Hello All!

DOS, OS/2, Win32, Win64 versions uploaded to: http://sf.net/projects/golded-plus/files/Golded%2B%201.1.5%20snapshots/ http://download.golded.org

changelog:
x64 fixes

* Originally in RU.GOLDED
* Crossposted in GOLDED

WBR, Max.
--- GoldED+/W32-MSVC 1.1.5-b20161221
* Origin: Uptime 561 day(s) 6:18:59.651 (2:5057/19)

Hi Max,

On 21 Dec 16 14:26, Max Vasilyev wrote to All:
about: "GoldED+ snapshot 2016.12.21":

DOS, OS/2, Win32, Win64 versions uploaded to: http://sf.net/projects/golded-plus/files/Golded%2B%201.1.5%20snapshots
/
http://download.golded.org

changelog:
x64 fixes

This message is written with: gpv64-115-20161221.zip

So it works here... Thanks!


Wilfred.

--- FMail-W32 1.73.10.56-B20161219
* Origin: point@work (2:280/464.112)

Hello Max,

On 21 Dec 16 14:26, Max Vasilyev wrote to All:

Hello All!

Just a heads up, Golded+ compiled successfully on ArchLinuxARM 32bit using gcc 6.2.1. However, quite a few times during the compile I thought it would fail due to massive amounts of warnings regarding:

"warning: ISO C++ forbids converting a string contant to `char*' [-Wpedantic]"

..in multiple files during most of the compile. I hope that's enough information for anyone that wants to look into it.

Regards,
Nick

... "?? ????. ? ????? ?????? ???????."
--- GoldED+/LNX 1.1.5-b20161221
* Origin: thePharcyde_ distribution system (Wisconsin) (1:154/10)

Hi Nicholas!

21 Dec 2016 20:10, from Nicholas Boel -> Max Vasilyev:

Just a heads up, Golded+ compiled successfully on ArchLinuxARM 32bit
using gcc 6.2.1.

Heads up also on Debian x86_64.

CU, Ricsi

--- GoldED+/LNX
* Origin: Art for art's sake is a philosophy of the well-fed (2:310/31)

Hello Richard,

On 22 Dec 16 07:20, Richard Menedetter wrote to Nicholas Boel:

Just a heads up, Golded+ compiled successfully on ArchLinuxARM
32bit using gcc 6.2.1.

Heads up also on Debian x86_64.

Did you see all the errors I mentioned, also?

I would imagine Debian is still using gcc 4.9.x? If so, then I doubt it's a old
vs. new gcc warning (or something changed in newer version(s) anyway).

Regards,
Nick

... "?? ????. ? ????? ?????? ???????."
--- GoldED+/LNX 1.1.5-b20161221
* Origin: thePharcyde_ distribution system (Wisconsin) (1:154/10)

On 2016 Dec 22 08:18:22, you wrote to Richard Menedetter:

Just a heads up, Golded+ compiled successfully on ArchLinuxARM 32bit
using gcc 6.2.1.

Heads up also on Debian x86_64.

Did you see all the errors I mentioned, also?

warnings are not errors... that's why the compile continued on...


)\/(ark

Always Mount a Scratch Monkey
Do you manage your own servers? If you are not running an IDS/IPS yer doin' it wrong...
... Fight against violence!
---
* Origin: (1:3634/12.73)

Hello mark,

On Thu Dec 22 2016 18:22:56, mark lewis wrote to Nicholas Boel:

Did you see all the errors I mentioned, also?

warnings are not errors... that's why the compile continued on...

I miswrote that. I meant to say warnings. Either way, whatever was done recently added a huge amount of those warnings. And warnings, like errors, can be fixed.

