Messaggio archiviato #289

Da: "Lewis G Rosenthal" <> Intestazioni complete
Messaggio non codificato
Oggetto: Re: [eCS-ISP] links in the ticket
Data: Sun, 28 Mar 2021 22:54:54 -0500
A: eCS ISP Mailing List <>

Hi, again...

(Reply 1 of 2):

On 03/28/21 08:59 pm, Steven Levine wrote:
In <>, on 03/28/21
    at 09:52 PM, "Lewis G Rosenthal" <> said:

Hi there,

attaching something would be  prudent. (Of course, if anyone thinks this
list is "ridiculous," he is  welcome to unsubscribe instead of reporting
a problem; that's fine, too.  This list is free - as in beer *and* as in
My thoughts exactly, but I preferred to let the moderator state this.  If
this were a SCOUG list, I would said much the same thing. :-)


One would think, unless there was a remnant of a YUM/RPM base on the
system  and we didn't recognize it as being in need of a bootstrap.
I was trying to guess how Massimo got himself in to this state and now I
may have an idea how he fooled ANPM into not installing the bootstrap.

You probably skipped some of the more fun parts of this thread and others,

Yes, I was only glancing at messages here and there. The "ridiculous" comment stuck in my head, though, prompting my entry.

Massimo started out with an eCS 2.2b system and was manually copying items
from the \usr tree on his ArcaOS install to the \usr tree on the eCS 2.2b

This is rarely a good practice. In fact, unless manually copying to \usr\local - and keeping clear and contemporaneous notes as to what has been copied (and why) - sullying \usr is just a bad idea. This is a *managed* tree, as we know.

   After having a sufficient number of problems attempting to this
without sufficient understanding of how to ensure the prerequisites are
also installed, he decided to install ANPM.  For reasons known only to
Massimo, he decided to install 1.04, although I am not sure if 1.06 would
have avoided any of his subsequent issues.  He did finally install 1.06,
but I can't say what, if any, of his specific problems this fixed.

I doubt that 1.0.6 would have helped in such a situation.

He probably had manually installed rpm and yum to his \usr tree and that's
probably suffcient to fool ANPM into thinking it did not need to install
the bootstrap.  I was wondering how he managed to end up with platform
i386 and this would explain it.  Without the bootstrap there will
initially be no platform file and ANPM will default to i386.

Well, to be precise, with no platform file, yum will default to i386, and ANPM will just call yum to do its thing. However, ANPM 1.0.6 does make it relatively easy (point & click) to set a reasonable base platform. I don't recall offhand whether we introduced that with 1.0.4 or something later. Of course, one would have to know that the platform file was missing and that i386 was non-optimal, having been essentially orphaned in 2017 or so, in favor of i686.

Of course, I am doing some educated guessing based on Massimo's inexact
statements of what failed when he attempted certain actions.  We will
probably never know for sure.


FWIW, the bootstrap required check could be a bit more robust.  The
current code only checks for the existance of rpm.exe and yum.exe, but the
test will pass even if the executables cannot run.

More on this in my next reply.

Lewis G Rosenthal, CNA, CLP, CLE, CWTS, EA
Rosenthal & Rosenthal, LLC      
visit my IT blog      

Isriviti: Feed, Riassunto, Indice.
Scrivi a ListMaster