Subj : Zone 4
To   : Janis Kracht
From : Kees van Eeten
Date : Fri Jul 27 2018 01:25:03

Hello Janis!

22 Feb 13 18:41, you wrote to me:

 >> A short summary: Manuel made a quick addition to his segment, but did
 >> not test it. Apparently he had allowunpub enabled.
 >> ZC1 was either forwarned or has allowunpub enabled by default or
 >> whatever.

 JK> Neither.

 Just a thought, a dirty mind is a joy forever. ;)

 JK> It was not enabled here.  I've had 'MINphone 1' in my ctl for ages in

 Now you mention it, MINphone indeed has a an influence. The test for the
 minimal number of parts sees allowunpub a one part. I think the test should
 be skipped for unpublished, as it nukes the positive use of MINphone.

 JK> preparation of ALLOWunpub, and that is probably why Manuel's entry had
 JK> no problems.  Just a guess.

 The is a claim that tje Linux version behaves differently from the W32 
version.

 I can only test waht a have available and as I run a 64 bit Linux I my only
 option is to compile streight from the source tree where I am usually
 up to date.

 I performed 3 test with the allowubpub option and a near to actual nodelist.
 I had to correct the mistake Manuel made with the double 27 entries,
 otherwise the routine I made to disseminate the nodelist overwrites the
 first 27 with the second, wich was a fully legal line.
 I enables allowunpub on all intermediate levels for thier respective
 segment files, so the line in question in the Z4 segment was not flagged
 as an error at an earlier state.

 The current source code says that '0' is the default value for allowunpub.

 Now for the final composite file.
 1. No allowunpub in the ctl (default value) - An error message was produced
 2. allowunpub 0 in the ctl                  - An error message was produced
 3. allowunpub 1 in the ctl                  - No error message was produced.

 I donot know what software matches the behaviour you describe, but it is
 not in the current code.

 I have a few gripes that are not particularly directed at you.

 After so many years of developement and accommodation to perceive bugs and
 incompattiblities for dropin replacement with the Ben Baker version.
 And that all this was first needed before a switch could be made to makenl_ng.
 That that switch had to be complete before allowunpub could be implemneted.
 Now we have to look for a perceived bug, that is not a bug in the future
 we all say we are eager to implement. Moreover that is a reason not to
 do so until it is solved.

 I cannot speak for the developers, they may have other goals. But without
 an definate agreemnet on when things will be implemented, I can think of
 more usefull things to do with my time.

Kees

--- FPD v2.9.040207  GoldED+/LNX 1.1.5
 * Origin: Carefully open my messages, it's Flue season. (2:280/5003.4)

.