Unnecessary trailings dots in menus

[nncp.git] / doc / integration.texi

@node Integration
@unnumbered Integration with existing software

Here is some examples of how you can solve popular tasks with NNCP,
making them store-and-forward friendly.

@menu
* Index files for freqing: FreqIndex
* Postfix::
* Exim::
* Web feeds: Feeds
* Web pages: WARCs
* BitTorrent and huge files: BitTorrent
* Downloading service: DownloadService
* Git::
* Multimedia streaming: Multimedia
@end menu

@node FreqIndex
@section Index files for freqing

In many cases you do not know exact files list on remote machine you
want to freq from. Because files can be updated there. It is useful to
run cron-ed job on it to create files listing you can freq and search
for files in it:

@example
0  4  *  *  *  cd /storage ; tmp=`mktemp` ; \
    tree -f -h -N --du --timefmt \%Y-\%m-\%d |
    zstdmt -19 > $tmp && chmod 644 $tmp && mv $tmp TREE.txt.zst ; \
    tree -J -f --timefmt \%Y-\%m-\%d |
    zstdmt -19 > $tmp && chmod 644 $tmp && mv $tmp TREE.json.zst
@end example

@node Postfix
@section Integration with Postfix

This section is taken from @url{http://www.postfix.org/UUCP_README.html,
Postfix and UUCP} manual and just replaces UUCP-related calls with NNCP
ones.

@strong{Setting up a Postfix Internet to NNCP gateway}

Here is how to set up a machine that sits on the Internet and that forwards
mail to a LAN that is connected via NNCP.

@itemize

@item You need an @ref{nncp-exec} program that extracts the sender
address from mail that arrives via NNCP, and that feeds the mail into
the Postfix @command{sendmail} command.

@item Define a @command{pipe(8)} based mail delivery transport for
delivery via NNCP:
@example
/usr/local/etc/postfix/master.cf:
nncp      unix  -       n       n       -       -       pipe
  flags=Rqhu user=nncp argv=nncp-exec -quiet $nexthop sendmail $recipient
@end example

This runs the @command{nncp-exec} command to place outgoing mail into
the NNCP queue after replacing @var{$nexthop} by the receiving NNCP
node and after replacing @var{$recipient} by the recipients. The
@command{pipe(8)} delivery agent executes the @command{nncp-exec}
command without assistance from the shell, so there are no problems with
shell meta characters in command-line parameters.

Pay attention to @code{flags}, containing @code{R}, telling Postfix to
include @code{Return-Path:} header. Otherwise that envelope sender
information may be lost. Possibly you will also need somehow to
preserve that header on the receiving side, because @command{sendmail}
command will replace it. For example you can rename it before feeding to
@command{sendmail} with
@code{reformail -R Return-Path: X-Original-Return-Path: | sendmail}, or
extract with:

@verbatiminclude sendmail.sh

Also pay attention that @command{maildrop} does not like @code{From_}
mbox-style header, so you possibly want:

@example
mailbox_command = reformail -f0 | maildrop -d $@{USER@}
@end example

@item Specify that mail for @emph{example.com}, should be delivered via
NNCP, to a host named @emph{nncp-host}:

@example
/usr/local/etc/postfix/transport:
    example.com     nncp:nncp-host
    .example.com    nncp:nncp-host
@end example

See the @command{transport(5)} manual page for more details.

@item Execute the command @command{postmap /etc/postfix/transport}
whenever you change the @file{transport} file.

@item Enable @file{transport} table lookups:

@example
/usr/local/etc/postfix/main.cf:
    transport_maps = hash:$config_directory/transport
@end example

@item Add @emph{example.com} to the list of domains that your site is
willing to relay mail for.

@example
/usr/local/etc/postfix/main.cf:
    relay_domains = example.com ...other relay domains...
@end example

See the @option{relay_domains} configuration parameter description for
details.

@item Execute the command @command{postfix reload} to make the changes
effective.

@end itemize

@strong{Setting up a Postfix LAN to NNCP gateway}

Here is how to relay mail from a LAN via NNCP to the Internet.

@itemize

@item You need an @ref{nncp-exec} program that extracts the sender
address from mail that arrives via NNCP, and that feeds the mail into
the Postfix @command{sendmail} command.

