[pkg-ntp-maintainers] Bug#502864: acknowledged by developer

[Peter T. Breuer]

"Also sprach Debian Bug Tracking System:"
> This is an automatic notification regarding your Bug report
> #502864: ntpdate: if-up.d file is missing wait_for_file,
> which was filed against the ntpdate package.
> 
> It has been marked as closed by one of the developers, namely
> kurt at roeckx.be (Kurt Roeckx).
> 
> You should be hearing from them with a substantive response shortly,
> in case you haven't already. If not, please contact them directly.

OK, likely tomorrow.

I guess that people simply waited for udev to become compulsory.

At the time of the bug report, I was NOT using udev, which would
be why

  It is defined in /lib/udev/hotplug.functions, which that file
  includes.

would have cut no ice (see below).  The package in question did not
depend on udev

  Versions of packages ntpdate depends on:
  ii  libc6                    2.7-15          GNU C Library: Shared libraries
  ii  libcap1                  1:1.10-14build1 support for getting/setting POSIX.
  ii  libssl0.9.8              0.9.8g-13       SSL shared libraries
  ii  netbase                  4.34            Basic TCP/IP networking system


I regard the response

 I'm closing this bug as there was no response and the circumstances
 sound implausible.

as continuing to be inappropriate and insulting.  Plainly the function
was called:

 # This is for the case that /usr will be mounted later.
   if [ -r /lib/udev/hotplug.functions ]; then
      . /lib/udev/hotplug.functions
      wait_for_file /usr/sbin/ntpdate-debian
      ^^^

and we can therefore deduce that I HAD a /lib/udev/hotplug.functions
file (since test -r succeeds), but not one that contained the
wait_for_file function.

What I guess is that I had an old udev installed, but not mounted
(or else uninstalled but not purged .. who knows) and nothing from udev
running. The ntpdate file spotted the readable file and assumed it was
for it, but it wasn't a version of the file that it needed.

The real BUG was that ntpdate depended on udev and didn't say so.

The real fix would have been to make it use something not from udev
so that it didn't depend on it. Wait_for_file is not hard to write.
Any programmer would write its five lines quicker than it takes to
breathe.

A maintainer's hack would have been to make ntpdate depend on udev.

This response is correct:

  # Should either depend on udev or have the function itself
  reopen 502864 

As to

   Looking at it again, I don't see how it can fail.  Either udev is

Look again. It's clear. Old udev file present without the function
in, and no dependency declared by ntpdate.

   installed and the function should exist, or udev is not installed
   and the function is not called. 
   
You miss a third alternative - some udev files are there but either
udev is not installed or not up to date, which ntpdate won't care about
in either case since udev is not declared as a dependency.
   
   The reporter also seem to be mixing debian and ubuntu packages.

No, not so. I really wish this excuse would go away! The dependencies
were satisfied by Debian.  Even if it were true, so what!  Making code
work in a hostile environment is EXACTLY what software maintenance is
all about, and is what causes code to evolve to become better.
Control the environment as well as the application and you end up
writing for just yourself - a la MS. One always wants to make ones
application as accepting of different environments as possible.

As to Ubuntu, years ago I used ubuntu to bootstrap debian onto my
portable. That reportbug still gets a ubuntu version number from
somewhere is a monument to reportbug's obscurity. You tell me 
what file has to be changed to get it to give up doing so. Strace
doesn't.


   close 502864 

I would be grateful if you DID something.  Maybe there is no programmer
involved in the work here and that's the problem?

Peter