#bangheadhere — Public Fediverse posts

"SD-WAN" (over MPLS no less!) with IPSec tunnels (not transports!), hard static routes for link monitoring, and floating static routes for tunnel failover... because that's less complicated than using a routing protocol. #NetEng #BangHeadHere

"SD-WAN" (over MPLS no less!) with IPSec tunnels (not transports!), hard static routes for link monitoring, and floating static routes for tunnel failover... because that's less complicated than using a routing protocol. #NetEng #BangHeadHere

"SD-WAN" (over MPLS no less!) with IPSec tunnels (not transports!), hard static routes for link monitoring, and floating static routes for tunnel failover... because that's less complicated than using a routing protocol. #NetEng #BangHeadHere

"SD-WAN" (over MPLS no less!) with IPSec tunnels (not transports!), hard static routes for link monitoring, and floating static routes for tunnel failover... because that's less complicated than using a routing protocol. #NetEng #BangHeadHere

"SD-WAN" (over MPLS no less!) with IPSec tunnels (not transports!), hard static routes for link monitoring, and floating static routes for tunnel failover... because that's less complicated than using a routing protocol. #NetEng #BangHeadHere

First meeting of the day.

Title: "Review Software Test Plan".

Actual topic of meeting: "Review this thing my LLM spat out, which looks a bit like a software test plan, but completely fails to include any details of how we actually do things, or even what the project is about."

#AI #LLM #BangHeadHere

Customer: "Hey, we found a bug in your stuff. Here's a .pcap which triggers it. Can you fix it?"

Us: "Sure, we fixed it, and we added your .pcap as a test case!"

<repeat several times>

<years elapse>

Me: "This .pcap looks like a capture from a live network. Are we sure we don't have people's personal data in here?"

Project manager: "Oh. I never thought of that."

#GDPR #DataProtection #BangHeadHere

Starting to wonder why I ever chose to look into #OAuth in any detail.

I don't think I've ever seen another protocol which attracts so many blowhards and reply guys.

Tip: When an RFC starts with a section labelled "1. Introduction", it's a good idea to read that section, to get some idea of how it fits into the ecosystem.

#BangHeadHere

My first week for work for #2026 draws to a close ... and most of the discussion with my boss has been on the topic of "what do we do about this bloody #idiot #slopmonger we've ended up with on the team"?

I hope the year takes a turn for the better, ideally very soon.

#AISlop #AICoding #NoToAI #BangHeadHere

My first week for work for #2026 draws to a close ... and most of the discussion with my boss has been on the topic of "what do we do about this bloody #idiot #slopmonger we've ended up with on the team"?

I hope the year takes a turn for the better, ideally very soon.

#AISlop #AICoding #NoToAI #BangHeadHere

My first week for work for #2026 draws to a close ... and most of the discussion with my boss has been on the topic of "what do we do about this bloody #idiot #slopmonger we've ended up with on the team"?

I hope the year takes a turn for the better, ideally very soon.

#AISlop #AICoding #NoToAI #BangHeadHere

My first week for work for #2026 draws to a close ... and most of the discussion with my boss has been on the topic of "what do we do about this bloody #idiot #slopmonger we've ended up with on the team"?

I hope the year takes a turn for the better, ideally very soon.

#AISlop #AICoding #NoToAI #BangHeadHere

My first week for work for #2026 draws to a close ... and most of the discussion with my boss has been on the topic of "what do we do about this bloody #idiot #slopmonger we've ended up with on the team"?

I hope the year takes a turn for the better, ideally very soon.

#AISlop #AICoding #NoToAI #BangHeadHere

example_client/api/default/put_object.py:34:62: SyntaxError: Simple statements must be separated by newlines or semicolons
   |
33 |     _kwargs["json"] = _body
34 |     headers["Content-Type"] = "application/ld+json; profile="https://www.w3.org/ns/activitystreams""
   |                                                              ^

example_client/api/default/put_object.py:34:62: SyntaxError: Simple statements must be separated by newlines or semicolons
   |
33 |     _kwargs["json"] = _body
34 |     headers["Content-Type"] = "application/ld+json; profile="https://www.w3.org/ns/activitystreams""
   |                                                              ^

example_client/api/default/put_object.py:34:62: SyntaxError: Simple statements must be separated by newlines or semicolons
   |
33 |     _kwargs["json"] = _body
34 |     headers["Content-Type"] = "application/ld+json; profile="https://www.w3.org/ns/activitystreams""
   |                                                              ^

Jesus...

Oh no, nothing could possibly go wrong here. Nothing at all...

#BangHeadHere

Watching a presentation with a slide that says...

"Humanoids: The Next Frontier In Global Productivity!"

Complete with a number-go-up graph showing that humanoid robots - powered by AI, of course! - will be worth 100 billion dollars by the year 2040.

Please, no ... what fresh hell is this?

#BangHeadHere

I thought I had a reasonably sound understanding of #Git, but current events at work have led me to appreciate it on a whole new level.

Specifically, that it's easy to be productive with it even when one of your colleagues is being a complete numpty and repeatedly pushing changes that make no sense and break the entire system for everyone around him.

This is presumably by design. I bet Linus sometimes has this problem too.

#BangHeadHere

Today's project: figure out how to reset a piece of hardware.

Sounds easy, right?

This thing has no less than 63 reset inputs.

SIXTY-THREE.

#FPGA #BangHeadHere

So you want to build a piece of high performance, deeply pipelined hardware, and write a device driver for it?

