Hello!

If I reply to a message with long-lines and insert the quoted text, the
too long lines will be broken up, but the quote sign ">" is missing in
the next line.

You can see that by trying to reply to that (please do that only in a
test group).
Message-ID: <v25lnk$cocr$2@solani.org>
https://pi-dach.dorfdsl.de/rocksolid/article-flat.php?id=1003&group=de.soc.recht.misc#1003

This results in
|> nach mehr oder weniger sinnfreien "Diskussionen" darüber, seit wann
|denn nun wo die Bahnsteigkarten Geschichte sind, mal was Sinnvolles:

--
kind regards
Marco

Send spam to 1715933509muell@cartoonies.org

On Fri, 17 May 2024 10:14:34 +0200, Marco Moock wrote:

> Hello!
>
> If I reply to a message with long-lines and insert the quoted text, the
> too long lines will be broken up, but the quote sign ">" is missing in
> the next line.
>
> You can see that by trying to reply to that (please do that only in a
> test group).
> Message-ID: <v25lnk$cocr$2@solani.org>
> https://pi-dach.dorfdsl.de/rocksolid/article-flat.php?id=1003&group=de.soc.recht.misc#1003
>
> This results in
>|> nach mehr oder weniger sinnfreien "Diskussionen" darüber, seit wann
>|denn nun wo die Bahnsteigkarten Geschichte sind, mal was Sinnvolles:

This was fixed in a previous commit. I'll try to find the commit, but it
has been fixed.

On Fri, 17 May 2024 02:23:56 -0700, Retro Guy wrote:

> On Fri, 17 May 2024 10:14:34 +0200, Marco Moock wrote:
>
>> Hello!
>>
>> If I reply to a message with long-lines and insert the quoted text, the
>> too long lines will be broken up, but the quote sign ">" is missing in
>> the next line.
>>
>> You can see that by trying to reply to that (please do that only in a
>> test group).
>> Message-ID: <v25lnk$cocr$2@solani.org>
>> https://pi-dach.dorfdsl.de/rocksolid/article-flat.php?id=1003&group=de.soc.recht.misc#1003
>>
>> This results in
>>|> nach mehr oder weniger sinnfreien "Diskussionen" darüber, seit wann
>>|denn nun wo die Bahnsteigkarten Geschichte sind, mal was Sinnvolles:
>
> This was fixed in a previous commit. I'll try to find the commit, but it
> has been fixed.

Here it is:
https://gitlab.com/rslight-public/rocksolid-light/-/commit/dfbee69ee8a48fbe141b022fddfb6f82377ec2f0

On 17.05.2024 um 02:28 Uhr Retro Guy wrote:

> Here it is:
> https://gitlab.com/rslight-public/rocksolid-light/-/commit/dfbee69ee8a48fbe141b022fddfb6f82377ec2f0

Ok, so is the current version considered stable to be used in a
productive environment?
If so, I will upgrade ASAP.

--
kind regards
Marco

Send spam to 1715905713muell@cartoonies.org

Marco Moock wrote:

> On 17.05.2024 um 02:28 Uhr Retro Guy wrote:

>> Here it is:
>> https://gitlab.com/rslight-public/rocksolid-light/-/commit/dfbee69ee8a48fbe141b022fddfb6f82377ec2f0

> Ok, so is the current version considered stable to be used in a
> productive environment?
> If so, I will upgrade ASAP.

There will always be bugs, it's just that I'm not aware of any critical bugs in the latest commit.

I'm running latest on all my sites and not having problems. On one site I've disabled memcache just to make sure it runs fine with it disabled (it does).

Or, you could just make the changes to post.php as seen in the commit. It's a simple change, just changing 'addslashes(mb_wordwrap($body, 75))' to 'addslashes($body)' in two lines.

It's up to you.

--
Retro Guy

On 17.05.2024 um 11:06 Uhr Retro Guy wrote:

> It's up to you.

I've now upgraded (git clone and then upgrade-script) and the result is
the same.