Regards,
Nick

... "?? ????. ? ????? ?????? ???????."
--- GoldED+/LNX 1.1.5-b20161221
* Origin: thePharcyde_ distribution system (Wisconsin) (1:154/10)

changelog:
x64 fixes

This message is written with: gpv64-115-20161221.zip

This message is written with: gpv64-115-20161221.zip into the hudson base. :)

'Tommi

--- GEcho/32 1.20/Pro
* Origin: GEcho win32 test point (2:221/6.2)

Hi,

On 2016-12-23 08:37:00, Tommi Koivula wrote to Wilfred van Velzen:
about: "GoldED+ snapshot 2016.12.21":

This message is written with: gpv64-115-20161221.zip

This message is written with: gpv64-115-20161221.zip into the hudson base. :)

That's probably why your seconds are :00 on this message... ;)

Bye, Wilfred.


--- FMail-W32 1.73.10.56-B20161219
* Origin: Native IPv6 connectable node (2:280/464)

This message is written with: gpv64-115-20161221.zip

This message is written with: gpv64-115-20161221.zip into the hudson
base.
:)

That's probably why your seconds are :00 on this message... ;)

Yes indeed! All messages here have have seconds :00 :)

Hexdump of Hudson message header and text +----------------------------------------------------

Msgbase : K:\BBS\MSGBASE\HMB\
BoardNo : 11
From : Wilfred van Velzen
To : Tommi Koivula
Subject : GoldED+ snapshot 2016.12.21
DateTime : 12-23-16 10:45

'Tommi

--- GEcho/32 1.20/Pro
* Origin: GEcho win32 test point (2:221/6.2)

Hi,

On 2016-12-23 13:51:00, Tommi Koivula wrote to Wilfred van Velzen:
about: "GoldED+ snapshot 2016.12.21":

That's probably why your seconds are :00 on this message... ;)

Yes indeed! All messages here have have seconds :00 :)

Hexdump of Hudson message header and text +----------------------------------------------------

Msgbase : K:\BBS\MSGBASE\HMB\
BoardNo : 11
From : Wilfred van Velzen
To : Tommi Koivula
Subject : GoldED+ snapshot 2016.12.21
DateTime : 12-23-16 10:45

The original hudson message base doesn't store seconds. It only stores the date
time as shown above in ascii format. There is a later extension that stores the
date/time as time_t 32 bits integer, but that isn't used everywhere...

Btw: I'm still looking for good documentation for the hudson message base format, if anyone's got it?

Bye, Wilfred.


--- FMail-W32 1.73.10.56-B20161219
* Origin: Native IPv6 connectable node (2:280/464)

23 Dec 16 13:19:10, you wrote to me:

On 2016-12-23 13:51:00, Tommi Koivula wrote to Wilfred van Velzen:
about: "GoldED+ snapshot 2016.12.21":

That's probably why your seconds are :00 on this message... ;)

Yes indeed! All messages here have have seconds :00 :)

Hexdump of Hudson message header and text
+----------------------------------------------------

Msgbase : K:\BBS\MSGBASE\HMB\
BoardNo : 11
From : Wilfred van Velzen
To : Tommi Koivula
Subject : GoldED+ snapshot 2016.12.21
DateTime : 12-23-16 10:45

The original hudson message base doesn't store seconds. It only stores the date time as shown above in ascii format. There is a later extension that stores the date/time as time_t 32 bits integer, but that isn't used everywhere...

Hmm.. Have to check how Concord does that...

'Tommi

---
* Origin: 2001:470:1f15:cb0:f1d0:2:221:6 (2:221/6)

On 2016 Dec 22 23:31:14, you wrote to me:

Did you see all the errors I mentioned, also?

warnings are not errors... that's why the compile continued on...

I miswrote that. I meant to say warnings. Either way, whatever was done recently added a huge amount of those warnings. And warnings, like errors, can be fixed.

sure... turn off the setting enabling that warning and you won't see them any more even though nothing in the code has changed...

they're just warnings and that particular one was a pedantic warning talking about ISO C which is a different dialect than ANSI C which is different than C95 which is different than C99 and so on and on and on... that pedantic warning may not even apply if the code is written as ANSI C or C11 or others...

do a search for "c programming dialects" to see a full list of all the various ones there are...

