Hello Евгений,

From your text in Github:

A text is taken for UTF-8 when every byte above 0x7F belongs to a well-formed sequence, there is at least one, and every character so
encoded is one a message might plausibly contain. That last test came
out of measuring the rule against a base of a hundred thousand messages: three lower-case CP866 letters form a well-formed sequence now and then,
and what they decode to is Tai Le.

Interesting. Yes, I can imagine that a well formed sequence test may give false positive. Not just with cp866, but with other alfabets as well. Greek for example. But that there will be no ill formed sequences to negate that false positive seems impossible to me.

"might plausibly contain" seems a bit tricky to me....

Anyway, the first test of your autodetect was OK.

Cheers, Michiel

--- GoldED+/W32-MINGW 1.1.5-b20260904
* Origin: Nieuw Schnøørd (2:280/5555)

Hello Евгений,

04 Sep 26 21:57, I wrote to Евгений Субботин:

Writing long header to base



Cheers, Michiel

--- GoldED+/W32-MINGW 1.1.5-b20260904
* Origin: Nieuw Schnøørd (2:280/5555)

Hello Евгений,

04 Sep 26 21:58, I wrote to you:

Writing FSP-1030...


Cheers, Michiel

--- GoldED+/W32-MINGW 1.1.5-b20260904
* Origin: Nieuw Schnøørd (2:280/5555)

Hello Michiel!

04 Sep 26 22:02, you wrote to Евгений Субботин Т:

@TID: FMail-W32 2.3.0.1-B20240319
@RFC-X-No-Archive: Yes
@TZUTC: 0200
@UCSTO: Евгений Субботин Тест

???

Previous message did not show the complete name at the end, some wierd characters
this message was ok, but saw @UCSTO with 'wierd chars'


Dennis


... We need to schedule a reboot.
--- GoldED+/LNX 1.1.5-b20260904
* Origin: ---- BOFH: Problem solved, user deleted. (2:280/2060)

Hello Michiel!

04 Sep 26 22:02, you wrote to Евгений Субботин Т:

@TID: FMail-W32 2.3.0.1-B20240319
@RFC-X-No-Archive: Yes
@TZUTC: 0200
@UCSTO: Евгений Субботин Тест
@CHRS: UTF-8 4
@MSGID: 2:280/5555 6a9b2427
@REPLY: 2:280/5555 6a9b2354
Hello Евгений,

04 Sep 26 21:58, I wrote to you:

Writing FSP-1030...

Works for me.

Yegor
--- AmberEdit/linux 0.6.6
* Origin: to err is human, but to really f.. things up you need AI (2:382/736)

Hello Dennis!

05 Sep 26 08:46, I wrote to Michiel van der Vlist:

@TID: FMail-W32 2.3.0.1-B20240319
@RFC-X-No-Archive: Yes
@TZUTC: 0200
@UCSTO: Евгений Субботин Тест

???

Previous message did not show the complete name at the end, some wierd characters this message was ok, but saw @UCSTO with 'wierd chars'

ah .. I added 2 new config lines and now UCSTO is working as expected.



Dennis


... Why does the Wi-Fi hate you? I wonder.
--- GoldED+/LNX 1.1.5-b20260904
* Origin: ---- BOFH: Problem solved, user deleted. (2:280/2060)

Hello Michiel!

04 Sep 26 22:02, you wrote to Евгений Субботин Т:

@TID: FMail-W32 2.3.0.1-B20240319
@RFC-X-No-Archive: Yes
@TZUTC: 0200
@UCSTO: Евгений Субботин Тест
@CHRS: UTF-8 4
@MSGID: 2:280/5555 6a9b2427
@REPLY: 2:280/5555 6a9b2354
Hello Евгений,

with the xlatutfautodetect override it worked.


Dennis


... It's not my fault; it's user-driven entropy.
--- GoldED+/LNX 1.1.5-b20260904
* Origin: ---- BOFH: Problem solved, user deleted. (2:280/2060)

Hello Евгений,

04 Sep 26 22:02, I wrote to you:

Writing FSP-1030...

Now testing both writing long header to base plus writing FTS-1030 kludge.


Cheers, Michiel

--- GoldED+/W32-MINGW 1.1.5-b20260904
* Origin: Nieuw Schnøørd (2:280/5555)

Hello Michiel!

05 Sep 26 11:45, you wrote to ╨Х╨▓╨│╨╡╨╜╨╕╨╣ ╨б╤Г╨▒╨▒╨╛╤В╨╕╨╜ ╨в╨:

@TID: FMail-W32 2.3.0.1-B20240319
Hello ╨Х╨▓╨│╨╡╨╜╨╕╨╣,

04 Sep 26 22:02, I wrote to you:

Writing FSP-1030...

Now testing both writing long header to base plus writing FTS-1030
kludge.

No kludges except TID, garbage in "To:" field.

Yegor
--- AmberEdit/linux 0.6.6
* Origin: to err is human, but to really f.. things up you need AI (2:382/736)

Hello Dennis,

On 05 Sep 26 10:36, you wrote to you:

Previous message did not show the complete name at the end, some
wierd characters this message was ok, but saw @UCSTO with 'wierd
chars'

That is to be expected. When the long header is written to the JAM base, the truncation to 35 bytes is done by the tosser. Fmail in this case. Fmail does not know about UTF-8 (yet?) so the truncation can happen in the middle of a well formed UTF-8 sequence. Hence the garbage character at the end of he name.

ah .. I added 2 new config lines and now UCSTO is working as expected.

