• Record length

    From Ward Dossche@2:292/854 to Andrew Leary on Fri Jul 27 01:25:02 2018
    Andrew,

    Will makenl barf if "long record lines" would be used for nodelist segments?

    \%/@rd

    --- D'Bridge 3.98
    * Origin: Many Glacier / Protect - Preserve - Conserve (2:292/854)
  • From Michiel van der Vlist@2:280/5555.6 to Andrew Leary on Fri Jul 27 01:25:02 2018
    Hello Andrew,

    On Tuesday February 11 2014 20:43, you wrote to Ward Dossche:

    Will makenl barf if "long record lines" would be used for
    nodelist segments?

    The code should handle nodelist lines up to 512 bytes in length.

    And gracefully handle longer lines I presume? (Instead of crashing like the old
    MakeNl)

    [ FTSC bureaucrat mode on ]

    From FTS-5000.004:

    For compatibility with certain broken segment processors, nodelist
    maintainers must limit all lines to at most 157 characters (plus
    terminator). It is very likely this limitation will be removed in
    the future, therefore future software implementations should be able
    to process lines of up to 1024 characters, and individual fields of
    up to 255 characters (unless otherwise specified).

    [ off ]

    It is no big deal of course, I do not expect to see lines >512 bytes in the nodelist any time soon. But as you are an FTSC member, you might consider upping the max line length next time you fire up the compiler...

    Just a suggestion.

    Cheers, Michiel

    --- GoldED+/W32-MINGW 1.1.5-b20110320
    * Origin: http://www.vlist.eu (2:280/5555.6)
  • From Michiel van der Vlist@2:280/5555.6 to Kees van Eeten on Fri Jul 27 01:25:02 2018
    Hello Kees,

    On Wednesday February 12 2014 11:49, you wrote to me:

    But do we need longer lines to list technical facillities, or just so
    we can boast all our hats and feathers.

    Apparently we do need lines longer than 157 bytes, as accidents have happened when segments with longer lines were submitted in the past. They crashed the old makenl and complete regions and even an entire zone disappeared.

    This is not the place to discuss the desirability of "vanity" in the nodelist. That is no concern of the Makenl_Ng crew. Here we should discuss the technical issues with regard to makenl. FTSC complience is a technical issue.

    What is actually allowed in the nodelist is for the nodelist clercks to decide.


    Cheers, Michiel

    --- GoldED+/W32-MINGW 1.1.5-b20110320
    * Origin: http://www.vlist.eu (2:280/5555.6)
  • From Ward Dossche@2:292/854 to Kees van Eeten on Fri Jul 27 01:25:02 2018
    Kees,

    But do we need longer lines to list technical facillities, or just so we can boast all our hats and feathers.

    There has already been a non-descript RC of Region-23 I seem to remember who royally bypassed the 157-boundary when deploying all of his feathers at some time.

    \%/@rd

    --- D'Bridge 3.98
    * Origin: Many Glacier / Protect - Preserve - Conserve (2:292/854)
  • From Michiel van der Vlist@2:280/5555.6 to Kees van Eeten on Fri Jul 27 01:25:02 2018
    Hello Kees,

    On Wednesday February 12 2014 13:57, you wrote to me:

    MvdV>> Here we should discuss the technical issues with regard to makenl.
    MvdV>> FTSC complience is a technical issue.

    The technical requirement is born out of documenting the the
    processes in Fidonet.

    The 1024 line limit first showed up in FTS-5000.002 dated Jan 2005. IIRC it was
    a compromise between the "unlimited" faction and those not wanting to saddle the programmers with the burden of "infinite" record length. While larger than "current practise", 1024 was considered "large enough for anybody".

    One of wich is showing off vanity in the nodelist.

    I fought that battle in the past and lost. I don't want to raise the issue again, it would be an exercise in futility. The nodelist is shrinking, memory gets cheaper, processing power is manyfold it used to be. I have no new arguments end whatever arguments I had then, have lost their strength.

    Indeed and they may have more to worry about.

    So let THEM worry about it.

    Just for the record:
    The limit for a nodelist line in the Ifmail package is 256.
    in Golded package, I have found it to be 512.
    For MBSEbbs the nodelist line lenght is set to 1024.

    These packages are however still maintained, so ugrades can
    be made available any time.

    I don't see the above as an argument against raising the max for MakeNl to 1024.


    Cheers, Michiel

    --- GoldED+/W32-MINGW 1.1.5-b20110320
    * Origin: http://www.vlist.eu (2:280/5555.6)
  • From Michiel van der Vlist@2:280/5555.6 to Kees van Eeten on Fri Jul 27 01:25:02 2018
    Hello Kees,

    On Wednesday February 12 2014 19:20, you wrote to me:

    Indeed and they may have more to worry about.
    MvdV>> So let THEM worry about it.

    Dn't worry, they probably do.

    My point was: let's not try to do their job, let us do what we are here for and
    just hand them the tools. In this case the tool of MakeNl.

    A sensible strategy when buying or making a new tool is to see that it can do more than what is needed at the moment. When you need a pump for the tyres of your car and that car needs 250 kPa, you don't buy one that can do just that and no more. You buy one that can do more, in case you need more in the future.

    Let's upgrade Makenl to 1024 lines to comply with FTS-5000. It does not mean that *C's must allow lines that long and if it is never needed, there is no harm done.

    These packages are however still maintained, so ugrades can
    be made available any time.

    MvdV>> I don't see the above as an argument against raising the max
    MvdV>> for MakeNl to 1024.

    No, it is just a start of an inventory, of were changes are needed and where there a limitations, as the software cannot be modified. This is
    not aimed at Makenl, but at the use of longer lines in the nodelist.

    Makes sense, but you are barking up the wrong tree. Here we make the tools. How
    they are used by the *Cs is a different issue that belongs in an area where the
    sysops can meet with the *Cs.

    It is wonderfull if all levels can produce a nodelist with long lines,


    Decoupling Pvt and -Unpublished- was not the only reason for the MakeNl_Ng project. Another and perhaps more important reason at the time was the 157 bytes bug. The old MakeNl CRASHED at lines > 157 bytes. Very Annoying.

    but if it fails on crucial systems on the net,

    That remains to be seen. Where is the spirit of the old days when we just tried
    it and saw what happened?

    where is the gain?

    157 bytes has been demonstated to be too short. The gain in allowing THIS tool to handle longer lines is that at least with THIS tool we will not run into that problem again.

    What is lost?


    Cheers, Michiel

    --- GoldED+/W32-MINGW 1.1.5-b20110320
    * Origin: http://www.vlist.eu (2:280/5555.6)
  • From Ward Dossche@2:292/854 to Michiel van der Vlist on Fri Jul 27 01:25:02 2018
    Michiel,

    157 bytes has been demonstated to be too short.

    I fully concur ...

    \%/@rd

    --- D'Bridge 3.98
    * Origin: Many Glacier / Protect - Preserve - Conserve (2:292/854)