)\/(ark

Always Mount a Scratch Monkey
Do you manage your own servers? If you are not running an IDS/IPS yer doin' it wrong...
... Old typesetters never die. They just decompose.
---
* Origin: (1:3634/12.73)

* Originally from Tommi Koivula to Wilfred van Velzen :

The original hudson message base doesn't store seconds. It

only stores the

date time as shown above in ascii format. There is a later

extension that

stores the date/time as time_t 32 bits integer, but that isn't

used

everywhere...

Hmm.. Have to check how Concord does that...

Nope, No seconds.. ;(

'Tommi

--- Concord/2 001B30d (Reg.)
* Origin: - rbb - telnet://rbb.homeip.net - (2:221/360)

Hello mark,

On Fri Dec 23 2016 07:10:40, mark lewis wrote to Nicholas Boel:

I miswrote that. I meant to say warnings. Either way, whatever
was done recently added a huge amount of those warnings. And
warnings, like errors, can be fixed.

sure... turn off the setting enabling that warning and you won't see
them any more even though nothing in the code has changed...

Why? The point was that the last time I compiled, those warnings weren't there.
Since there were some recent updates to the code, I decided to compile again and report the massive amounts of warnings, in case someone wants to fix them.

I did also say it finished compiling, so if it's not that important, so be it. I was only trying to help in showing what I witnessed.

do a search for "c programming dialects" to see a full list of all the various ones there are...

Uhh, thanks for the info? :/

Regards,
Nick

... "?? ????. ? ????? ?????? ???????."
--- GoldED+/LNX 1.1.5-b20161221
* Origin: thePharcyde_ distribution system (Wisconsin) (1:154/10)

Hello Nicholas,

On Friday December 23 2016 07:53, you wrote:

Why? The point was that the last time I compiled, those warnings
weren't there. Since there were some recent updates to the code, I
decided to compile again and report the massive amounts of warnings,
in case someone wants to fix them.

My experience with C is that - depenndig on compiler - it is unwise to just ignore compiler warnings. Especially with the Borland compiler they should be taken seriously, they may point to a serious problem.

Just my ?0.02,.


Cheers, Michiel

--- GoldED+/W32-MSVC 1.1.5-b20130111
* Origin: Blof T?non (2:280/5555)

Hi Nicholas!

22 Dec 2016 08:18, from Nicholas Boel -> Richard Menedetter:

Just a heads up, Golded+ compiled successfully on ArchLinuxARM
32bit using gcc 6.2.1.

Heads up also on Debian x86_64.

Did you see all the errors I mentioned, also?

Did not watch the build process ... so I do not know.

I would imagine Debian is still using gcc 4.9.x?

Correct
root@vserv:~# gcc --version
gcc (Debian 4.9.2-10) 4.9.2

CU, Ricsi

--- GoldED+/LNX
* Origin: Today is a good day to bribe a high-ranking official (2:310/31)

Hello Michiel,

On Fri Dec 23 2016 15:46:00, Michiel van der Vlist wrote to Nicholas Boel:

Why? The point was that the last time I compiled, those warnings
weren't there. Since there were some recent updates to the code,
I decided to compile again and report the massive amounts of
warnings, in case someone wants to fix them.

My experience with C is that - depenndig on compiler - it is unwise to just ignore compiler warnings. Especially with the Borland compiler
they should be taken seriously, they may point to a serious problem.

Just my ?0.02,.

I agree. If they're not there before the code was modified, and then they're there (in abundance, mind you) afterwards, I would think it's better to let someone know it's happening at the very least - and not just ignore them because it finished compiling. *shrug*

Regards,
Nick

... "?? ????. ? ????? ?????? ???????."
--- GoldED+/LNX 1.1.5-b20161221
* Origin: thePharcyde_ distribution system (Wisconsin) (1:154/10)

Hello Richard,

On Fri Dec 23 2016 16:06:56, Richard Menedetter wrote to Nicholas Boel:

Did not watch the build process ... so I do not know.

