:: Re: [DNG] ..heads-up: systemd 253+ …
トップ ページ
このメッセージを削除
このメッセージに返信
著者: Curtis
日付:  
To: Steve Litt
CC: dng
題目: Re: [DNG] ..heads-up: systemd 253+ promises to block forensic imaging and booting of new hardware
seems to me tha Linux is quickly becomin a bootloader for systemd. meh with systemd you’re at the mercy of how well it multitasks since it runs everything as a foreground process. I won’t even get into logging issues.

—Curtis

Sent from my iPhone

> On Feb 23, 2023, at 10:24 PM, Steve Litt <slitt@???> wrote:
>
> altoid via Dng said on Wed, 22 Feb 2023 13:38:39 -0300
>
>>> On 22 Feb 2023 at 14:25, Enrico Weigelt, metux IT consult wrote:
>>>
>>> Or maybe he forgot that nvme isn't the only storage around ?
>> Nope ...
>>
>> What Poettering forgot or probably never learned about (a distinct
>> possibility) are the basic underpinnings of Unix/Linux philosophy.
>
> Yes. In fact, in interviews he looks at Linux with scorn, and implies
> that he's the person to bring Linux up to the wonderful level of
> Windows and MacOS. He has stated he doesn't give a dam about Posix.
>
>>
>> Usually referred to as "do one thing and do it well".
>>
>
> And there's a reason for "do one thing and do it well". When each
> component has one or a very few inputs, and one or a very few outputs,
> and every input and output has a very thin interface, such components
> are incredibly easy to connect together to make whatever is needed.
> Such components are trivial to completely test in isolation so before
> they're put together they're already known to be free of flaws. When
> put together, such assemblies of simple components have clearly defined
> test points for probing and injection, making troubleshooting very
> straightforward.
>
> For instance, on the process level, one can make most of a log analysis
> program by piping together a few greps and awks to get rid of
> extraneous stuff that would be very time consuming in, let's say, Perl,
> and then format and total individual reports with a little
> Perl/Python/Lua/Ruby/AWK with some sorts. If you're doing your reports
> across a slow network, rsync can be included to record changes in the
> logs each day so you're always working with the latest information,
> without repeatedly downloading the same stuff over and over again.
> Doing this in pure Python/Perl/Lua/Ruby/AWK or, heaven forbid C, would
> be a huge, error prone project, probably with all sorts of hard to find
> bugs. Unfortunately, people who don't regularly repair things are
> clueless to the benefits of "do one thing and do it well."
>
> Poettering gets his kicks not by producing robust software, but by
> producing complex interlocking Chinese puzzles in C. He's
> not the slightest bit interested in troubleshooting and repair (in
> other words debugging), so debugging is left to others. The
> interlocking puzzle nature of such software explains all
> the WONTFIX designations in the systemd bug tracking system.
>
>
>> If we go by his "achievements", it would seem he's clueless.
>
> Poettering's crowning achievement is to serve as a useful idiot for
> Redhat, IBM and Microsoft.
>
> SteveT
>
> Steve Litt
> Autumn 2022 featured book: Thriving in Tough Times
> http://www.troubleshooters.com/bookstore/thrive.htm
> _______________________________________________
> Dng mailing list
> Dng@???
> https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng