I would like to make a request to LKML archives. It would be a highly
useful feature to have an archive site where one can algorithmically
produce a URL from the Message-ID, so one can post a clickable URL from
a delivered message without having to look it up every time.
For example, if a message had:
Message-ID: <[email protected]>
... and I could just generate a URL:
http://linux-kernel.example.com/msgid/[email protected]
... that would be highly convenient.
(Extra points if all the vger lists are available, not just LKML.)
-hpa
On 2007.06.20 08:37:48 -0700, H. Peter Anvin wrote:
> I would like to make a request to LKML archives. It would be a highly
> useful feature to have an archive site where one can algorithmically
> produce a URL from the Message-ID, so one can post a clickable URL from
> a delivered message without having to look it up every time.
>
> For example, if a message had:
>
> Message-ID: <[email protected]>
>
> ... and I could just generate a URL:
>
> http://linux-kernel.example.com/msgid/[email protected]
>
> ... that would be highly convenient.
>
> (Extra points if all the vger lists are available, not just LKML.)
marc.info supports that.
http://marc.info/[email protected]
Bj?rn
Bj?rn Steinbrink wrote:
>
> marc.info supports that.
>
> http://marc.info/[email protected]
>
Excellent! Didn't see that documented anywhere.
-hpa
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Wed, 20 Jun 2007, H. Peter Anvin wrote:
> Bj?rn Steinbrink wrote:
>>
>> marc.info supports that.
>>
>> http://marc.info/[email protected]
>
> Excellent! Didn't see that documented anywhere.
=)
You're right, it's not really. In fact very little is documented.
You'd have to discover it -- in the message view, the Message-ID is
actually a self-referential link using the above syntax.
A few relevant points:
- - Such links never actually show you the corresponding message. Instead:
- If the message-id is unique, you are sent a 302 redirect to the
message using the normal MARC message URL (l=foo&msgid=123)
- If the message-id is not unique (such as crossposting, or deliberate
attempts to cause collisions) you'll be presented with a list of
messages, showing date, subject, author, and listname, to choose from.
- - MARC doesn't currently touch message bodies for spam-obfuscation
purposes, but it does for headers, specifically From: and Message-ID:.
When MARC generates message-id-based links, they are obfuscated.
But it can read in links formed either way. So, if you copy/paste the
target of a Message-ID: link out of MARC, it will look slightly
different than if you just made your own http://marc.info/?i=what@not,
but either form will work.
- - We don't use < > surrounding the message-id, but again, will accept
URLs formed either way.
- - We don't have Message-ID's for messages added before Dec, 2003.
Please feel free to ping me (offlist) with questions/concerns, hopefully
I will manage to respond :-P
Thanks,
Hank
-----BEGIN PGP SIGNATURE-----
iD8DBQFGeVgYqP26fHCT+PMRAsFHAJ9G2v+vrQ4g7VUAWEr1+Tb3dHDi3wCfRfho
sCYaEfLf9cJjGZtgpsCAPII=
=PpNt
-----END PGP SIGNATURE-----
"H. Peter Anvin" <[email protected]> writes:
> For example, if a message had:
>
> Message-ID: <[email protected]>
>
> ... and I could just generate a URL:
>
> http://linux-kernel.example.com/msgid/[email protected]
gmane can do that:
http://mid.gmane.org/[email protected]
> (Extra points if all the vger lists are available, not just LKML.)
gmane has thousands of free software mailing lists. I don't know
whether it has all of the vger lists.
--
Ben Pfaff
http://benpfaff.org
* Newsgroups: gmane.linux.kernel
* Date: Wed, 20 Jun 2007 11:10:53 -0700
>
>> For example, if a message had:
>>
>> Message-ID: <[email protected]>
>>
>> ... and I could just generate a URL:
>>
>> http://linux-kernel.example.com/msgid/[email protected]
>
> gmane can do that:
> http://mid.gmane.org/[email protected]
To have Gmane (news<=>e-mail [and web]) engine[s] on the kernel.org will
be the best thing ever!
The "News<=>e-mail" part is major because of things, i've learned after
using it with (_mature_ news reader) slrn. Any web interface, basic like
in marcs.info, or comprehensive, like Gmane's one can't replace
functionality news server(Gmane)/reader is giving to you (summary):
* You don't receive (or annoyingly bounce) huge traffic.