Okay, no problem. I'm fairly certain I used the same version of gcc I'm using now to compile Golded when I moved to the RasPi3. So not having the warnings last time, and this time seeing hundreds of them if not more - I may as well let someone know about it.

I would imagine Debian is still using gcc 4.9.x?

Correct
root@vserv:~# gcc --version
gcc (Debian 4.9.2-10) 4.9.2

That's what I thought. Thank you for confirming.

Regards,
Nick

... "?? ????. ? ????? ?????? ???????."
--- GoldED+/LNX 1.1.5-b20161221
* Origin: thePharcyde_ distribution system (Wisconsin) (1:154/10)

Hi,

On 2016-12-23 07:53:52, Nicholas Boel wrote to mark lewis:
about: "GoldED+ snapshot 2016.12.21":

Why? The point was that the last time I compiled, those warnings
weren't there. Since there were some recent updates to the code, I
decided to compile again and report the massive amounts of warnings,
in case someone wants to fix them.

Did your gcc version change since the last time? Because not that much has changed in the golded code in the last couple of years?

Btw: I noticed them too, when I compiled the golded code this week. But there were just too many, otherwise I might have taken a look at the source code that
caused them. ;)

Bye, Wilfred.


--- FMail-W32 1.73.10.56-B20161219
* Origin: Native IPv6 connectable node (2:280/464)

On 2016 Dec 23 13:19:10, you wrote to Tommi Koivula:

Btw: I'm still looking for good documentation for the hudson message
base format, if anyone's got it?

i'm not aware of any documentation on the HMB at all... the only thing i have that even comes close is source code... some would say that that is documentation enough ;)

)\/(ark

Always Mount a Scratch Monkey
Do you manage your own servers? If you are not running an IDS/IPS yer doin' it wrong...
... Cats have the simplest of taste - the best will suffice
---
* Origin: (1:3634/12.73)

On 2016 Dec 23 07:53:52, you wrote to me:

I miswrote that. I meant to say warnings. Either way, whatever was
done recently added a huge amount of those warnings. And warnings,
like errors, can be fixed.

sure... turn off the setting enabling that warning and you won't see
them any more even though nothing in the code has changed...

Why? The point was that the last time I compiled, those warnings
weren't there.

what i'm saying is that the make file instructions may have been updated to turn on all those warnings when they may have been turned off previously...

Since there were some recent updates to the code, I decided to compile again and report the massive amounts of warnings, in case someone
wants to fix them.

that's fine... i have no problem with that... i do the same with a simulator that i'm involved with...

I did also say it finished compiling, so if it's not that important,
so be it. I was only trying to help in showing what I witnessed.

you're fine, man... its all good :)

do a search for "c programming dialects" to see a full list of all
the various ones there are...

Uhh, thanks for the info? :/

there's a lot of them... i was only pointing out that that particular warning was saying that ISO C doesn't like that but other dialects may love it...

)\/(ark

Always Mount a Scratch Monkey
Do you manage your own servers? If you are not running an IDS/IPS yer doin' it wrong...
... That explains it. I'm Canadian and you are not.
---
* Origin: (1:3634/12.73)

On 2016 Dec 23 16:06:56, you wrote to Nicholas Boel:

I would imagine Debian is still using gcc 4.9.x?

Correct
root@vserv:~# gcc --version
gcc (Debian 4.9.2-10) 4.9.2

$ cat /etc/issue; gcc --version
Ubuntu 14.04.5 LTS

gcc (Ubuntu 4.8.4-2ubuntu1~14.04.3) 4.8.4
Copyright (C) 2013 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

i have two cmake's, though... the default installed one...

$ cmake --version
cmake version 2.8.12.2

and a newer one required to build the simulator i'm involved with...

~/flightgear-dev/next/install/cmake/bin/cmake
cmake version 3.7.20161029-gcc998

CMake suite maintained and supported by Kitware (kitware.com/cmake).