Hmm, it must be something else. Om my point system it works as expected without any modifications to the config. The UCSTO kludge is picked up by Golded and displayed in the header automatically.


Cheers, Michiel

--- GoldED+/W32-MINGW 1.1.5-b20260904
* Origin: Nieuw Schnøørd (2:280/5555)

Hello Евгений,

05 Sep 26 11:45, I wrote to you:

Now testing both writing long header to base plus writing FTS-1030
kludge.

That does not work as I expected. Apparently it is one or the other. Enabeling long header, disables the USC* kludges.

I have to sleep on this.


Cheers, Michiel

--- GoldED+/W32-MINGW 1.1.5-b20260904
* Origin: Nieuw Schnøørd (2:280/5555)

Hello Michiel!

Saturday September 05 2026 14:06, you wrote to Dennis Slagers:

MvdV> That is to be expected. When the long header is written to the JAM
MvdV> base, the truncation to 35 bytes is done by the tosser. Fmail in this
MvdV> case. Fmail does not know about UTF-8 (yet?) so the truncation can
MvdV> happen in the middle of a well formed UTF-8 sequence. Hence the
MvdV> garbage character at the end of he name.

I’ve had a lot of problems with FMail. First, it doesn’t support BSD systems for building (this is easy to fix, but requires testing). Second, there’s no support for Amiga Style Outbound. Third, there’s no option to import from the husky fidoconfig/hpt config file. And yes, it also can’t handle character encodings, whereas hpt already has some code for that.

Otherwise, FMail was my first and main tosser for a long time in the late ’90s and early 2000s, and overall, I’ve always liked it.

Eugene

... It's full of stars!
--- GoldED+/BSD 1.1.5-b20260905 (NetBSD 11.0 Intel Core Haswell)
* Origin: FireFox Station (2:5075/35)

Hello Eugene,

On 06 Sep 26 11:46, you wrote to me:

I’ve had a lot of problems with FMail.

Othe than most Fidonet software Fmail is till under development. Just contact Wilfred.

First, it doesn’t support BSD systems for building (this is easy to
fix, but requires testing).

The linux version is relatively new. Apparently there has not been much demand yet among the Fmail users or else someone would have picked that up...

AFAIK the source is available...

Second, there’s no support for Amiga Style Outbound.

Wilfred is the head of what is left of Amiga net. So once again there does not seem to be a demand for that. Apparently even the head of what is left of the Amiga crowd dows not need it. So why do you need ASO? Does anyone still use it?

Third, there’s no option to import from the husky fidoconfig/hpt
config file.

Never needed that either... ;-)

And yes, it also can’t handle character encodings, whereas hpt already has some code for that.

Of course in the 8 bit age - which regarding Fidonet we have hardly left behind - there never was a need for handling character encodings in a TOSSER. On the contrary a tosser that is fully 8 bit transparent is all you need and want. If it is 8 bit agnostic, it can deal with messages in ANY encoding. Or so I thought.

But now with UTF-8 it goes wrong when long strings need to be truncated. I never realised there was a problem with that, up until now I thought that full 8 bit transparency was enough to deal wit UTF-8. But now I discover that Fmail also truncates when converting from a base with large headers to *.pkt.

I think Wilfred can fix that...

Otherwise, FMail was my first and main tosser for a long time in the
late ’90s and early 2000s, and overall, I’ve always liked it.

In the very beginning I had Gecho. I dropped it in favour of Fmail when Garard suddenly stopped with Gecho. Never regretted that choice.


Cheers, Michiel

--- GoldED+/W32-MINGW 1.1.5-b20260904
* Origin: Nieuw Schnøørd (2:280/5555)

Hi Michiel,

On 2026-09-10 13:04:42, you wrote to Eugene Subbotin:

Second, there’s no support for Amiga Style Outbound.

MvdV> Wilfred is the head of what is left of Amiga net. So once again there does not
MvdV> seem to be a demand for that. Apparently even the head of what is left of the
MvdV> Amiga crowd dows not need it. So why do you need ASO? Does anyone still use
MvdV> it?

I know there are some AmigaNet nodes who use it, but that is on real Amiga hardware (or emulated), with all Amiga software.

Never had a request for this for FMail...

Are there Windows or Linux mailers that only support ASO?


Bye, Wilfred.

--- FMail-lnx64 2.3.4.1-B20260520
* Origin: FMail development HQ (2:280/464)

Hello Michiel!

10 Sep 26 13:04, you wrote to Eugene Subbotin:

Second, there’s no support for Amiga Style Outbound.

Wilfred is the head of what is left of Amiga net. So once again there
does not seem to be a demand for that. Apparently even the head of what
is left of the Amiga crowd dows not need it. So why do you need ASO? Does anyone still use it?

I'm not equipped with a hex2dec calculator in my head, so 2.382.736.0.flo is more convenient for debugging than 017e02e0.flo. It's not about Amiga devices (never had one) - it's about the 8.3 filename format having been dead for decades.

Yegor
--- AmberEdit/linux 0.8.1
* Origin: to err is human, but to really f.. things up you need AI (2:382/736)