Wolf480pl @[email protected]

People say this place is a street.
But they behave like it's their living room.
While in reality, it's a soapbox on the main square.

Meanwhile all I wanted is a watercooler.

Theory:
1. Reinvent a square wheel
2. Understand what's so hard about making wheels
3. Observe how a round wheel someone invented a while ago performs better than your square one.
4. Appreciate the brilliant and simple solutions to all wheel design issues that a round wheel implements.
5. Throw away your square wheel.

Practice:
1. Reinvent a square wheel
2. Push it to production

So there's this AI for GMing RPGs:
https://play.aidungeon.io/

I've been playing around with it, and it's pretty impressive:

me: "I'm getting this compilation error - it says it can't find the socklen_t type" the student replies. This looks like a common mistake. You ask: "Did you put sys/types.h as the first include?"

AI: and he replies that he did. You proceed to say "Well, you need sys/socket.h to use sockets in linux." and he replies that he included that file too.

Is allowlist/denylist just a shibboleth?

Today I convinced one merchant on the polish auction site to fix the mislabeling on one of their offers.

139 999 merchants to go...

I wonder what is the band between 2300MHz and 2400MHz used for in Poland.

Cause, I'd think it wouldn't be too hard to make an optimizaiton pass like:
1. convert all conditions to DNF form
2. split them at ORs, copy the rest of the query, and UNION them
3. remove useless LEFT JOINs

Of course it will only be beneficial if the conditions are highly selective. And I don't know how hard it is to estimate selectiveness of complex conditions.

Any idea if this is still the case with PostgreSQL 12?

TIL postgresql 9.6 doesn't like ORs in conditions.

Converting ORs to UNIONs speed up the query by 3 orders of magnitude.

Tfw. Dell PowerEdge servers come with two types of power supplies: ones with PMBUS support, and ones without.

And no amount of firmware updates will cause power draw measurement to suddenly start working.

Btw. does anyone have a bunch of 0NW455A00 laying around?

You can also use this opportunity to replace tty1 with ttyS1 in /etc/inittab or something, because console=ttyS1 doesn't guarantee the userspace will provide any kind of serial console down the line

How to workaround wierd Linux kernel commandline length limitations:

1. find the earliest point in the init process where you can get a shell, and how to trigger it
2. trigger it
3. find the next script that runs after that
4. in the early init shell, edit that script and add the variables that didn't fit in kernel commandline
5. continue booring

Tomorrow I'll try:
- newer SBUU
- newer SUU
- running a particular update package by hand from a shell in the SBUU
- maybe updating just the BMC and not BIOS (but first I'll try to find out whether it's save to do so)

It still did nothing.
It seemed like it worked, but BIOS is still old :/

Meanwhile, I discovered that instead of trying to find anything on the shitty support & downloads page, I could just go to:

http://downloads.dell.com/published/pages/poweredge-1950.html

It's a blessing.
A static HTML page with all downloads for a particular server model.
And they have pages like that for most of the models AFAIK.