Good morning

For some time, I use to replace the usual quotation marks ' " ' by those that are expected in the current language, like ' « » ' for the French language or '
„” ' for German. I sometimes mix up things in English, but ' „ ” ' should be
okay... I guess.

Anyway, FLNews appears to have difficulties, when I use these special characters in the subject header. I do currently not see any need for other symbols than those occasional quote marks.

_Question_: Are we expected to avoid them in headers in general, especially in the subject-header or would there be a way to allow – what do you call them –
typographic quote marks with FLNews – in a future version, maybe? I will test with ' – ', too, but I do not care for more special characters at this moment.

Thank you anyway

Michael
--
Es ist an der Zeit

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Supersedes because I guessed wrong.

Good morning

For some time, I use to replace the usual quotation marks ' " ' by those that are expected in the current language, like ' « » ' for the French language or '
„” ' for German. I sometimes mix up things in English, but ' “ ” ' should be
okay... I guess.

Anyway, FLNews appears to have difficulties, when I use these special characters in the subject header. I do currently not see any need for other symbols than those occasional quote marks.

_Question_: Are we expected to avoid them in headers in general, especially in the subject-header or would there be a way to allow – what do you call them –
typographic quote marks with FLNews – in a future version, maybe? I will test with ' – ', too, but I do not care for more special characters at this moment.

Thank you anyway

Michael
--
Es ist an der Zeit

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Michael Uplawski wrote:

[...]
For some time, I use to replace the usual quotation marks ' " ' by those that are expected in the current language, like ' « » ' for the French language or '
„” ' for German. I sometimes mix up things in English, but ' “ ” ' should be
okay... I guess.

Anyway, FLNews appears to have difficulties, when I use these special characters in the subject header. I do currently not see any need for other symbols than those occasional quote marks.

If you use an external postprocessor, the MIME encoding for the header
is executed before postprocessing. This means that the postprocessor
must encode anything beyond US-ASCII according to RFC 2047.

Quoted from the flnews manual page:
|
| o post_proc (String) Pathname of an optional article postprocessor
|
| [...]
| The header is already in RFC 2047 format. If the postprocessor
| modifies an existing header field, this format must be preserved.
| If a new header field is added, the postprocessor is responsible
| for creating valid RFC 2047 encoding and Unicode normalization to NFC.

This was most likely the reason that it has not worked as expected.

_Question_: Are we expected to avoid them in headers in general, especially in
the subject-header or would there be a way to allow – what do you call them –
typographic quote marks with FLNews – in a future version, maybe? I will test
with ' – ', too, but I do not care for more special characters at this moment.

I think it should be avoided to modify the "Subject" header field
in general. Exceptions are:
- Repairing broken encoding
Every article should be RFC 5536 conformant.
- Explicit switching with the form "New subject (was: Old subject)"

For newsgroups with English language, I would suggest to at least limit
such modifications to articles that already use MIME encoding.
In other words: If the article was plain US-ASCII, this should still
be true after postprocessing.
Otherwise you may create problems for users of old newsreaders with
no or broken MIME support (that are still used for newsgroups with
English language by many people).

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Supersedes for presentation.

Mahlzeit

and thank you for the clarifications.

Michael Bäuerle wrote in news.software.readers:

Anyway, FLNews appears to have difficulties, when I use these special characters in the subject header. I do currently not see any need for other symbols than those occasional quote marks.

If you use an external postprocessor, the MIME encoding for the header
is executed before postprocessing. This means that the postprocessor
must encode anything beyond US-ASCII according to RFC 2047.

If I understand correctly, this applies even when I *do not wish to alter the Subject*. It would explain everything.

Quoted from the flnews manual page:

| The header is already in RFC 2047 format. If the postprocessor
| modifies an existing header field, this format must be preserved.
| If a new header field is added, the postprocessor is responsible
| for creating valid RFC 2047 encoding and Unicode normalization to NFC.

Hm. The problem arouse in an initial post, not in a follow-up and the subject is set in the flnews-internal editor. The post-processor does not touch the Subject.

