b/Documentation/public-inbox-overview.pod
@@ -4,7 +4,7 @@ public-inbox-overview - an overview of public-inbox
  
  =head1 DESCRIPTION
  
-public-inbox consists of many parts which may be used
+public-inbox consists of many pieces which may be used
  independently or in conjunction of each other for:
  
  =over 4
@@ -15,11 +15,11 @@ Mirroring existing public-inboxes.
  
  =item 2
  
-Mirroring mailing lists directly.
+Mirroring mailing lists.
  
  =item 3
  
-Hosting standalone.
+Hosting standalone inboxes.
  
  =back
  
@@ -30,35 +30,52 @@ started.  Your mirror will remain dependent on the REMOTE_URL
  you are mirroring and you only need to use two new commands in
  addition to common L<git(1)> commands.
  
-        git clone --mirror REMOTE_URL /path/to/repo.git
+Instructions are different depending on whether the inbox
+is L<public-inbox-v1-format(5)> or L<public-inbox-v2-format(5)>.
+See the "Archives are clonable:" part of the WWW interface
+of a given inbox for cloning instructions specific to
+that inbox.  The instructions are roughly:
  
-        # The following should create the necessary entry in
-        # ~/.public-inbox/config
-        public-inbox-init NAME /path/to/repo.git MY_URL LIST_ADDRESS
+  # for v1 inboxes:
+  git clone --mirror URL INBOX_DIR
  
-        # Optional but strongly recommended for hosting HTTP
-        # (and required for NNTP)
-        # enable search (requires Search::Xapian and DBD::SQLite)
-        public-inbox-index /path/to/repo.git
+  # for v2 inboxes (each epoch needs to be cloned):
+  git clone --mirror URL/EPOCH INBOX_DIR/git/EPOCH.git
  
-        # Periodically update the repo with the following commands
-        # to update the git repo and index new messages:
-        cd /path/to/repo.git && git fetch && public-inbox-index
+  # The following should create the necessary entry in
+  # ~/.public-inbox/config, use "-V2" only for v2 inboxes:
+  public-inbox-init [-V2] NAME INBOX_DIR MY_URL LIST_ADDRESS
  
-See L</"Hosting public-inboxes"> below for info on how to expose
+  # Optional but strongly recommended for hosting HTTP
+  # (and required for NNTP)
+  # enable overview (requires DBD::SQLite) and, if Search::Xapian is
+  # available, search:
+  public-inbox-index INBOX_DIR
+
+  # Periodically fetch the repo using git-fetch(1)
+  # for v1 inboxes:
+  git --git-dir=INBOX_DIR fetch
+
+  # for v2 (in most cases, only the newest epoch needs to be fetched):
+  git --git-dir=INBOX_DIR/git/EPOCH.git fetch
+
+  # index new messages after fetching:
+  public-inbox-index INBOX_DIR
+
+See L</"Serving public-inboxes"> below for info on how to expose
  your mirror to other readers.
  
-=head2 Mirroring mailing lists directly
+=head2 Mirroring mailing lists
  
-Mirroring existing mailing lists may be done by any reader
+Mirroring mailing lists may be done by any reader
  of a mailing list using L<public-inbox-watch(1)>.
  
          # This will create a new git repository:
-        public-inbox-init NAME /path/to/repo.git MY_URL LIST_ADDRESS
+        public-inbox-init -V2 NAME INBOX_DIR MY_URL LIST_ADDRESS
  
  Then, see the L<public-inbox-watch(1)> manual for configuring
-C<watch>, C<watchheader>, and the optional C<spamcheck> and
-C<watchspam> entries.
+C<watch>, C<watchheader>, C<listid> and the optional C<spamcheck>
+and C<watchspam> entries.
  
  You will need to leave L<public-inbox-watch(1)> running to
  keep the mailbox up-to-date as messages are delivered to
@@ -69,15 +86,20 @@ is recommended.  L<public-inbox-watch(1)> will automatically
  maintain the indices if they were created by
  L<public-inbox-index(1)>
  
-        public-inbox-index /path/to/repo.git
+        public-inbox-index INBOX_DIR
+
+Instead of using L<public-inbox-watch(1)>, using
+L<public-inbox-mda(1)> with the C<--no-precheck> option and
+relying on the C<listid> directive in L<public-inbox-config(5)>
+is also an option.
  
-=head2 Hosting standalone
+=head2 Hosting standalone inboxes
  
  Using L<public-inbox-init(1)> to initialize the inbox as in the
  other methods is recommended.  See L<public-inbox-mda(1)> for
  more details; but this also requires MTA-specific knowledge.
  
-=head2 Hosting public-inboxes
+=head2 Serving public-inboxes
  
  Since public-inboxes are git repositories, they may be served to
  remote clients via L<git-daemon(1)> as well as specialized HTTP