This is how post.php looks:

$message = message_post(quoted_printable_encode($subject), $nemail . " (" . quoted_printable_encode($name) . ")", $newsgroups, $references_array, addslashes($body), $_POST['encryptthis'], $_POST['encryptto'], strtolower($name), $_POST['fromname'], null, true);
} else {
$message =
message_post(quoted_printable_encode($subject), $nemail
. " (" . quoted_printable_encode($name) . ")",
$newsgroups, $references_array, addslashes($body),
$_POST['encryptthis'], $_POST['encryptto'],
strtolower($name), $_POST['fromname']);

--
kind regards
Marco

Send spam to 1715936814muell@cartoonies.org

On Fri, 17 May 2024 15:01:27 +0200, Marco Moock wrote:

> On 17.05.2024 um 11:06 Uhr Retro Guy wrote:
>
>> It's up to you.
>
> I've now upgraded (git clone and then upgrade-script) and the result is
> the same.
>
> This is how post.php looks:
>
> $message = message_post(quoted_printable_encode($subject), $nemail . " (" . quoted_printable_encode($name) . ")", $newsgroups, $references_array, addslashes($body), $_POST['encryptthis'], $_POST['encryptto'], strtolower($name), $_POST['fromname'], null, true);
> } else {
> $message =
> message_post(quoted_printable_encode($subject), $nemail
> . " (" . quoted_printable_encode($name) . ")",
> $newsgroups, $references_array, addslashes($body),
> $_POST['encryptthis'], $_POST['encryptto'],
> strtolower($name), $_POST['fromname']);

This is correct. mb_wordwrap is now missing from the function call, so the
line should not be wrapped.

Please make sure you don't have a cached page, or an article that has
already been wrapped. Hopefully it's working properly. I did test it here
when I made the change and the problem went away.

Retro Guy wrote:

> Marco Moock wrote:

>> On 17.05.2024 um 11:06 Uhr Retro Guy wrote:

>>> It's up to you.

>> I've now upgraded (git clone and then upgrade-script) and the result is
>> the same.

>> This is how post.php looks:

>> $message = message_post(quoted_printable_encode($subject), $nemail . " (" . quoted_printable_encode($name) . ")", $newsgroups, $references_array, addslashes($body), $_POST['encryptthis'], $_POST['encryptto'], strtolower($name), $_POST['fromname'], null, true);
>> } else {
>> $message =
>> message_post(quoted_printable_encode($subject), $nemail
>> . " (" . quoted_printable_encode($name) . ")",
>> $newsgroups, $references_array, addslashes($body),
>> $_POST['encryptthis'], $_POST['encryptto'],
>> strtolower($name), $_POST['fromname']);

> I'm replying to this to prove myself wrong (it's a common occurrence :)

> Let's see if this wraps incorrectly, then I can go back and try to fix it again. Grrrr....

Seems ok. The long line is not wrapped. It appears wrapped in the display of the article, but if I change the browser window size, it wraps in a different location. The line is only wrapped by the browser in display, but it's not a wrapped line.

Again, I await being wrong :)

--
Retro Guy

Marco Moock wrote:

> On 17.05.2024 um 11:06 Uhr Retro Guy wrote:

>> It's up to you.

> I've now upgraded (git clone and then upgrade-script) and the result is
> the same.

> This is how post.php looks:

> $message = message_post(quoted_printable_encode($subject), $nemail . " (" . quoted_printable_encode($name) . ")", $newsgroups, $references_array, addslashes($body), $_POST['encryptthis'], $_POST['encryptto'], strtolower($name), $_POST['fromname'], null, true);
> } else {
> $message =
> message_post(quoted_printable_encode($subject), $nemail
> . " (" . quoted_printable_encode($name) . ")",
> $newsgroups, $references_array, addslashes($body),
> $_POST['encryptthis'], $_POST['encryptto'],
> strtolower($name), $_POST['fromname']);

I'm replying to this to prove myself wrong (it's a common occurrence :)

Let's see if this wraps incorrectly, then I can go back and try to fix it again. Grrrr....

