Hello!
https://pi-dach.dorfdsl.de/rocksolid/overboard.php?thisgroup=de.alt.folklore.computer

The post itself looks fine when viewed directly, but in the "latest"
view, the encoding of special German characters is broken.

This affects
<AABmyvIvTIsAAAoF.A3.flnews@WStation7.micha.freeshell.org>
Message-ID: <slrnvckmpm.55rd.hjp-usenet4@trintignant.hjp.at>

--
kind regards
Marco

On Sun, 25 Aug 2024 11:32:03 +0200, Marco Moock wrote:

> Hello!
> https://pi-dach.dorfdsl.de/rocksolid/overboard.php?thisgroup=de.alt.folklore.computer
>
> The post itself looks fine when viewed directly, but in the "latest"
> view, the encoding of special German characters is broken.
>
> This affects
> <AABmyvIvTIsAAAoF.A3.flnews@WStation7.micha.freeshell.org>
> Message-ID: <slrnvckmpm.55rd.hjp-usenet4@trintignant.hjp.at>

Thank you. I've just added the group to my test server and will get on it.

I appreciate feedback from users about character display issues as the
groups I read are all English (guess why)

The first user, years ago, that got me on the right path to displaying
multibyte characters properly was Russian and was very helpful (very
specific feedback) in getting things working.

On 25.08.2024 um 15:12 Uhr Marcel Logen wrote:

> These articles use "charset=ISO-8859-1" for the body.

Is that RFC-conform or is there an issue with the article itself?

--
kind regards
Marco

Send spam to 1724591544muell@cartoonies.org

Marco Moock in rocksolid.nodes.help:

>On 25.08.2024 um 15:12 Uhr Marcel Logen wrote:

>> These articles use "charset=ISO-8859-1" for the body.
>
>Is that RFC-conform or is there an issue with the article itself?

The use of ISO-8859-1 is perfectly correct and RFC compliant,
but has become rather rare (in favor of UTF-8).

Marcel 1m2b8 (1771880)
--
╭───╮ ╭──╮ ╭─╮ ╭──────╮ ╭──╮ ╭──────╮
│ ╰───╯ │ │ ╰─╮ ╰────╮ ╰─────╯ ╰───╯ ╭───╯
╮ ╭─╮ ╭──╯ ╭─╯ ╭─╯ ╰──╮ ╭─╯ ╰───╮
╰────╯ ╰─────╯ ╰───╯ ╰──╯ b4a418 ╰──────

On Sun, 25 Aug 2024 16:05:00 +0200, Marcel Logen wrote:

> Marco Moock in rocksolid.nodes.help:
>
>>On 25.08.2024 um 15:12 Uhr Marcel Logen wrote:
>
>>> These articles use "charset=ISO-8859-1" for the body.
>>
>>Is that RFC-conform or is there an issue with the article itself?
>
> The use of ISO-8859-1 is perfectly correct and RFC compliant,
> but has become rather rare (in favor of UTF-8).

Yes, this charset is fine and should not be the problem.

I am working on this, but it will take some time. Part of the issue, but
not all of it, is decoding a portion of the body to create the "snippet". I
know it sounds simple to just decode the entire body and create the
snippet, but there is more to it than that.

It will take a bit of time to work out.

On Sun, 25 Aug 2024 15:13:13 +0000, Retro Guy wrote:

> On Sun, 25 Aug 2024 16:05:00 +0200, Marcel Logen wrote:
>
>> Marco Moock in rocksolid.nodes.help:
>>
>>>On 25.08.2024 um 15:12 Uhr Marcel Logen wrote:
>>
>>>> These articles use "charset=ISO-8859-1" for the body.
>>>
>>>Is that RFC-conform or is there an issue with the article itself?
>>
>> The use of ISO-8859-1 is perfectly correct and RFC compliant,
>> but has become rather rare (in favor of UTF-8).
>
> Yes, this charset is fine and should not be the problem.
>
> I am working on this, but it will take some time. Part of the issue, but
> not all of it, is decoding a portion of the body to create the
> "snippet". I
> know it sounds simple to just decode the entire body and create the
> snippet, but there is more to it than that.
>
> It will take a bit of time to work out.

Does this look better? https://postimg.cc/Zv4gwcVP

or did I make it worse somehow? :)

--
Retro Guy

On 25.08.2024 um 18:42 Uhr Retro Guy wrote:

> Does this look better? https://postimg.cc/Zv4gwcVP

Looks fine, umlauts are now working.

--
kind regards
Marco

Send spam to 1724604157muell@cartoonies.org

On Sun, 25 Aug 2024 19:49:46 +0000, Marco Moock wrote:

> On 25.08.2024 um 18:42 Uhr Retro Guy wrote:
>
>> Does this look better? https://postimg.cc/Zv4gwcVP
>
> Looks fine, umlauts are now working.

Great!

It turned out to be an issue specific to the headers for this poster,
and I'm sure some others that I haven't noticed. The character set was
not being properly extracted from the header due to the specific way the
header line was generated. This is now checked and the line is parsed
for this.

The fix will only work for newly spooled articles, as the "snippet" is
saved in the database for articles, it is not generated every time it is
displayed.

So new articles should display properly in Latest, but existing articles
will not change.

I have just pushed this fix, and several others, to devel.

--
Retro Guy

On 25.08.2024 um 20:45 Uhr Retro Guy wrote:

> I have just pushed this fix, and several others, to devel.

Is this currently considered stable for production or should I wait
before switching to the current version?

--
kind regards
Marco

Send spam to 1724611512muell@cartoonies.org

On Mon, 26 Aug 2024 13:02:02 +0200, Marco Moock wrote:

> On 25.08.2024 um 20:45 Uhr Retro Guy wrote:
>
>> I have just pushed this fix, and several others, to devel.
>
> Is this currently considered stable for production or should I wait
> before switching to the current version?

It should be fine. I just now merged devel into master, and pushed master
to gitlab.

So master is latest at the moment.

Please, always back up first.

You may wish to restart your nntp server(s) after upgrade, as
rslight-lib.php is modified. You'll need to kill them manually and let cron
restart them. In the upgraded version, restarting them is much simpler :)