Re: A meditation on the Antithesis of the VMS Ethos

On 2024-07-22, Lawrence D'Oliveiro <ldo@nz.invalid> wrote:

On Mon, 22 Jul 2024 08:54:36 -0400, Arne Vajhøj wrote:
>

It was config for and impacting behavior of kernel code.

>
And it was not subject to the configuration option for turning off
automatic updates. Updates for these files were forced through anyway.

Un-bloody-believable. :-( :-(

I hope this doesn't turn out to be some clueless cretin who thought
they knew better than anyone else when creating an update and have now
just discovered the hard way they did not.

I read somewhere the file that got pushed had nulls in it and that the
file that got pushed was not identical to the one that was tested. :-(

Also turns out their fully-privileged kernel mode driver didn't do the
proper level of validation on this file. (So once again, we are back to
clueless cretin who thought they knew better than anyone else, but only
this time we are talking about the kernel-mode driver writer. :-( )

And no, that is _NOT_ with the benefit of hindsight. When you are writing
this kind of code, you don't trust _anything_ external (or even your own
code :-) ), and you instead validate and perform cross-checks accordingly.
And yes, this _is_ standard practice for any code I write.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.