--
Retro Guy

On 17.05.2024 um 13:39 Uhr Retro Guy wrote:

> Seems ok. The long line is not wrapped. It appears wrapped in the
> display of the article, but if I change the browser window size, it
> wraps in a different location. The line is only wrapped by the
> browser in display, but it's not a wrapped line.

I talk about the reply edit window.

If I try to insert the quote, it will wrap the line, but still misses
the >.

I used another browser, I don't think this is a caching problem.

If you need, you can try it out on my instance.

It is really strange because it doesn't affect my test post.
https://pi-dach.dorfdsl.de/rocksolid/post.php?id=304&type=reply&group=de.test

But it does affect this post:
https://pi-dach.dorfdsl.de/rocksolid/post.php?id=1003&type=reply&group=de.soc.recht.misc

--
kind regards
Marco

Send spam to 1715945990muell@cartoonies.org

Marco Moock wrote:

> On 17.05.2024 um 13:39 Uhr Retro Guy wrote:

>> Seems ok. The long line is not wrapped. It appears wrapped in the
>> display of the article, but if I change the browser window size, it
>> wraps in a different location. The line is only wrapped by the
>> browser in display, but it's not a wrapped line.

> I talk about the reply edit window.

> If I try to insert the quote, it will wrap the line, but still misses
> the >.

> I used another browser, I don't think this is a caching problem.

> If you need, you can try it out on my instance.

> It is really strange because it doesn't affect my test post.
> https://pi-dach.dorfdsl.de/rocksolid/post.php?id=304&type=reply&group=de.test

> But it does affect this post:
> https://pi-dach.dorfdsl.de/rocksolid/post.php?id=1003&type=reply&group=de.soc.recht.misc

Are you inserting the quote using the Quote button, or pasting the quote?

Pasting the quote will break the line as I believe it will add linefeeds during cut/paste.

--
Retro Guy

On 17.05.2024 um 15:25 Uhr Retro Guy wrote:

> Are you inserting the quote using the Quote button, or pasting the
> quote?

Yes, I use that button.

--
kind regards
Marco

Send spam to 1715952302muell@cartoonies.org

On Fri, 17 May 2024 21:10:43 +0200, Marco Moock wrote:

> On 17.05.2024 um 15:25 Uhr Retro Guy wrote:
>
>> Are you inserting the quote using the Quote button, or pasting the
>> quote?
>
> Yes, I use that button.

I'm going to add the group to my test site.

I may need to reply to a post there as a test, but I'll state that in the
reply (in English, that's all I got, lol)

On 17.05.2024 um 13:12 Uhr Retro Guy wrote:

> I may need to reply to a post there as a test, but I'll state that
> in the reply (in English, that's all I got, lol)

The error occurs when inserting the text with the quote button, so it is
not needed to send the reply after that.

You can also just copy the original post into a test post in a test
group.

--
kind regards
Marco

Send spam to 1715944363muell@cartoonies.org

Marco Moock wrote:

> On 17.05.2024 um 13:12 Uhr Retro Guy wrote:

>> I may need to reply to a post there as a test, but I'll state that
>> in the reply (in English, that's all I got, lol)

> The error occurs when inserting the text with the quote button, so it is
> not needed to send the reply after that.

> You can also just copy the original post into a test post in a test
> group.

I've performed this test in a local group last night. I created a message with a very very long line, then sent it.

Then I clicked reply and quote and the message came up. The long line looks like it's wrapped, but it is only for the display. When I sent the reply the long line is still there, but it's still one long line, no added linefeeds at all. Quote characters '>' are in the correct location.

I believe, but could be wrong of course, that what you're seeing is word wrap in the browser and in the reply window. The message is never getting extra linefeeds added.

