Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.debian.user > #224476 > unrolled thread
| Started by | "Will Mengarini" <seldon@eskimo.com> |
|---|---|
| First post | 2020-07-07 07:20 +0200 |
| Last post | 2020-07-07 17:10 +0200 |
| Articles | 3 — 3 participants |
Back to article view | Back to linux.debian.user
Real bounces from debian-user "Will Mengarini" <seldon@eskimo.com> - 2020-07-07 07:20 +0200
Re: Real bounces from debian-user Kushal Kumaran <kushal@locationd.net> - 2020-07-07 16:40 +0200
Re: Real bounces from debian-user David Wright <deblis@lionunicorn.co.uk> - 2020-07-07 17:10 +0200
| From | "Will Mengarini" <seldon@eskimo.com> |
|---|---|
| Date | 2020-07-07 07:20 +0200 |
| Subject | Real bounces from debian-user |
| Message-ID | <ApObo-5tf-3@gated-at.bofh.it> |
TL;DR: How can I use *one* query (web or otherwise)
to retrieve *all* my recently-bounced debian-user mail,
or a list of URLs by which they can be retrieved?
DETAILS:
In the past 3 months I've averaged about 5 messages monthly
saying "lists.debian.org has received bounces from you".
These don't tell me WHICH messages bounced, only HOW MANY.
They contain a link to retrieve only the most recent bounce.
I've read https://wiki.debian.org/DebianMailingLists/,
and know how to retrieve individual messages from
<https://lists.debian.org/msgid-search/>, though I didn't notice
documentation saying whether that includes Resent-Message-ID.
But none of that suffices to fix broken threads without laboriously
searching the archives. I'd need to code a hairy script to:
- Enumerate recently received debian-user messages in my inbox,
saving Resent-Message-ID if supported, else Message-ID,
as well as author name in a form recognizable on web rendering
(which might require decoding) (because a full archive list shows
authors and subjects, but neither kind of message ID).
- Use archives to count messages from each author.
(Subjects can change during threads, so just add complexity.)
- FOR AUTHORS WHOSE MESSAGE COUNTS DON'T MATCH, RETRIEVE
EACH OF THEIR FULL MESSAGES JUST TO GET MESSAGE-ID!
- Compare all retrieved messages' Message-ID headers to those I'd
enumerated from my inbox to see what's missing.
- For each missing message, mail it to myself, using mutt -H to
strip web cruft (says the wiki - haven't tried it).
Gleep!
Is there a simpler solution to this problem?
If bounce messages included links for *all* bounces, instead of just
the most recent, they would suffice; a much simpler script could
retrieve each message and mail it to me.
If there's some URL that can retrieve a list of *all* recent bounces for
a user, that'd suffice as well; any "you got bounces" messages would
take one hit to the my-bounces URL, then one hit for each message.
If there's some URL that can say "retry all my recent bounces
again now", that'd be best; most often these messages result
from problems my ISP is having with hardware instability of
specific {leading or bleeding}-edge kernels on new servers, and
by the time I start getting mail again, the problem has been
fixed with a new kernel, so asking for all my retries would work.
If this has to be coped with from luser-land, it probably needs that
hairy script I described at the beginning. Has anybody implemented
anything similar, or any useful parts of that solution?
But solutions from list admins would be better, because
they'd entail much less web traffic to cope with unavoidable
bounces. How hard would it be for you-got-bounces messages
to include links for all (or some reasonable number) of
the bounced messages, or to include a link returning all
bounces as attachments, or as one attachment in mbox format?
How hard would it be for an alternative archive listing to
include message IDs or URLs that could be directly compared
with mail headers to identify missing messages? I see that
individual messages are fully retrievable with the List-Archive
link, but comparing those with messages in the existing archive
listings (e.g <https://lists.debian.org/debian-user/2020/07/>)
seems unnecessarily difficult. I did think to look at the
source of that HTML, where I see nice stuff like name="00231"
href="msg00231.html"; but I don't know how to translate from the
List-Archive header to the sequence number for a particular month.