@item Specify that all remote mail must be sent via the @command{nncp}
mail transport to your NNCP gateway host, say, @emph{nncp-gateway}:

@example
/usr/local/etc/postfix/main.cf:
    relayhost = nncp-gateway
    default_transport = nncp
@end example

Postfix 2.0 and later also allows the following more succinct form:

@example
/usr/local/etc/postfix/main.cf:
    default_transport = nncp:nncp-gateway
@end example

@item Define a @command{pipe(8)} based message delivery transport for
mail delivery via NNCP:

@example
/usr/local/etc/postfix/master.cf:
nncp      unix  -       n       n       -       -       pipe
  flags=Fqhu user=nncp argv=nncp-exec -quiet $nexthop sendmail $recipient
@end example

This runs the @command{nncp-exec} command to place outgoing mail into
the NNCP queue. It substitutes the hostname (@emph{nncp-gateway}, or
whatever you specified) and the recipients before execution of the
command. The @command{nncp-exec} command is executed without assistance
from the shell, so there are no problems with shell meta characters.

@item Execute the command @command{postfix reload} to make the changes
effective.

@end itemize

@node Exim
@section Integration with Exim

This section is unaltered copy-paste of
@url{https://changelog.complete.org/archives/10165-asynchronous-email-exim-over-nncp-or-uucp, Asynchronous Email: Exim over NNCP (or UUCP)}
article by John Goerzen, with his permission.

@strong{Sending from Exim to a smarthost}

One common use for async email is from a satellite system: one that
doesn't receive mail, or have local mailboxes, but just needs to get
email out to the Internet. This is a common situation even for
conventionally-connected systems; in Exim speak, this is a "satellite
system that routes mail via a smarthost". That is, every outbound
message goes to a specific target, which then is responsible for
eventual delivery (over the Internet, LAN, whatever).

This is fairly simple in Exim.

We actually have two choices for how to do this: bsmtp or rmail mode.
bsmtp (batch SMTP) is the more modern way, and is essentially a
derivative of SMTP that explicitly can be queued asynchronously.
Basically it's a set of SMTP commands that can be saved in a file. The
alternative is "rmail" (which is just an alias for sendmail these days),
where the data is piped to rmail/sendmail with the recipients given on
the command line. Both can work with Exim and NNCP, but because we're
doing shiny new things, we'll use bsmtp.

These instructions are loosely based on the
@url{https://people.debian.org/~jdg/bsmtp.html, Using outgoing BSMTP with Exim HOWTO}.
Some of these may assume Debianness in the configuration, but should be
easily enough extrapolated to other configs as well.

First, configure Exim to use satellite mode with minimal DNS lookups
(assuming that you may not have working DNS anyhow).

Then, in the Exim primary router section for smarthost
(@file{router/200_exim4-config_primary} in Debian split configurations),
just change @code{transport = remote_smtp_smarthost to transport = nncp}.

Now, define the NNCP transport. If you are on Debian, you might name this
@file{transports/40_exim4-config_local_nncp}:

@example
nncp:
  debug_print = "T: nncp transport for $local_part@@$domain"
  driver = pipe
  user = nncp
  batch_max = 100
  use_bsmtp
  command = /usr/local/nncp/bin/nncp-exec -noprogress -quiet hostname_goes_here rsmtp
.ifdef REMOTE_SMTP_HEADERS_REWRITE
  headers_rewrite = REMOTE_SMTP_HEADERS_REWRITE
.endif
.ifdef REMOTE_SMTP_RETURN_PATH
  return_path = REMOTE_SMTP_RETURN_PATH
.endif
@end example

This is pretty straightforward. We pipe to @command{nncp-exec}, run it
as the nncp user. @command{nncp-exec} sends it to a target node and runs
whatever that node has called @command{rsmtp} (the command to receive
bsmtp data). When the target node processes the request, it will run the
configured command and pipe the data in to it.

@strong{More complicated: Routing to various NNCP nodes}

Perhaps you would like to be able to send mail directly to various NNCP
nodes. There are a lot of ways to do that.

Fundamentally, you will need a setup similar to the UUCP example in
@url{https://www.exim.org/exim-html-current/doc/html/spec_html/ch-the_manualroute_router.html,
Exim's manualroute manual}, which lets you define how to reach various
hosts via UUCP/NNCP. Perhaps you have a star topology (every NNCP node
exchanges email with a central hub). In the NNCP world, you have two
choices of how you do this. You could, at the Exim level, make the
central hub the smarthost for all the side nodes, and let it
redistribute mail. That would work, but requires decrypting messages at
the hub to let Exim process. The other alternative is to configure NNCP
to just send to the destinations via the central hub; that takes
advantage of onion routing and doesn't require any Exim processing at
the central hub at all.