--
Retro Guy

On 18.05.2024 um 09:54 Uhr Retro Guy wrote:

> I believe, but could be wrong of course, that what you're seeing is
> word wrap in the browser and in the reply window. The message is
> never getting extra linefeeds added.

Sounds reasonable.
Is there a reason to do that?
Normally, those lines should be wrapped at 72 chars. Is there a reason
not to do that?

--
kind regards
Marco

Send spam to 1716018848muell@cartoonies.org

On Sat, 18 May 2024 14:24:08 +0200, Marco Moock wrote:

> On 18.05.2024 um 09:54 Uhr Retro Guy wrote:
>
>> I believe, but could be wrong of course, that what you're seeing is
>> word wrap in the browser and in the reply window. The message is
>> never getting extra linefeeds added.
>
> Sounds reasonable.
> Is there a reason to do that?
> Normally, those lines should be wrapped at 72 chars. Is there a reason
> not to do that?

Yes, because it's difficult :)

I need to write a proper function to do it properly, but haven't yet taken
the time to get to work on it. It's not as simple as just wrapping, as
that's where the problem shows up of wrapped quoted lines without '>'.

The function must differentiate between quoted lines, maintain quote depth,
and wrap accordingly. It will take a bit of time for me to get this done,
so for now we just don't wrap.

Retro Guy wrote:

> On Sat, 18 May 2024 14:24:08 +0200, Marco Moock wrote:

>> On 18.05.2024 um 09:54 Uhr Retro Guy wrote:
>>
>>> I believe, but could be wrong of course, that what you're seeing is
>>> word wrap in the browser and in the reply window. The message is
>>> never getting extra linefeeds added.
>>
>> Sounds reasonable.
>> Is there a reason to do that?
>> Normally, those lines should be wrapped at 72 chars. Is there a reason
>> not to do that?

> Yes, because it's difficult :)

> I need to write a proper function to do it properly, but haven't yet
> taken
> the time to get to work on it. It's not as simple as just wrapping, as
> that's where the problem shows up of wrapped quoted lines without '>'.

> The function must differentiate between quoted lines, maintain quote
> depth,
> and wrap accordingly. It will take a bit of time for me to get this
> done,
> so for now we just don't wrap.

I think we're about 90% there. Wrapping quoted text is working, and
maintaining the quote depth.

There still are some bugs, like what specifically should I use for a wrap
length, but it is working reasonably well right now.

This is being sent from my test server, so if this message looks horrible,
it's because I have failed :) Note: I did not manually wrap the lines in my
response here, I'm leaving that for rslight.

--
Retro Guy

Retro Guy wrote:

> Retro Guy wrote:

>> On Sat, 18 May 2024 14:24:08 +0200, Marco Moock wrote:

>>> On 18.05.2024 um 09:54 Uhr Retro Guy wrote:
>>>
>>>> I believe, but could be wrong of course, that what you're seeing
>>>> is
>>>> word wrap in the browser and in the reply window. The message is
>>>> never getting extra linefeeds added.
>>>
>>> Sounds reasonable.
>>> Is there a reason to do that?
>>> Normally, those lines should be wrapped at 72 chars. Is there a
>>> reason
>>> not to do that?

>> Yes, because it's difficult :)

>> I need to write a proper function to do it properly, but haven't yet
>> taken
>> the time to get to work on it. It's not as simple as just wrapping,
>> as
>> that's where the problem shows up of wrapped quoted lines without
>> '>'.

>> The function must differentiate between quoted lines, maintain quote
>> depth,
>> and wrap accordingly. It will take a bit of time for me to get this
>> done,
>> so for now we just don't wrap.

> I think we're about 90% there. Wrapping quoted text is working, and
> maintaining the quote depth.

> There still are some bugs, like what specifically should I use for a
> wrap
> length, but it is working reasonably well right now.

> This is being sent from my test server, so if this message looks
> horrible,
> it's because I have failed :) Note: I did not manually wrap the lines
> in my
> response here, I'm leaving that for rslight.