)\/(ark

Always Mount a Scratch Monkey
Do you manage your own servers? If you are not running an IDS/IPS yer doin' it wrong...
... Here comes Sanity, here comes Sanity, right down therapy lane.
---
* Origin: (1:3634/12.73)

Hi,

On 2016-12-23 14:45:00, mark lewis wrote to Wilfred van Velzen:
about: "GoldED+ snapshot 2016.12.21":

Btw: I'm still looking for good documentation for the hudson message
base format, if anyone's got it?

i'm not aware of any documentation on the HMB at all...

I've searched the net, but all I could find is the MBLIB10.RAR file. It has very little documentation, and it only knows about the old hudson mb version, not the later extensions.

the only thing i have that even comes close is source code...

That's what I use now, but it takes some "studying"...

some would say that that is documentation enough ;)

Up to a point it is. The structures are very clear. But the details of what goes into the fields, and the relations between them, is sometimes a bit obscure...

Bye, Wilfred.


--- FMail-W32 1.73.10.56-B20161219
* Origin: Native IPv6 connectable node (2:280/464)

Hello Nicholas!

23 Dec 16 09:30, you wrote to Richard Menedetter:

Okay, no problem. I'm fairly certain I used the same version of gcc
I'm using now to compile Golded when I moved to the RasPi3. So not
having the warnings last time, and this time seeing hundreds of them
if not more - I may as well let someone know about it.

I compiled this version of golded (in tearline) using your instructions
and on my ripi3 i had hundreds of warnings. I didn't worry about them as
in my experience c programs always complain about something when compiling.
I do agree with you that they should be reported though just so someone knows what's going on.

Shawn

... The House of Lords has a value.it is good evidence of life after death.
--- GoldED+/LNX 1.1.5-b20160322
* Origin: Tiny's BBS - www.tinysbbs.com (1:229/452.3)

Hello Wilfred,

On Fri Dec 23 2016 18:46:56, Wilfred van Velzen wrote to Nicholas Boel:

Why? The point was that the last time I compiled, those warnings
weren't there. Since there were some recent updates to the code,
I decided to compile again and report the massive amounts of
warnings, in case someone wants to fix them.

Did your gcc version change since the last time? Because not that much
has changed in the golded code in the last couple of years?

I don't believe it had changed. The last time I believe I compiled was in November, and now just recently to get the latest HMB updates you sent patches for. So maybe it had something to do with that?

Btw: I noticed them too, when I compiled the golded code this week.

Using what version of gcc? At least it's nice to know it's not specific to my system. But the possibility of it being both Golded and/or gcc are still there.

But there were just too many, otherwise I might have taken a look at
the source code that caused them. ;)

That's why I said something. There were so many I actually figured it would eventually fail - and was quite surprised that it didn't.

Regards,
Nick

... "?? ????. ? ????? ?????? ???????."
--- GoldED+/LNX 1.1.5-b20161221
* Origin: thePharcyde_ distribution system (Wisconsin) (1:154/10)

Hello mark,

On Fri Dec 23 2016 15:02:54, mark lewis wrote to Nicholas Boel:

Why? The point was that the last time I compiled, those warnings
weren't there.

what i'm saying is that the make file instructions may have been
updated to turn on all those warnings when they may have been turned
off previously...

Anything is possible. However this was a crazy amount of warnings - probably more than I have ever seen when compiling Golded - all referencing the exact same warning. So in case it was something that was changed recently, I tried to
let someone know about it. Looks like Wilfred saw the same this week.

do a search for "c programming dialects" to see a full list of
all the various ones there are...

Uhh, thanks for the info? :/

there's a lot of them... i was only pointing out that that particular warning was saying that ISO C doesn't like that but other dialects may love it...

That's why I was asking about gcc as well. Recently we had to specify the dialect (std=gnu++98) in Synchronet's JS engine makefile as it was failing with
the newer gcc 6.1.2, which changed to yet another dialect.