@strong{Receiving mail from NNCP}

On the receiving side, first you need to configure NNCP to authorize the
execution of a mail program. In the section of your receiving host where
you set the permissions for the client, include something like this:

@example
exec: @{
    rsmtp: ["/usr/sbin/sendmail", "-bS"]
@}
@end example

The -bS option is what tells Exim to receive BSMTP on @code{stdin}.

Now, you need to tell Exim that nncp is a trusted user (able to set From
headers arbitrarily). Assuming you are running NNCP as the nncp user,
then add @code{MAIN_TRUSTED_USERS = nncp} to a file such as
@file{/etc/exim4/conf.d/main/01_exim4-config_local-nncp}. That's it!

Some hosts, of course, both send and receive mail via NNCP and will need
configurations for both.

@node Feeds
@section Integration with Web feeds

RSS and Atom feeds could be collected using
@url{https://github.com/wking/rss2email, rss2email} program. It converts
all incoming feed entries to email messages. Read about how to integrate
@ref{Postfix}/@ref{Exim} with email. @command{rss2email} could be run in
a cron, to collect feeds without any user interaction. Also this program
supports ETags and won't pollute the channel if remote server supports
them too.

After installing @command{rss2email}, create configuration file:

@example
$ r2e new rss-robot@@address.com
@end example

and add feeds you want to retrieve:

@example
$ r2e add http://www.git.cypherpunks.ru/?p=nncp.git;a=atom
@end example

and run the process:

@example
$ r2e run
@end example

@node WARCs
@section Integration with Web pages

Simple HTML web page can be downloaded very easily for sending and
viewing it offline after:

@example
$ wget http://www.example.com/page.html
@end example

But most web pages contain links to images, CSS and JavaScript files,
required for complete rendering.
@url{https://www.gnu.org/software/wget/, GNU Wget} supports that
documents parsing and understanding page dependencies. You can download
the whole page with dependencies the following way:

@example
$ wget \
    --page-requisites \
    --convert-links \
    --adjust-extension \
    --restrict-file-names=ascii \
    --span-hosts \
    --random-wait \
    --execute robots=off \
    http://www.example.com/page.html
@end example

that will create @file{www.example.com} directory with all files
necessary to view @file{page.html} web page. You can create single file
compressed tarball with that directory and send it to remote node:

@example
$ tar cf - www.example.com | zstd |
    nncp-file - remote.node:www.example.com-page.tar.zst
@end example

But there are multi-paged articles, there are the whole interesting
sites you want to get in a single package. You can mirror the whole web
site by utilizing @command{wget}'s recursive feature:

@example
$ wget \
    --recursive \
    --timestamping \
    -l inf \
    --no-remove-listing \
    --no-parent \
    [...]
    http://www.example.com/
@end example

There is a standard for creating
@url{https://en.wikipedia.org/wiki/Web_ARChive, Web ARChives}:
@strong{WARC}. Fortunately again, @command{wget} supports it as an
output format.

@example
$ wget \
    --warc-file www.example_com-$(date '+%Y%M%d%H%m%S') \
    --no-warc-compression \
    --no-warc-keep-log \
    [...]
    http://www.example.com/
@end example

That command will create uncompressed @file{www.example_com-XXX.warc}
web archive. By default, WARCs are compressed using
@url{https://en.wikipedia.org/wiki/Gzip, gzip}, but, in example above,
we have disabled it to compress with stronger and faster
@url{https://en.wikipedia.org/wiki/Zstd, zstd}, before sending via
@command{nncp-file}.

There are plenty of software acting like HTTP proxy for your browser,
allowing to view that WARC files. However you can extract files from
that archive using @url{https://pypi.python.org/pypi/Warcat, warcat}
utility, producing usual directory hierarchy:

@example
$ python3 -m warcat extract \
    www.example_com-XXX.warc \
    --output-dir www.example.com-XXX \
    --progress
@end example

@node BitTorrent
@section BitTorrent and huge files

If dealing with @ref{Git}, @ref{Feeds, web feeds} and @ref{Multimedia,
multimedia} goes relatively fast, then BitTorrent and huge files
consumes much time. You can not wait for downloads finish, but want to
queue them after.

@url{http://aria2.github.io/, aria2} multi-protocol download utility
could be used for solving that issue conveniently. It supports HTTP,
HTTPS, FTP, SFTP and BitTorrent protocols, together with
@url{http://tools.ietf.org/html/rfc5854, Metalink} format. BitTorrent
support is fully-featured: UDP trackers, DHT, PEX, encryption, magnet
URIs, Web-seeding, selective downloads, LPD. @command{aria2} can
accelerate HTTP*/*FTP downloads by segmented multiple parallel
connections.