For example, headers in the original mailing list (i.e SMTP) version
of <https://lists.debian.org/debian-user/2020/07/msg00231.html> have
--------------------------------
List-Archive: https://lists.debian.org/msgid-search/20200706222029.GS29877@randomstring.org
Message-ID: <20200706222029.GS29877@randomstring.org>
Resent-Message-ID: <7G4bOlRoN4O.A.hbB.6O6AfB@bendel>
X-Mailing-List: <debian-user@lists.debian.org> archive/latest/764776
Date: Mon, 6 Jul 2020 18:20:29 -0400
From: Dan Ritter <dsr@randomstring.org>
Subject: Re: Firefox non-ESR update needed
--------------------------------
but the Message-ID is from randomstring.org, and though it
begins with a date, the following 222029.$blurf doesn't have any
relationship to Debian data; and the only valid hex string in
the Resent-Message-ID is "A", which seems unlikely to help. I
didn't try base 36, because FML; was that a mistake? Note that
once I know "00231" the problem is solved, but the inefficiency
(many extra web accesses) occurs because I *don't* know it based
on the SMTP mail headers. I did notice the mysteriously alluring
"archive/latest/764776" after the X-Mailing-List header, but ...
404: <https://lists.debian.org/debian-user/archive/latest/764776>
404: <https://lists.debian.org/archive/latest/764776>
404: <https://lists.debian.org/debian-user/latest/764776>
... sigh. Anyway, if that could magically translate into
some URL, it doesn't help ipso facto, because it comes from
mail I *did* get; I need the URLs for mail I *didn't* get.
SUMMARY: How can I either:
- Retrieve all my recent debian-user bounces, or a list thereof;
or:
- Minimizing web traffic, retrieve a list of all recent debian-user
mail that I can then, minimizing hack time, compare with all
debian-user mail in my inbox to see what's missing?
--
Will Mengarini <seldon@eskimo.com>
Free software: the Source will be with you, always.
"I'm too old for this." -- 5000-year-old Methos, /Highlander/ S4E14
[toc] | [next] | [standalone]
| From | Kushal Kumaran <kushal@locationd.net> |
|---|---|
| Date | 2020-07-07 16:40 +0200 |
| Message-ID | <ApWVj-2gC-1@gated-at.bofh.it> |
| In reply to | #224476 |
"Will Mengarini" <seldon@eskimo.com> writes:
> TL;DR: How can I use *one* query (web or otherwise)
> to retrieve *all* my recently-bounced debian-user mail,
> or a list of URLs by which they can be retrieved?
>
I don't know if this would be too radical a change, but you can also use
a news reader to access the mailing lists. The wiki link you sent
mentions the linux.debian.* heirarchy. Since this works by your NNTP
client reaching out to the server to fetch messages, that eliminates any
bounces.
You can configure your news client to send messages by email to the
regular debian-user email address
> DETAILS:
>
> In the past 3 months I've averaged about 5 messages monthly
> saying "lists.debian.org has received bounces from you".
> These don't tell me WHICH messages bounced, only HOW MANY.
> They contain a link to retrieve only the most recent bounce.
>
> I've read https://wiki.debian.org/DebianMailingLists/,
> and know how to retrieve individual messages from
> <https://lists.debian.org/msgid-search/>, though I didn't notice
> documentation saying whether that includes Resent-Message-ID.
>
> But none of that suffices to fix broken threads without laboriously
> searching the archives. I'd need to code a hairy script to:
> - Enumerate recently received debian-user messages in my inbox,
> saving Resent-Message-ID if supported, else Message-ID,
> as well as author name in a form recognizable on web rendering
> (which might require decoding) (because a full archive list shows
> authors and subjects, but neither kind of message ID).
> - Use archives to count messages from each author.
> (Subjects can change during threads, so just add complexity.)
> - FOR AUTHORS WHOSE MESSAGE COUNTS DON'T MATCH, RETRIEVE
> EACH OF THEIR FULL MESSAGES JUST TO GET MESSAGE-ID!
> - Compare all retrieved messages' Message-ID headers to those I'd
> enumerated from my inbox to see what's missing.
> - For each missing message, mail it to myself, using mutt -H to
> strip web cruft (says the wiki - haven't tried it).
> Gleep!
>
> Is there a simpler solution to this problem?
>
> If bounce messages included links for *all* bounces, instead of just
> the most recent, they would suffice; a much simpler script could
> retrieve each message and mail it to me.
>
> If there's some URL that can retrieve a list of *all* recent bounces for
> a user, that'd suffice as well; any "you got bounces" messages would
> take one hit to the my-bounces URL, then one hit for each message.