I meant to send this from this test site. The other copy is from
novabbs.org. They look the same, lol.

I have changed the wrap length to 72 characters.

Question: Should I break long words (words with no space) or let them
be? It would mostly wrap urls I would think.

--
Retro Guy

Retro Guy wrote:

> On Sat, 18 May 2024 14:24:08 +0200, Marco Moock wrote:

>> On 18.05.2024 um 09:54 Uhr Retro Guy wrote:
>>
>>> I believe, but could be wrong of course, that what you're seeing is
>>> word wrap in the browser and in the reply window. The message is
>>> never getting extra linefeeds added.
>>
>> Sounds reasonable.
>> Is there a reason to do that?
>> Normally, those lines should be wrapped at 72 chars. Is there a reason
>> not to do that?

> Yes, because it's difficult :)

> I need to write a proper function to do it properly, but haven't yet taken
> the time to get to work on it. It's not as simple as just wrapping, as
> that's where the problem shows up of wrapped quoted lines without '>'.

> The function must differentiate between quoted lines, maintain quote depth,
> and wrap accordingly. It will take a bit of time for me to get this done,
> so for now we just don't wrap.

I have changed the wrap length to 72 characters.

Question: Should I break long words (words with no space) or let them be? It would mostly wrap urls I would think.

--
Retro Guy

On 18.05.2024 um 18:39 Uhr Retro Guy wrote:

> Question: Should I break long words (words with no space) or let them
> be? It would mostly wrap urls I would think.

Keep them together. Proper hyphenation is much more complicated.
If a URL is longer, it should be kept that.

--
kind regards
Marco

Send spam to 1716050376muell@cartoonies.org

Marco Moock wrote:

> On 18.05.2024 um 18:39 Uhr Retro Guy wrote:

>> Question: Should I break long words (words with no space) or let
>> them
>> be? It would mostly wrap urls I would think.

> Keep them together. Proper hyphenation is much more complicated.
> If a URL is longer, it should be kept that.

I agree. I'll just leave long words alone, and have decided on 72
characters.

I'm not fully confident to run this mod yet, but will push it to devel
in a few minutes if you feel like testing with it.

It's one line added to post.php, and a new function in newsportal.php

Replying using the mod again here...

--
Retro Guy

On 18.05.2024 um 19:08 Uhr Retro Guy wrote:

> I'm not fully confident to run this mod yet, but will push it to devel
> in a few minutes if you feel like testing with it.

Don't hurry, it isn't a big problem.

--
kind regards
Marco

Send spam to 1716052095muell@cartoonies.org

Marco Moock wrote:

> On 18.05.2024 um 19:08 Uhr Retro Guy wrote:

>> I'm not fully confident to run this mod yet, but will push it to
>> devel
>> in a few minutes if you feel like testing with it.

> Don't hurry, it isn't a big problem.

Sounds good. It's in devel now, and I'm running it on
www.rocksolidbbs.com and www.novabbs.org, but not yet on www.novabbs.com
(as that's the site that would get the most people upset if it didn't
work as expected)

So, I'll be posting with it for now, as I usually post from
novabbs.org.

On a side note, all my posts in rocksolid.* are either posted via
rslight, or 40tude Dialog, which uses the rslight nntp server, not an
inn server. That allows me to hopefully find bugs earlier than I
otherwise would.

This mod will only wrap messages posted via rslight. It will not modify
posts made with a newsreader (so my 40tude posts do not have any
modifications applied, they are just how my newsreader posts)

</end useless information>

--
Retro Guy

Marco Moock wrote:

> On 18.05.2024 um 18:39 Uhr Retro Guy wrote:

>> Question: Should I break long words (words with no space) or let them
>> be? It would mostly wrap urls I would think.

> Keep them together. Proper hyphenation is much more complicated.
> If a URL is longer, it should be kept that.

Quick mod to discard the depth count before wrapping. Hopefully this
will resolve the one word lines that appear.

Let's see.

--
Retro Guy