This was most likely the reason that it has not worked as expected.

Okay, I have to look at my Ruby-code, although I scrutinize the existing headers only in the attempt to find the empty line, below.

I think it should be avoided to modify the "Subject" header field
in general. Exceptions are:
- Repairing broken encoding
Every article should be RFC 5536 conformant.
- Explicit switching with the form "New subject (was: Old subject)"

Okay, I have to verify all this, be it only to master my own crap a little better.

For newsgroups with English language, I would suggest to at least limit
such modifications to articles that already use MIME encoding.
In other words: If the article was plain US-ASCII, this should still
be true after postprocessing.

A complete description for a test-case. ;)

Otherwise you may create problems for users of old newsreaders with
no or broken MIME support (that are still used for newsgroups with
English language by many people).

I am aware of that but tend to become more and more fatalistic. Mime encoding shall however be my occupation for a while.

TY again.

Cheerio.
--
He's an ill cook that breeds contempt

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Mahlzeit

and thank you for the clarifications.

Michael Bäuerle wrote in news.software.readers:

Anyway, FLNews appears to have difficulties, when I use these special characters in the subject header. I do currently not see any need for other symbols than those occasional quote marks.

If you use an external postprocessor, the MIME encoding for the header
is executed before postprocessing. This means that the postprocessor
must encode anything beyond US-ASCII according to RFC 2047.

If I understand correctly, this applies even when I *do not wish to alter the Subject. It would explain everything.

Quoted from the flnews manual page:

| The header is already in RFC 2047 format. If the postprocessor
| modifies an existing header field, this format must be preserved.
| If a new header field is added, the postprocessor is responsible
| for creating valid RFC 2047 encoding and Unicode normalization to NFC.

Hm. The problem arouse in an initial post, not in a follow-up and the subject is set in the flnews-internal editor. The post-processor does not touch the Subject.

This was most likely the reason that it has not worked as expected.

Okay, I have to look at my Ruby-code, although I scrutinize the existing headers only in the attempt to find the empty line, below.

I think it should be avoided to modify the "Subject" header field
in general. Exceptions are:
- Repairing broken encoding
Every article should be RFC 5536 conformant.
- Explicit switching with the form "New subject (was: Old subject)"

Okay, I have to verify all this, be it only to master my own crap a little better.

For newsgroups with English language, I would suggest to at least limit
such modifications to articles that already use MIME encoding.
In other words: If the article was plain US-ASCII, this should still
be true after postprocessing.

A complete description for a test-case. ;)

Otherwise you may create problems for users of old newsreaders with
no or broken MIME support (that are still used for newsgroups with
English language by many people).

I am aware of that but tend to become more and more fatalistic. Mime encoding shall however be my occupation for a while.

TY again.

Cheerio.
--
He's an ill cook that breeds contempt

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Supersedes for presentation.
Supersedes for orthopaed .. stuff (no content altered)

Mahlzeit

and thank you for the clarifications.

Michael Bäuerle wrote in news.software.readers:

Anyway, FLNews appears to have difficulties, when I use these special characters in the subject header. I do currently not see any need for other symbols than those occasional quote marks.

If you use an external postprocessor, the MIME encoding for the header
is executed before postprocessing. This means that the postprocessor
must encode anything beyond US-ASCII according to RFC 2047.

If I understand correctly, this applies even when I *do not wish to alter the Subject*. It would explain everything.

Quoted from the flnews manual page:

| The header is already in RFC 2047 format. If the postprocessor
| modifies an existing header field, this format must be preserved.
| If a new header field is added, the postprocessor is responsible
| for creating valid RFC 2047 encoding and Unicode normalization to NFC.

Hm. The problem arose in an initial post, not in a follow-up and the subject
is set in the flnews-internal editor. The post-processor does not touch the Subject.

This was most likely the reason that it has not worked as expected.

Okay, I have to look at my Ruby-code, although I scrutinize the existing headers only in the attempt to find the empty line, below.