. big backlogs are handled easily in many cases, like getting
e-mails, replying in interesting threads, just don't care about
flames. I've just removed (by one key-press) 300, 200, 100 messages
of Dual-Licensing thread, because i trust Linus.
* Interesting discussions are easily saved(by threads) on your side, so
you can search them or read (any mbox format reader, like `mutt -f`)
* (slrn) Searching Message-Id as easy as one key-combination and one
copy/paste {1}
* (slrn) Restoring of whole thread from one e-mail (or {1}) -- one
key-combination
* (slrn + Gmane) headers have "Archived-At" entry, so i can just
copy/paste it for web-based users
* (slrn + Gmane) listing of news-groups (actually ML<=>Gmane-NEWS)
as easy as all above: patterns, one key-combination
* i bet much more, than i've just told.
And that's pretty why i'm happy to be with. After abandoning Mozilla
(buggy, memory hungry ever stuff), X and having console, screen, slrn,
lynx, emacs-nox i'm a pretty much more efficient that any time back.
All web, rss, other XML/WEB-2.0 crap does will not help. Get your
developers back -- use and publish healthy ways and tools to use!
I mean, come on, that's easy, rather than develop IMHO useless
e-mail/patch brokenness (WRT plain-text) checkers and docs.
Publish your way of being productive and saving environment! (tm:)
>> (Extra points if all the vger lists are available, not just LKML.)
>
> gmane has thousands of free software mailing lists. I don't know
> whether it has all of the vger lists.
There are many, that i'm aware of: kernel, commits, ide, net, usb,
archs(amd64 at least), sparse, git...
Namespace in Gmane is a little bit wrong. For example
gmane.linux.kernel and gmane.linux.debian.*
I.e. Linux kernel (with sub-parts) is in noise of distributions, while
latter can have different kernel actually (e.g. kfreebsd in the Debian).
____
* Date: Thu, 21 Jun 2007 08:21:03 +0200
>
> To have Gmane (news<=>e-mail [and web]) engine[s] on the kernel.org will
> be the best thing ever!
Some more "Why":
>
> * You don't receive (or annoyingly bounce) huge traffic.
> . big backlogs are handled easily in many cases, like getting
> e-mails, replying in interesting threads, just don't care about
> flames. I've just removed (by one key-press) 300, 200, 100 messages
> of Dual-Licensing thread, because i trust Linus.
* No [for] stupid[s] ML trailes in every sigle e-mail
* Gmane News server handles crossposting, thus you will not hit same
message in LKML and linux-ide, for example (How i could forget
such useful feature? Nice things you are getting used to so fast
without even noticing it. :) <http://gmane.org/about.php>
* Spam is not only well-filtered, it is also removed, thus even if
you see such headers (what is depends on mailing list, gmane
doesn't remove them) you will not receive the body. After some
time they are removed completely.
> * Interesting discussions are easily saved(by threads) on your side, so
> you can search them or read (any mbox format reader, like `mutt -f`)
>
> * (slrn) Searching Message-Id as easy as one key-combination and one
> copy/paste {1}
>
> * (slrn) Restoring of whole thread from one e-mail (or {1}) -- one
> key-combination
>
> * (slrn + Gmane) headers have "Archived-At" entry, so i can just
> copy/paste it for web-based users
>
> * (slrn + Gmane) listing of news-groups (actually ML<=>Gmane-NEWS)
> as easy as all above: patterns, one key-combination
>
> * i bet much more, than i've just told.
<http://gmane.org/features.php>
* While it's service, it's free software available for download, but
"is next to impossible to set up properly".
Kudos to Lars Magne Ingebrigtsen!
____
On Wed, 2007-06-20 at 08:37 -0700, H. Peter Anvin wrote:
> I would like to make a request to LKML archives. It would be a highly
> useful feature to have an archive site where one can algorithmically
> produce a URL from the Message-ID, so one can post a clickable URL from
> a delivered message without having to look it up every time.
>
> For example, if a message had:
>
> Message-ID: <[email protected]>
>
> ... and I could just generate a URL:
>
> http://linux-kernel.example.com/msgid/[email protected]
>
> ... that would be highly convenient.
>
> (Extra points if all the vger lists are available, not just LKML.)
