Christopher Howard <christopher@librehacker.com> writes:

> Hi, as mentioned earlier, I'm happy to report that the elpher SOCKS wrapper is
> working great. However, there is one little hangup for me. Personally, I like
> to have elpher-gemini-TLS-cert-checks set to ON — as it doesn't bother me to
> have to explicitly accept a new cert, as opposed to the alternative of not
> checking certs at all. However, it seems that enabling elpher-socks-always
> somehow causes cert checking to be once again disabled. Do you know if this is
> an issue in the elpher code, or something happening back in socks.el?

The problem is on the Emacs side. As things stand, you can't have `nsm'
checks without also leaking DNS. See [1] (in particular, the first two
patches), and reply to that bug thread if you'd like to see any action
on this front. Thanks.

Hi Christopher,

Christopher Howard <christopher@librehacker.com> writes:

> Hi, I want to switch over to elpher as my primary gemini client, but I have a
> requirement that I need to be able to route the connections through my TOR
> daemon, available at 127.0.0.1:9050. I see that elpher has SOCKS5 support but
> it doesn't seem to be working for me. Or perhaps I am not setting it up
> correctly.
>
> I didn't see a guide for this, but I set these variables (through Custom):
>
>  '(elpher-socks-always t)
>  '(socks-server '("Default server" "socks" 9050 5))

This adds an -events option to log JOINs, PARTs, etc., as mentioned in [1].
The log formatting in WriteMessage is copied from formatMessage in
soju/msgstore/znclog. It differs only in ZNC 1.9-style "Joins:" lines for
logged-in users [2]. "Joins:" parsing in UnmarshalLine should not be affected.

[1] https://todo.sr.ht/~emersion/chathistorysync/1
[2] https://github.com/znc/znc/pull/1870
---
 client.go   | 14 ++++++++++++
 main.go     |  4 ++++
 msgstore.go | 62 ++++++++++++++++++++++++++++++++++++++++++++++++++++-
 3 files changed, 79 insertions(+), 1 deletion(-)

diff --git a/client.go b/client.go
[message trimmed]

Hi, these changes appear to work for my purposes, but I'm only just trying them
out now. Perhaps others needing this functionality would like to as well. Thanks.

Simon Ser <contact@emersion.fr> writes:

> Would you be able to run chathistorysync on a timer instead?

D'oh! Found it:

  https://sr.ht/~emersion/chathistorysync/

Yeah, that should work (cue walk of shame music). Cheers.

Simon Ser <contact@emersion.fr> writes:

> Would you be able to run chathistorysync on a timer instead?

Er, um, hopefully? Here's the part where I admit I don't know what that
means. :/

I guess I was under the (mistaken) impression "chathistorysync" meant
having a client fetch the latest history, but now I'm fairly convinced
you're referring to something rather specific. (FWIW, I *did* grep the
code base for that term, but alas...)

This adds the helper script contrib/unmigrate-logs, which complements
contrib/migrate-logs to offer round-trip mobility of database history to and
from msgstore 'fs' logs. It can run as a cron job to update an ancillary store
for things like maintaining a centralized file-based search index.
---
 contrib/unmigrate-logs/main.go                | 388 ++++++++++++++++++
 contrib/unmigrate-logs/main_test.go           | 362 ++++++++++++++++
 .../soju-test-user/testnet/#a/2024-06-04.log  |   5 +
 .../soju-test-user/testnet/#a/2024-06-05.log  |   1 +
 .../soju-test-user/testnet/#b/2024-06-04.log  |   1 +
 .../soju-test-user/testnet/#b/2024-06-05.log  |   1 +
 .../testnet/memoserv/2024-06-04.log           |   2 +
 .../testnet/tester/2024-06-04.log             |   2 +
 .../testnet/tester/2024-06-05.log             |   2 +
[message trimmed]

Enabling the option 'Stable' specifies a secondary sort key when querying the
database for a list of messages or targets. Its meaning differs slightly
depending on the nature of the query. In ListMessages, it breaks ties by
primary (row) ID, which normally corresponds to original insertion order.
However, in ListMessageLastPerTarget, the secondary sort is lexical, by target
name, which is unrelated to the order in which targets were originally added.
Note that target names are effectively unique when performing queries per
network, as done by this function.
---
 database/database.go |  1 +
 database/postgres.go | 12 ++++++++++++
 database/sqlite.go   | 12 ++++++++++++
 3 files changed, 25 insertions(+)
[message trimmed]

Hi, I mentioned this in the Libera channel a few days ago. Basically, I have
occasion to keep a secondary 'fs' msgstore alongside a bouncer running a 'db'
one. My search and retrieval solution going back years operates on a shared
archive of centrally indexed ZNC logs. While I do plan on eventually
transitioning to something in-band, like soju.im/search, I need to stick with
the status quo for now. Thus, I've been experimenting with ways to periodically
sync logs, as was suggested in the channel. I'm currently running the script
shown in the second patch on a systemd.timer unit. So far, no issues, but it'd
be nice if others in the same boat could try it and offer feedback. I'd also
appreciate any insights from those familiar with the code base, since I clearly
don't know what I'm doing. Also, if anyone more competent has a superior
solution or wants to take over, please be my guest. I don't need any credit and
would be eternally grateful. Many thanks.

Hi Troels,

Troels Henriksen <athas@sigkill.dk> writes:

> Sometimes I click on a gopher:// link to a binary file in ERC (the Emacs
> IRC client), which will then invoke elpher to download it.  However,
> it'll eventually fail in `elpher-render-download` because
> `(elpher-page-address elpher-current-page)` returns NIL, and this is not
> a valid value for `elpher-address-gopher-p`.  I think the problem is
> that whatever mechanism invokes elpher through ERC (which I didn't
> manage to trace) doesn't actually switch to an elpher buffer.  Any ideas
> on how to fix it?

By default, ERC relies on `browse-url-button-open-url' for visiting most