~cnx

Ulsan, South Korea

https://cnx.gdn

Vietnamese free software enthusiast

Join the FSF

~cnx/loang

Last active 9 days ago

~cnx/ipwhl-discuss

Last active 30 days ago

~cnx/misc

Last active 8 months ago

~cnx/ipwhl-announce

Last active 8 months ago

~cnx/ipwhl-devel

Last active 9 months ago

~cnx/blackshades

Last active 10 months ago

~cnx/site

Last active 1 year, 2 months ago

~cnx/tzigame

Last active 1 year, 5 months ago

~cnx/palace

Last active 1 year, 6 months ago

~cnx/nwsim

Last active 1 year, 11 months ago
View more

Recent activity

Fwd: [vpsFree.cz] Outage - node4.brq - 2023-03-16 08:15 CET 9 days ago

From Nguyễn Gia Phong to ~cnx/loang

Forwarded message from vpsFree.cz on 2023-03-16 at 08:36+01:00:
> we would like to inform you about an outage
> on vpsFree.cz infrastructure.
>
> Directly affected VPS:
>
>   - VPS #21857 - brno (on node4.brq)
>
> Date and time: 2023-03-16 08:15 CET
> Expected duration: 35 minutes
> Affected systems: node4.brq
> Outage type: vps_reset
> Reason: Kernel lockup
> Handled by: Jakub Skokan

[PATCH aerc] parse msg-id lists more liberally 14 days ago

From Nguyễn Gia Phong to ~rjarry/aerc-devel

Some user agents deliberately generate non-standard
message identifier lists in In-Reply-To and References headers.

Instead of failing silently, aerc now falls back to a desperate parser
scavaging whatever looking like <id-left@id-right>.

As the more liberal parser being substituted with, References header
are now stored for IMAP not only when there's a parsing error.

Fixes: 31d2f5be3cec ("message-info: add explicit References field")
Signed-off-by: Nguyễn Gia Phong <mcsinyx@disroot.org>
---
> Hmm I don't understand what this does. Can you give some examples of the
> expected input/output of this function? It feels like there is a simpler
[message trimmed]

[RFC PATCH] Parse msg-id lists more liberally 16 days ago

From Nguyễn Gia Phong to ~rjarry/aerc-devel

Some user agents deliberately generate non-standard
message identifier lists in In-Reply-To and References headers.

Instead of failing silently, aerc now falls back to a desperate parser
scavaging whatever looking like <id-left@id-right>.
---
I'm unsure about the change in worker/imap/cache.go
 commands/msg/reply.go | 11 +++--------
 lib/parse/header.go   | 42 ++++++++++++++++++++++++++++++++++++++++++
 models/models.go      | 13 ++++---------
 worker/imap/cache.go  |  7 ++++---
 worker/imap/fetch.go  |  6 ++----
 worker/lib/parse.go   | 17 ++++-------------
 6 files changed, 59 insertions(+), 37 deletions(-)
[message trimmed]

Re: aerc handling of message ID list in headere 23 days ago

From Nguyễn Gia Phong to ~rjarry/aerc-discuss

On 2023-03-02 at 23:45+06:00, Kirill Isakov wrote:
> From looking at go-message bugtracker (which is used by aerc
> to parse email) I got the impression that workarounds for third-party
> brain damage will not be merged.

Considering aerc's workaround for In-Reply-To
in https://lists.sr.ht/~rjarry/aerc-devel/patches/36782
(falling back to the entire string as one message ID),
I suppose the same could be done for References.

It would break threading though, so alternatively
a more liberal parser for message ID lists could
be implemented as a fallback.

Re: aerc putting invalid message ID in references header 23 days ago

From Nguyễn Gia Phong to ~rjarry/aerc-discuss

On 2023-03-02 at 11:23-06:00, Tim Culverhouse wrote:
> Mine should at least be fixed now.

Thanks, I can confirm that.

On 2023-03-02 at 11:23-06:00, Tim Culverhouse wrote:
> In one of the RFCs somewhere I read you should be liberal in what you
> accept and strict in what you send.  We are strict in what we accept
> *and* send, because we solely rely on go-message for that parsing. I
> don't think ~emersion wants to change the library to accept incorrect
> fields though, so it would be something we need to do internally to
> aerc.

Yea the robustness principle was stated in (now obsolete) RFC 761

Re: aerc putting invalid message ID in references header 23 days ago

From Nguyễn Gia Phong to ~rjarry/aerc-discuss

On 2023-03-02 at 11:01-06:00, Tim Culverhouse wrote:
> On Thu Mar 2, 2023 at 4:08 AM EST, Nguyễn Gia Phong wrote:
> > I got some messages hidden from aerc listing recently
>
> Is it only messages from me? I have a local patch
> that is messing with this which is likely causing the issue...

Unfortunately not, the others are from my institute's office 365
(so I guess an out look MUA of some kind?).  They actually violate
RFC 5322 (and 2822 for that matter) by intentionally separating
msg-id using commas, as if micro shaft is allergic to standards
or something.

aerc putting invalid message ID in references header 23 days ago

From Nguyễn Gia Phong to ~rjarry/aerc-discuss

Hi,

I got some messages hidden from aerc listing recently
and funny enough some are sent from aerc, e.g.
https://lists.sr.ht/~rjarry/aerc-discuss/%3CCQV9KPKE41TE.WEKVERBCAY3D%40spunky%3E/raw

This is due to invalid headers, in this case a message ID of <>
in the references header.  Other than that aerc should always
construct messages in valid form, I think it should also be more robust,
i.e. an invalid header should be ignored instead of causing its message
to be unviewable.

Maintainers and users of aerc, what do you think?

Re: Any ETA for sourcehut as a registar/DNS provider? 24 days ago

From Nguyễn Gia Phong to ~sircmpwn/sr.ht-discuss

On 2023-02-28 at 21:35-05:00, Amolith wrote:
> Once other people begin relying on what you host,
> redundancy becomes much more important.  At that point,
> I would say a minimum of two self- or professionally-managed
> [secondary] nameservers is ideal.

Thanks for the advise!

On 2023-02-28 at 21:35-05:00, Amolith wrote:
> On 2023-03-01 at 11:29+09:00, Nguyễn Gia Phong wrote:
> > I always wonder why (almost) no public free software service host
> > offers authoritative DNS.  Is there a security implication
> > or just the lack of demand?
>

Re: Any ETA for sourcehut as a registar/DNS provider? 24 days ago

From Nguyễn Gia Phong to ~sircmpwn/sr.ht-discuss

On 2023-02-28 at 19:00-05:00, Amolith wrote:
> In my experience, authoritative DNS is quite trivial
> for most minimally experienced admins to run themselves.

As a minimally experienced admin, I can confirm this.
I run Knot instead of PowerDNS though.

On 2023-02-28 at 19:00-05:00, Amolith wrote:
> the other two [VPSes] as auto-secondaries

How risky is it to only run a primary service and rely
on a free secondary service like ns-global.zone?
(On top of that, I also run other services on the same system,
which I'm aware to not be the best choice, but the current VPS

IPWHL user survey and furture of mirror on loang.net 30 days ago

From Nguyễn Gia Phong to ~cnx/loang

Good morning,

IPFS has been eating increasingly more memory lately on loang.net
to a point it triggers OOM every once in a while.  It is currently
only hosting https://git.sr.ht/~cnx/ipwhl-data releases.

For the stability of other services hosted on loang.net,
I intend to spin down the IPFS service at the end of this march.
If you use this IPWHL mirror but aren't capable of self-hosting,
please respond to this message and and hopefully we can figure
something out.

Good afternoon, good evening, and good night,
McSinyx