Even more extra points if there's a mailto: link allowing people to
reply, which actually gets the threading information right. E.g.:
mailto:[email protected],[email protected]?Subject=Re:%20Request %20for%20Linux%20Kernel%20Mailing%20List%20archives&[email protected]
--
dwmw2
On 2007.06.21 23:06:50 +0800, David Woodhouse wrote:
> On Wed, 2007-06-20 at 08:37 -0700, H. Peter Anvin wrote:
> > I would like to make a request to LKML archives. It would be a highly
> > useful feature to have an archive site where one can algorithmically
> > produce a URL from the Message-ID, so one can post a clickable URL from
> > a delivered message without having to look it up every time.
> >
> > For example, if a message had:
> >
> > Message-ID: <[email protected]>
> >
> > ... and I could just generate a URL:
> >
> > http://linux-kernel.example.com/msgid/[email protected]
> >
> > ... that would be highly convenient.
> >
> > (Extra points if all the vger lists are available, not just LKML.)
>
> Even more extra points if there's a mailto: link allowing people to
> reply, which actually gets the threading information right. E.g.:
>
> mailto:[email protected],[email protected]?Subject=Re:%20Request %20for%20Linux%20Kernel%20Mailing%20List%20archives&[email protected]
Or maybe alternatively a feature that forwards the mail to you. I find
it preferable to actually have the email you reply to. No copy'n'paste
for quoting, no manual quote marking, no manual "On $date wrote $author"
line, ...
lkml.org has that.
Bj?rn
On Thu, 2007-06-21 at 17:54 +0200, Björn Steinbrink wrote:
> Or maybe alternatively a feature that forwards the mail to you.
A mailing list where you actually _receive_ the mail. Such an innovative
idea... :)
But yes, I agree that's quite cute. You could even do it as an anonymous
read-only IMAP mailbox, perhaps?
> I find it preferable to actually have the email you reply to. No
> copy'n'paste for quoting, no manual quote marking, no manual "On $date
> wrote $author" line, ...
You can put all that into a mailto: URL :)
--
dwmw2
David Woodhouse wrote:
> On Thu, 2007-06-21 at 17:54 +0200, Björn Steinbrink wrote:
>> Or maybe alternatively a feature that forwards the mail to you.
>
> A mailing list where you actually _receive_ the mail. Such an innovative
> idea... :)
>
> But yes, I agree that's quite cute. You could even do it as an anonymous
> read-only IMAP mailbox, perhaps?
Or NNTP.
-hpa
On Wed, Jun 20, 2007 at 12:38:48PM -0400, Hank Leininger wrote:
> On Wed, 20 Jun 2007, H. Peter Anvin wrote:
> >Bj?rn Steinbrink wrote:
> >>
> >>marc.info supports that.
> >>http://marc.info/[email protected]
> >Excellent! Didn't see that documented anywhere.
>
> You're right, it's not really. In fact very little is documented.
> You'd have to discover it -- in the message view, the Message-ID is
> actually a self-referential link using the above syntax.
>
> A few relevant points:
> - - Such links never actually show you the corresponding message. Instead:
> - If the message-id is unique, you are sent a 302 redirect to the
> message using the normal MARC message URL (l=foo&msgid=123)
> - If the message-id is not unique (such as crossposting, or deliberate
> attempts to cause collisions) you'll be presented with a list of
> messages, showing date, subject, author, and listname, to choose from.
> - - MARC doesn't currently touch message bodies for spam-obfuscation
> purposes, but it does for headers, specifically From: and Message-ID:.
> When MARC generates message-id-based links, they are obfuscated.
> But it can read in links formed either way. So, if you copy/paste the
> target of a Message-ID: link out of MARC, it will look slightly
> different than if you just made your own http://marc.info/?i=what@not,
> but either form will work.
> - - We don't use < > surrounding the message-id, but again, will accept
> URLs formed either way.
> - - We don't have Message-ID's for messages added before Dec, 2003.
Thanks for that usefull info.
Attached is a script containing a quick hack to enable look up of
a message-id at MARC with mutt(ng).
Posted to LKML just in case some other mutt(ng) user subscribed here
finds it useful.
Regards,
Andreas
--
Operating | AMD Saxony Limited Liability Company & Co. KG,
System | Wilschdorfer Landstr. 101, 01109 Dresden, Germany
Research | Register Court Dresden: HRA 4896, General Partner authorized
Center | to represent: AMD Saxony LLC (Wilmington, Delaware, US)
(OSRC) | General Manager of AMD Saxony LLC: Dr. Hans-R. Deppe, Thomas McCoy