Regards,
Nick

... "?? ????. ? ????? ?????? ???????."
--- GoldED+/LNX 1.1.5-b20161221
* Origin: thePharcyde_ distribution system (Wisconsin) (1:154/10)

Hello Shawn,

On Fri Dec 23 2016 17:27:02, Shawn Highfield wrote to Nicholas Boel:

I compiled this version of golded (in tearline) using your
instructions and on my ripi3 i had hundreds of warnings. I didn't
worry about them as in my experience c programs always complain about something when compiling. I do agree with you that they should be
reported though just so someone knows what's going on.

That may not be related to what I experienced this week. This one specific warning was repeated hundreds of times, sometimes 30-40 lines one after the other, even. Just not something normally seen when compiling Golded, is all.

This only started happening with the latest code for me, and I was using a version later than yours (November I think?) prior to this one. But yeah, definitely thought it was worth letting someone know about it. :)

Regards,
Nick

... "?? ????. ? ????? ?????? ???????."
--- GoldED+/LNX 1.1.5-b20161221
* Origin: thePharcyde_ distribution system (Wisconsin) (1:154/10)

Hi,

On 2016-12-23 17:39:38, Nicholas Boel wrote to Wilfred van Velzen:
about: "GoldED+ snapshot 2016.12.21":

Did your gcc version change since the last time? Because not that
much has changed in the golded code in the last couple of years?

I don't believe it had changed. The last time I believe I compiled was in November, and now just recently to get the latest HMB updates you sent patches for. So maybe it had something to do with that?

Some structure members types changed, so that could trigger some new warnings. But going through the warnings I don't notice any of those at first glance. The
bulk of them are the:

warning: deprecated conversion from string constant to `char*' [-Wwrite-strings]

And those have nothing to do with my changes.

Btw: I had to deal with those too, when I recently switcht from the Borland compiler to Mingw (gcc 4.9.2) for FMail... ;)

An other possibility would be the compiler options in the makefile(s) changed. But going through the version history there weren't any changes of importance for at least the last year or so...

Btw: I noticed them too, when I compiled the golded code this week.

Using what version of gcc? At least it's nice to know it's not specific to my system. But the possibility of it being both Golded and/or gcc are

still

there.

gcc --version

gcc (SUSE Linux) 4.8.5


Bye, Wilfred.


--- FMail-W32 1.73.10.56-B20161219
* Origin: Native IPv6 connectable node (2:280/464)

Hello Wilfred,

On Sat, 24 Dec 2016 13:39:24 +0100, Wilfred van Velzen -> Nicholas Boel wrote:

Some structure members types changed, so that could trigger some new warnings. But going through the warnings I don't notice any of those at first glance. The bulk of them are the:

warning: deprecated conversion from string constant to `char*' [-Wwrite-strings]

I've always seen those, I'm pretty sure. The new ones I mentioned about the ISO
C++ warnings was what I was referring to. So maybe you didn't see those (you definitely would have, as it would cough up about 30 in a row at times)? If not, it very well could be an issue with the latest version(s) of gcc and it switching from -std=gnu++98 to -std=gnu++14 mode introduced in gcc 6.1, although I don't believe my version has changed since the last time I compiled Golded.

An other possibility would be the compiler options in the makefile(s) changed. But going through the version history there weren't any changes of importance for at least the last year or so...

Then it isn't that.

gcc --version

gcc (SUSE Linux) 4.8.5

Does SUSE stay further behind Debian? Or haven't you upgraded in awhile?

Regards,
Nick

--- slrn/1.0.3 (Linux)
* Origin: thePharcyde_ distribution system (1:154/10)

Hi,

On 2016-12-24 09:42:58, Nicholas Boel wrote to Wilfred van Velzen:
about: "GoldED+ snapshot 2016.12.21":

warning: deprecated conversion from string constant to `char*'
[-Wwrite-strings]

I've always seen those, I'm pretty sure. The new ones I mentioned about

the

ISO C++ warnings was what I was referring to. So maybe you didn't see

those

(you definitely would have, as it would cough up about 30 in a row at times)?

