doc: generate manpages for some Xapian commands

They're nowhere to be found on Xapian.org, and links to
external services are either too long (for manpages.debian.org)
or have privacy-invasive tracking JS on them.

3 files changed, 66 insertions, 2 deletions

diff --git a/Documentation/.gitignore b/Documentation/.gitignore
index e78a0d33..4d2ca22f 100644
--- a/Documentation/.gitignore
+++ b/Documentation/.gitignore
@@ -1,2 +1,3 @@
 /public-inbox-*.txt
 /standards.txt
+/.x
diff --git a/Documentation/extman.perl b/Documentation/extman.perl
new file mode 100755
index 00000000..a72caca8
--- /dev/null
+++ b/Documentation/extman.perl
@@ -0,0 +1,33 @@
+#!/usr/bin/perl -w
+# Copyright (C) 2019 all contributors <meta@public-inbox.org>
+# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
+# prints a manpage to stdout
+use strict;
+my $xapmsg = 'See https://xapian.org/ for more information on Xapian';
+my $usage = "$0 /path/to/manpage.SECTION.txt";
+my $manpage = shift or die $usage;
+my $MAN = $ENV{MAN} || 'man';
+my @args;
+$manpage = (split('/', $manpage))[-1];
+$manpage =~ s/\.txt\z//;
+$manpage =~ s/\A\.//; # no leading dot (see Documentation/include.mk)
+$manpage =~ s/\.(\d+.*)\z// and push @args, $1; # section
+push @args, $manpage;
+
+# don't use UTF-8 characters which readers may not have fonts for
+$ENV{LC_ALL} = $ENV{LANG} = 'C';
+$ENV{COLUMNS} = '76'; # same as pod2text default
+$ENV{PAGER} = 'cat';
+my $cmd = join(' ', $MAN, @args);
+system($MAN, @args) and die "$cmd failed: $!\n";
+$manpage =~ /\A(?:copydatabase|xapian-compact)\z/ and
+        print "\n\n", $xapmsg, "\n";
+
+# touch -r $(man -w $section $manpage) output.txt
+if (-f \*STDOUT) {
+        open(my $fh, '-|', $MAN, '-w', @args) or die "$MAN -w broken?: $!\n";
+        chomp(my $path = <$fh>);
+        my @st = stat($path) or die "stat($path) failed: $!\n";
+        # 9 - mtime
+        utime($st[9], $st[9], \*STDOUT) or die "utime(STDOUT) failed: $!\n";
+}
diff --git a/Documentation/include.mk b/Documentation/include.mk
index 5b793fb5..b064f295 100644
--- a/Documentation/include.mk
+++ b/Documentation/include.mk
@@ -8,6 +8,11 @@ all::
 # Maybe it's not worth it to support non-GNU make, though...
 RSYNC = rsync
 RSYNC_DEST = public-inbox.org:/srv/public-inbox/
+MAN = man
+
+# same as pod2text
+COLUMNS = 76
+
 txt := INSTALL README COPYING TODO HACKING
 dtxt := design_notes.txt design_www.txt dc-dlvr-spam-flow.txt hosted.txt
 dtxt += marketing.txt
@@ -106,6 +111,26 @@ html: $(docs_html)
 gz_docs := $(addsuffix .gz, $(docs) $(docs_html))
 rsync_docs := $(gz_docs) $(docs) $(docs_html)
 
+# external manpages which we host ourselves, since some packages
+# (currently just Xapian) doesn't host manpages themselves.
+xtxt :=
+xtxt += .copydatabase.1
+xtxt += .xapian-compact.1
+xtxt := $(addprefix Documentation/.x/, $(addsuffix .txt, $(xtxt)))
+xdocs := $(xtxt)
+xdocs_html := $(addsuffix .html, $(subst .txt,,$(xtxt)))
+gz_xdocs := $(addsuffix .gz, $(xdocs) $(xdocs_html))
+rsync_xdocs := $(gz_xdocs) $(xdocs_html) $(xdocs)
+xdoc: $(xdocs) $(xdocs_html)
+
+Documentation/.x/%.txt::
+        @-mkdir -p $(@D)
+        $(PERL) -w Documentation/extman.perl $@ >$@+
+        mv $@+ $@
+
+Documentation/.x/%.html: Documentation/.x/%.txt
+        $(txt2pre)
+
 doc: $(docs)
 
 %.gz: %
@@ -114,14 +139,19 @@ doc: $(docs)
         mv $@+ $@
 
 gz-doc: $(gz_docs)
+
+gz-xdoc: $(gz_xdocs)
+
 rsync-doc:
         # /usr/share/doc/rsync/scripts/git-set-file-times{.gz} on Debian systems
         # It is also at: https://yhbt.net/git-set-file-times
         -git set-file-times $(docs) $(txt)
-        $(MAKE) gz-doc
-        $(RSYNC) --chmod=Fugo=r -av $(rsync_docs) $(RSYNC_DEST)
+        $(MAKE) gz-doc gz-xdoc
+        $(RSYNC) --chmod=Fugo=r -av $(rsync_docs) $(rsync_xdocs) $(RSYNC_DEST)
+
 clean-doc:
         $(RM) $(man1) $(man5) $(man7) $(gz_docs) $(docs_html) $(mantxt)
+        $(RM) $(gz_xdocs) $(xdocs_html) $(xdocs)
 
 clean :: clean-doc