>
> If there's some URL that can say "retry all my recent bounces
> again now", that'd be best; most often these messages result
> from problems my ISP is having with hardware instability of
> specific {leading or bleeding}-edge kernels on new servers, and
> by the time I start getting mail again, the problem has been
> fixed with a new kernel, so asking for all my retries would work.
>
> If this has to be coped with from luser-land, it probably needs that
> hairy script I described at the beginning. Has anybody implemented
> anything similar, or any useful parts of that solution?
>
> But solutions from list admins would be better, because
> they'd entail much less web traffic to cope with unavoidable
> bounces. How hard would it be for you-got-bounces messages
> to include links for all (or some reasonable number) of
> the bounced messages, or to include a link returning all
> bounces as attachments, or as one attachment in mbox format?
>
> How hard would it be for an alternative archive listing to
> include message IDs or URLs that could be directly compared
> with mail headers to identify missing messages? I see that
> individual messages are fully retrievable with the List-Archive
> link, but comparing those with messages in the existing archive
> listings (e.g <https://lists.debian.org/debian-user/2020/07/>)
> seems unnecessarily difficult. I did think to look at the
> source of that HTML, where I see nice stuff like name="00231"
> href="msg00231.html"; but I don't know how to translate from the
> List-Archive header to the sequence number for a particular month.
>
> For example, headers in the original mailing list (i.e SMTP) version
> of <https://lists.debian.org/debian-user/2020/07/msg00231.html> have
> --------------------------------
> List-Archive: https://lists.debian.org/msgid-search/20200706222029.GS29877@randomstring.org
> Message-ID: <20200706222029.GS29877@randomstring.org>
> Resent-Message-ID: <7G4bOlRoN4O.A.hbB.6O6AfB@bendel>
> X-Mailing-List: <debian-user@lists.debian.org> archive/latest/764776
> Date: Mon, 6 Jul 2020 18:20:29 -0400
> From: Dan Ritter <dsr@randomstring.org>
> Subject: Re: Firefox non-ESR update needed
> --------------------------------
> but the Message-ID is from randomstring.org, and though it
> begins with a date, the following 222029.$blurf doesn't have any
> relationship to Debian data; and the only valid hex string in
> the Resent-Message-ID is "A", which seems unlikely to help. I
> didn't try base 36, because FML; was that a mistake? Note that
> once I know "00231" the problem is solved, but the inefficiency
> (many extra web accesses) occurs because I *don't* know it based
> on the SMTP mail headers. I did notice the mysteriously alluring
> "archive/latest/764776" after the X-Mailing-List header, but ...
> 404: <https://lists.debian.org/debian-user/archive/latest/764776>
> 404: <https://lists.debian.org/archive/latest/764776>
> 404: <https://lists.debian.org/debian-user/latest/764776>
> ... sigh. Anyway, if that could magically translate into
> some URL, it doesn't help ipso facto, because it comes from
> mail I *did* get; I need the URLs for mail I *didn't* get.
>
> SUMMARY: How can I either:
> - Retrieve all my recent debian-user bounces, or a list thereof;
> or:
> - Minimizing web traffic, retrieve a list of all recent debian-user
> mail that I can then, minimizing hack time, compare with all
> debian-user mail in my inbox to see what's missing?
[toc] | [prev] | [next] | [standalone]
| From | David Wright <deblis@lionunicorn.co.uk> |
|---|---|
| Date | 2020-07-07 17:10 +0200 |
| Message-ID | <ApXol-2G7-3@gated-at.bofh.it> |
| In reply to | #224476 |
On Mon 06 Jul 2020 at 22:10:10 (-0700), Will Mengarini wrote: > TL;DR: How can I use *one* query (web or otherwise) > to retrieve *all* my recently-bounced debian-user mail, > or a list of URLs by which they can be retrieved? Would subscribing to the digest (as well) help? I haven't checked what arrives, but there are replies in debian-user that quote from it, and it looks as if the message-id is included in each message. From each digest, you can compile a list of message-id strings, and of course the digests are numbered so you know whether you have all the digests. If you find that Message-ID: <20200707051010.GB11861@eskimo.com> is missing from your own received emails, you can fetch it with https://lists.debian.org/msgid-search/20200707051010.GB11861@eskimo.com or use whatever that trick was to email it to yourself. Cheers, David.
[toc] | [prev] | [standalone]
Back to top | Article view | linux.debian.user
csiph-web