Which was it exactly, there were so many, I didn't pay much attention to the details. ;)

But the above one was the most common, after that probably:

warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing]

Which happend in several files. But I don't see warnings containing "ISO".

If not, it very well could be an issue with the latest version(s) of
gcc and it switching from -std=gnu++98 to -std=gnu++14 mode introduced
in gcc 6.1, although I don't believe my version has changed since the
last time I compiled Golded.

Possibly...

gcc --version

gcc (SUSE Linux) 4.8.5

Does SUSE stay further behind Debian? Or haven't you upgraded in awhile?

I'm one version behind. 42.2 Came out mid november.

On both the default is still gcc 4.8.

6.2.1 and other newer versions, are availabel for both, but have to be specifically installed...

Bye, Wilfred.


--- FMail-W32 1.73.10.56-B20161219
* Origin: Native IPv6 connectable node (2:280/464)

Hello Wilfred,

Btw: I'm still looking for good documentation for the hudson message
base format, if anyone's got it?

Me too, as I am running FD 2.02 NC and RA 1.11 for very many years at UniCorn BBS in Arnhem.nl with a Hudson msg-base.
I only know very well that it must never exceed the 16 MB length,
otherwise its msgbase and index are both corrupted.
So I invested many hours to automatically reduce its volume below that 16 MB length mark by using a good daily management.
Now my Point system is my archive for older messages.
If you found some information about the Hudson msg-base structure,
please tell me too.

Henri.

---
* Origin: Connectivity is the Future; UniCorn BBS 31 26 4425506 (2:280/1208)

Hello mark,

Btw: I'm still looking for good documentation for the hudson message
base format, if anyone's got it?

i'm not aware of any documentation on the HMB at all... the only thing i have that even comes close is source code... some would say that that is documentation enough ;)

It is ;-)
For my RISC OS !WimpLink Point system I have figured out how the message base was configured at low level, by simply looking at it with a hex editer
and thwe flag list in the manual.
I did the same to my RISC OS e-mail program !Marcel.
The main reason for this examination was to correct broken messages,
and after that reindex the msg-base to use them again, in stead of deleting. Sometimes I got tears from broken msg-bases ;-(.
The indexing system is more complex to figure out, but I am not a C
programmer.
What one can do is make a test system with an emty msg-base,
write some messages with every possible flag and kludge,
and receive some to it, and then examine the low level contents.
With the source code besides it, it must be possible to get the structure clear for at least 90 % I think.
May be the FTSC documents can help to get it all clear?

It could be something like this:

From AKA
To AKA
msg date in unix time
received or sent / polling date in unix time
From:
To:
Subject
EchoMail Area Tag
8 bytes msg flags
Kludges with ctrl A
unformatted msg text (mostly the biggest part ;-)
tagline
tearline and product name
Origin
SEEN-Bye's
Kludges with Ctrl A
Path aka's
Closing byte Hex 0 (zero)

That last one is important as msg separator.
This list has to be finetuned to create the right msg format.
Hm, Hudson, wasn't that a river near New York found by some Dutchman?
Good luck with it.

Henri.

---
* Origin: Connectivity is the Future; UniCorn BBS 31 26 4425506 (2:280/1208)

Hi Henri,

On 23 Jan 17 00:47, Henri Derksen wrote to Wilfred van Velzen:
about: "GoldED+ snapshot 2016.12.21":

If you found some information about the Hudson msg-base structure,
please tell me too.

I didn't find any extra info. But the structures inside the individual files are pretty well known. The FMail source already has a pretty good library to access the files. And there is the MBLIB10.ZIP (easy to find using google), that contains the structures as a .h C include file.

What I was looking for was some more indepth documentation by the original author(s), where they explain the design decisions, the rationale behind it, all the itty bitty details, and how it all comes together...

Wilfred.

--- FMail-W32 1.73.12.68-B20170121
* Origin: point@work (2:280/464.112)