= Сообщение: 7442 из 12555 ====================================== ENET.SYSOP = От : Michiel van der Vlist 2:280/5555 22 Sep 18 12:53:25 Кому : Kees van Eeten 22 Sep 18 12:53:25 Тема : Nlcheck report FGHI : area://ENET.SYSOP?msgid=2:280/5555+5ba62749 На : area://ENET.SYSOP?msgid=2:280/5003.4+5ba4c5fe = Кодировка сообщения определена как: CP850 ================================== ============================================================================== Hello Kees,
On Friday September 21 2018 11:39, you wrote to me:
MvdV>> When "Allowunpub" was introduced, MakeNl should have been MvdV>> equipped with a check for missing IP connect info and force MvdV>> "Pvt" if both the telephone number and IP info are absent. And MvdV>> refuse the entry in case of a Hub, Host, Region or Zone entry MvdV>> without connect info.
KE> You cannot drop an entry with a Hub, Host, Region or Zone entry. If KE> you do the hole segment has to be dropped.
So one has to reject the segment and use the latest correct segment...
KE> I have not checked with makenl, but what does it do without KE> Allowunpub, when a Host line contains -Unpublished- ?
It flags the line as ;E with "invalid phone number" as reason. The output is renamed to *.$$$ and makenl reports a fatal error. Effectively rejecting the segment and causing makenl to use the previously submitted segment.
KE> I agree, that makenl was crippled in it's checks to allow for a KE> changing needs. Crippling the checks in makenl was easy, getting KE> consenus on new limits of what was allowable was not, or not tried. KE> Common sense by those who edit the segments, was expected to be KE> enough.
And that, relying on common sense, unfortunaly turned out to have been a big mistake. :-(
KE> The problem is, that common sense is not common.
The problem is that the checks and balances that kept Fidonet technically sane, are no longer functioning. The FTSC is a paper tiger and the *Cs are making a mess of the nodelist. Your surveys of obviously outdated nodelist information is another indication that the nodelist clercks are failing in keeping the nodelist accurate, up to date and technically correct. :-(
KE> As for what makenl should check and what makenl should not check, be KE> aware that as far as makenl is concerned, a nodelist line ends with KE> the modem speed value. Makenl has no knowlede of flags.
KE> You may claim, that including the flags in tests is peanuts, as you KE> have done it in your util, including such rules in makenl, could be a KE> major effort.
I am not claiming that it would be peanuts. Just that it can be done, as indeed I have done it in my util. I am not familiar with the source code of MakeNl and so I have no idea how hard or easy it would be to integrate that functionality in MakeNl.
KE> And if done, will the main users of makenl then accept such changes to KE> that program.
The same nodelist clercks that knowingly and willingly list nodes without Pvt and without connect info will most likely reject an upated version of MakeNl that rejects those entries...
KE> Another thing would be to make the use of a pre/post filter programm KE> mandatory. History has show, that introducing such a programm fidonet KE> wide is next to impossible, but you are welcome to try.
I have better things to do than waste my valuable time on a mission impossible. I am afraid it is too late to repair the damage. The genie is out of the bottle. :-(