I think it should be avoided to modify the "Subject" header field
in general. Exceptions are:
- Repairing broken encoding
Every article should be RFC 5536 conformant.
- Explicit switching with the form "New subject (was: Old subject)"

Okay, I have to verify all this, be it only to master my own crap a little better.

For newsgroups with English language, I would suggest to at least limit
such modifications to articles that already use MIME encoding.
In other words: If the article was plain US-ASCII, this should still
be true after postprocessing.

A complete description for a test-case. ;)

Otherwise you may create problems for users of old newsreaders with
no or broken MIME support (that are still used for newsgroups with
English language by many people).

I am aware of that but tend to become more and more fatalistic. Mime encoding shall however be my occupation for a while.

TY again.

Cheerio.
--
He's an ill cook that breeds contempt

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Michael Uplawski <michael.uplawski@uplawski.eu> wrote:

Supersedes because I guessed wrong.

Good morning

For some time, I use to replace the usual quotation marks ' " ' by those
that are expected in the current language, like ' « » ' for the French >language or ' „” ' for German. I sometimes mix up things in English,
but ' “ ” ' should be okay... I guess.

If you write in English, use ASCII ' and ". There's no reason to use typographical quote marks, given how many Mail clients and newsreaders mishandle them.

Anyway, FLNews appears to have difficulties, when I use these special >characters in the subject header.

Can you be very specific about the difficulty?

You MUST encode non-ASCII characters on Subject. MUST

If you enter non-ASCII characters onto Subject, then your Mail client or newsreader is supposed to encode them using the encoded word technique described in RFC 2047. If the precursor message or article has encoded
word on Subject to which you reply or follow up, the client is supposed to decode for the composer, then re-encode prior to sending the reply or
the followup.

This online tool decodes encoded word.

https://dogmamix.com/MimeHeadersDecoder/

If the only encoding is for typographical quotes in English with ASCII substitutes ' and ", then after decoding, use this tool to perform the
ASCII substitution.

https://dan.hersam.com/tools/smart-quotes.php

Note that it performs the substitution for punctuation only, not for
combining characters like vowels with accents or characters with
diacritical marks generally.

There are modules you can download that perform the encoding or
re-encoding, but I've never spotted a tool on line that will take a
string of characters, some of which are non-ASCII, then perform the
encoding so you may paste the resulting encoded text onto Subject.

. . .

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Good morning

Thanks to Adam and Michael again for their input.
I found out that the behavior of flnews is correct and only implicitly leads to the difficulties that I encountered. Persuing the thread would probably render it off-topic. I resume however:

1.) My post-processor (and me) was not at all prepared for subjects that have to be wrapped by the news client. This fact alone is probably sufficient to render articles occasionally invalid and to prevent them from being posted.

2.) When the wrapping happens with an altered subject, the behavior of the post-processor becomes even more interesting, as it recognizes a new header in the value between the line-break and the localized representation of “(Was: ”.

I can intervene and am “correcting” the post processor. But I am more and more
in doubt about the pertinence of the endeavor. Hence my other thread, where I have asked for examples by other peoples. I may be developing stuff but I am not a developper these days.

Thank you anyway

Cheerio
p.s. no post-processing with this post.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Supersedes for bad English

Good morning

Thanks to Adam and Michael again for their input.
I found out that the behavior of flnews is correct and only implicitly leads to the difficulties that I encountered. Persuing the thread would probably render it off-topic. I resume however:

1.) My post-processor (and me) was not at all prepared for subjects that have to be wrapped by the news client. This fact alone is probably sufficient to render articles occasionally invalid and to prevent them from being posted.

2.) When the wrapping happens with an altered subject, the behavior of the post-processor becomes even more interesting, as it recognizes a new header in the value between the line-break and the localized representation of “(Was: ”.

I can intervene and am “correcting” the post processor, but am more and more
in doubt about the pertinence of the endeavor. Hence the other thread, where I have asked for examples by other people. I may be developing stuff but I am
not a developper these days.

Thank you anyway

Cheerio
p.s. no post-processing with this post.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)