You can queue you files after they are completely downloaded.
@file{aria2-downloaded.sh} contents:

@verbatiminclude aria2-downloaded.sh

Also you can prepare
@url{http://aria2.github.io/manual/en/html/aria2c.html#files, input file}
with the jobs you want to download:

@example
$ cat jobs
http://www.nncpgo.org/download/nncp-0.11.tar.xz
    out=nncp.txz
http://www.nncpgo.org/download/nncp-0.11.tar.xz.sig
    out=nncp.txz.sig
$ aria2c \
    --on-download-complete aria2-downloaded.sh \
    --input-file jobs
@end example

and all that downloaded (@file{nncp.txz}, @file{nncp.txz.sig}) files
will be sent to @file{remote.node} when finished.

@node DownloadService
@section Downloading service

Previous sections tell about manual downloading and sending results to
remote node. But one wish to remotely initiate downloading. That can be
easily solved with @ref{CfgExec, exec} handles.

@verbatim
exec: {
  warcer: ["/bin/sh", "/path/to/warcer.sh"]
  wgeter: ["/bin/sh", "/path/to/wgeter.sh"]
  aria2c: [
    "/usr/local/bin/aria2c",
    "--on-download-complete", "aria2-downloaded.sh",
    "--on-bt-download-complete", "aria2-downloaded.sh"
  ]
}
@end verbatim

@file{warcer.sh} contents:

@verbatiminclude warcer.sh

@file{wgeter.sh} contents:

@verbatiminclude wgeter.sh

Now you can queue that node to send you some website's page, file or
BitTorrents:

@example
$ echo http://www.nncpgo.org/Postfix.html |
    nncp-exec remote.node warcer postfix-whole-page
$ echo http://www.nncpgo.org/Postfix.html |
    nncp-exec remote.node wgeter postfix-html-page
$ echo \
    http://www.nncpgo.org/download/nncp-0.11.tar.xz
    http://www.nncpgo.org/download/nncp-0.11.tar.xz.sig |
    nncp-exec remote.node aria2c
@end example

@node Git
@section Integration with Git

@url{https://git-scm.com/, Git} version control system already has all
necessary tools for store-and-forward networking.
@url{https://git-scm.com/docs/git-bundle, git-bundle} command is
everything you need.

Use it to create bundles containing all required blobs/trees/commits and tags:

@example
$ git bundle create repo-initial.bundle master --tags --branches
$ git tag -f last-bundle
$ nncp-file repo-initial.bundle remote.node:repo-$(date % '+%Y%M%d%H%m%S').bundle
@end example

Do usual working with the Git: commit, add, branch, checkout, etc. When
you decide to queue your changes for sending, create diff-ed bundle and
transfer them:

@example
$ git bundle create repo-$(date '+%Y%M%d%H%m%S').bundle last-bundle..master
or maybe
$ git bundle create repo-$(date '+%Y%M%d').bundle --since=10.days master
@end example

Received bundle on remote machine acts like usual remote:

@example
$ git clone -b master repo-XXX.bundle
@end example

overwrite @file{repo.bundle} file with newer bundles you retrieve and
fetch all required branches and commits:

@example
$ git pull # assuming that origin remote points to repo.bundle
$ git fetch repo.bundle master:localRef
$ git ls-remote repo.bundle
@end example

Bundles are also useful when cloning huge repositories (like Linux has).
Git's native protocol does not support any kind of interrupted download
resuming, so you will start from the beginning if connection is lost.
Bundles, being an ordinary files, can be downloaded with native
HTTP/FTP/NNCP resuming capabilities. After you fetch repository via the
bundle, you can add an ordinary @file{git://} remote and fetch the
difference.

Also you can find the following exec-handler useful:

@verbatiminclude git-bundler.sh

And it allows you to request for bundles like that:
@code{echo some-old-commit..master | nncp-exec REMOTE bundler REPONAME}.

@node Multimedia
@section Integration with multimedia streaming

Many video and audio streams could be downloaded using
@url{http://yt-dl.org/, youtube-dl} program.
@url{https://rg3.github.io/youtube-dl/supportedsites.html, Look} how
many of them are supported, including @emph{Dailymotion}, @emph{Vimeo}
and @emph{YouTube}.

When you multimedia becomes an ordinary file, you can transfer it easily.

@example
$ youtube-dl \
    --exec 'nncp-file @{@} remote.node:' \
    'https://www.youtube.com/watch?list=PLd2Cw8x5CytxPAEBwzilrhQUHt_UN10FJ'
@end example