[ANN] ruby-core archive mirror [was: mirroring ruby-core archives off blade.nagaokaut.ac.jp]

[ANN] ruby-core archive mirror [was: mirroring ruby-core archives off blade.nagaokaut.ac.jp]

[Eric Wong]

Eric Wong <normalperson@yhbt.net> wrote:
> I will also need to make some tweaks in the public-inbox code to
> add an option to change the default reply-to-all behavior of the
> archives; and also to synchronize NNTP article numbers with the
> article numbers generated by lists.ruby-lang.org...

Actually, there were a number of changes needed for all that;
including implementing address obfuscation(*) since I assume
people here are conditioned to expect that from blade and gmane :<

So without further adieu: https://public-inbox.org/ruby-core/ and
http://ou63pmih66umazou.onion/ruby-core/ (for Tor users)
are both available.

I'll try to have other mirrors up soon for redundancy.

Since it is common to identify messages by [ruby-core:#{number}]
on this list, I made sure searching with the term prefix(**)
"core" will return the result for the message identified by
[ruby-core:12345]:

	https://public-inbox.org/ruby-core/?q=core:12345

I already added this feature ("altid") to support expired gmane
links for the git@vger.kernel.org list at https://public-inbox.org/git/


You can also clone all the messages (around 200 MB):

	git clone --mirror https://public-inbox.org/ruby-core

And the AGPL-3.0+ source code:

	git clone https://public-inbox.org/public-inbox

And hack + run it yourself.  I suggest doing this since I am an
unreliable sysadmin and have frequent reliability problems.


Here's the relevant portions for this list for the config file:
===> ~/.public-inbox/config <===
[publicinbox]
	; for public-inbox-nntpd(1)
	nntpserver = news.public-inbox.org
	nntpserver = ou63pmih66umazou.onion
[publicinboxlearn]
	; for public-inbox-watch(1)
	watchspam = maildir:/path/to/maildir/.INBOX.spam
[publicinboxwatch]
	spamcheck = spamc
[publicinbox "ruby-core"]
	address = ruby-core@ruby-lang.org
	url = //public-inbox.org/ruby-core
	mainrepo = /path/to/git/ruby-core.git
	newsgroup = inbox.comp.lang.ruby.core
	infourl = https://www.ruby-lang.org/en/community/mailing-lists/

	; for public-inbox-watch(1)
	watch = maildir:/path/to/maildir/.INBOX.ruby
	watchheader = List-Id:<ruby-core.ruby-lang.org>

	; new features implemented in June 2017:

	; altid is a preexisting feature, msgmap.sqlite3 is another
	; preexisting feature (for NNTP article numbers), but
	; combining them is new (via PublicInbox::Filter::RubyLang)
	; to maintain the existing serial numbers generated by the list
	altid = serial:core:file=msgmap.sqlite3

	; several input filters already exist (i.e. Vger), but a
	; new one was written to strip the "Unsubscribe:"
	; for RubyLang and maintain the altid <=> msgmap
	filter = PublicInbox::Filter::RubyLang

	; bleh(*)
	obfuscate = true

	; usually, this ends up just being reply-to-list, but
	; sometimes people not on the list are addressed
	; directly.  Adding "From" below would make it the same
	; as ":all" (which is what I'm used to, actually)
	replyto = :list,Cc,To,Reply-To
/EOF


Read-only newsgroups are also available over NNTP:

        nntp://news.public-inbox.org/inbox.comp.lang.ruby.core
        nntp://ou63pmih66umazou.onion/inbox.comp.lang.ruby.core

As implied above, NNTP article numbers correspond to the
ruby-core:#{number} generated by the list.

In the future, I will try to implement configurable auto-linkification,
so [ruby-core:#{number}] and [Bug ##{number}] can generate links
in the HTML (among many other features at https://public-inbox.org/TODO)

(*)   Address obfuscation causes accessibility problems for me
      and there are plenty of ways for spammers to get your
      address via other means.  Only filters like SpamAssassin
      (which combine multiple techniques) seems effective against
      spam.

(**)  It seems Xapian does not allow '-' in prefix, so I could not
      use "ruby-core", only "core" or "rubycore" as prefix, I chose
      the latter since it still seems to allow the extraneous "ruby-"
      to match based on the frequency of the term "ruby".
      (I am not a Xapian expert)


Anyways, thanks for reading!