The hardest problem is usually: "How do I bring it to a clean stop?"

Get this right. Early.

I hear SW folks proposing "optimisations" to stop the pipeline 10ms faster at the expense of corrupting kernel memory when unexpected errors occur, and I hear HW folks asking "Why do you need to do that? Just reset it!"

Don't be like these people.

#FPGA #ARGHHHHH #BangHeadHere

So you want to build a piece of high performance, deeply pipelined hardware, and write a device driver for it?

The hardest problem is usually: "How do I bring it to a clean stop?"

Get this right. Early.

I hear SW folks proposing "optimisations" to stop the pipeline 10ms faster at the expense of corrupting kernel memory when unexpected errors occur, and I hear HW folks asking "Why do you need to do that? Just reset it!"

Don't be like these people.

#FPGA #ARGHHHHH #BangHeadHere

So you want to build a piece of high performance, deeply pipelined hardware, and write a device driver for it?

The hardest problem is usually: "How do I bring it to a clean stop?"

Get this right. Early.

I hear SW folks proposing "optimisations" to stop the pipeline 10ms faster at the expense of corrupting kernel memory when unexpected errors occur, and I hear HW folks asking "Why do you need to do that? Just reset it!"

Don't be like these people.

#FPGA #ARGHHHHH #BangHeadHere

So you want to build a piece of high performance, deeply pipelined hardware, and write a device driver for it?

The hardest problem is usually: "How do I bring it to a clean stop?"

Get this right. Early.

I hear SW folks proposing "optimisations" to stop the pipeline 10ms faster at the expense of corrupting kernel memory when unexpected errors occur, and I hear HW folks asking "Why do you need to do that? Just reset it!"

Don't be like these people.

#FPGA #ARGHHHHH #BangHeadHere

So you want to build a piece of high performance, deeply pipelined hardware, and write a device driver for it?

The hardest problem is usually: "How do I bring it to a clean stop?"

Get this right. Early.

I hear SW folks proposing "optimisations" to stop the pipeline 10ms faster at the expense of corrupting kernel memory when unexpected errors occur, and I hear HW folks asking "Why do you need to do that? Just reset it!"

Don't be like these people.

#FPGA #ARGHHHHH #BangHeadHere

#AskFedi: Which characters are legal in #OpenIDConnect subject identifiers?

Which document/spec/RFC defines the set of legal characters?

The #OIDC spec just says "The sub value is a case-sensitive string." #BangHeadHere

Please boost if you know someone who might know!

The joys of working on the boundary...

The software team have just reviewed the hardware team's work.

One bright spark came back with:

"This block does not have enough registers. Please add more registers in the next release, so that we have something we can test."

#FPGA #BangHeadHere

Oh, joy, it's time for the annual ITAR / EAR training course.

I'm glad they reminded me that I'm not supposed to be exporting thermal imaging cameras to Belarus. I might have slipped up on this front if I hadn't taken the course.

#BangHeadHere

ARRRRGGGHHHH

Every time I upgrade #Ubuntu it becomes harder and harder to actually open any of my bloody files.

Note to #snap developers: there is a thing called #NFS, and it's traditional to mount it as "/nfs".

People do this in order to store their files in it. You know, those things they want to open using the application you're trying to package?

Even "/mnt/nfs" is starting to be blocked now.

WTF?

#SnapOff #BangHeadHere

First job of the day: a customer support request, which has wended its way over several days from the product application engineer, via the director of silicon design engineering, through two RTL engineers and finally into my lap.

There are two very different answers to the question, and which is best depends on whether they're trying to simulate the design, or run it on a real #FPGA.

Somehow, nobody has thought to ask "what is the customer actually trying to do?"

#BangHeadHere

If hardware engineers don't understand software, and software engineers don't understand hardware ... then how the hell did anyone ever make the bits in the middle, between the CPU and the peripherals, where all the caches, funky burst transactions etc take place?

It's complex, sure, but it's not black magic. I can assure you that not only can it be understood, it can also be optimised.

Or do you think we recovered it from a wrecked UFO in Area 51 or something?

#FPGA #BangHeadHere

Trying to get a Cisco ATA191-MPP connected to a #Microsoft #Teams SIP Gateway and it won’t authorize. It’s a supported device on a supported service, but support replies with “not in scope.” I suspect they don’t get what “supported” actually means. #BangHeadHere

@taral Gaaah.

#BangHeadHere

Reviewing some #specification documents for a project at work.

One of my #review comments: "Documents A and B use different, completely opposing definitions of the term 'wait state'. This is confusing and took me some time to understand. Can we change them to be consistent?"

The response: Document C now includes a parenthetical remark: "Note that documents A and B use different definitions of the term 'wait state'."

#BangHeadHere

Starting my day digging through a pcap file in Wireshark. Newly upgraded to Ubuntu 22.04 shibboleth severs can’t reach LDAP anymore. Yippee.

#tcpdump #Wireshark #Shibboleth #LDAP #BangHeadHere

Nothing says “good morning” like capturing fresh SSL packets via tcpdump and viewing them in Wireshark in an effort to understand why mod_cluster and Apache are misbehaving.

#SystemAdministration #Apache #tcpdump #Wireshark #BangHeadHere

Okay, I got a bit ranty with @[email protected] when I found out that the C9500-24X Network Essentials license doesn’t have VRRP enabled. This was premature. Looking at the data sheet, Network Essentials •does• include VRRP. Now to call TAC and find out why it’s not there. #BangHeadHere #X