chexum

chexum, 8 days ago

@didier Is that hardcoded, or just the way it's adjusting the stack? If I'm reading it right, this address seems to be in the middle of the bootstrap module in v1.3, at an instruction (but not function) boundary, which might be a way to continue obfuscate booting a disk with a demangled bootsector.
If the former, I would have thought it wasn't fashionable to use hardcoded addresses especially as even the A500 had v1.2 and v1.3 ROMs; maybe they thought everyone can upgrade to v1.3.

chexum, 8 days ago

@didier I don't see it actually manipulating the return address, so it might be portable (perhaps to 1.3+ at least). I've never seen this, but it seems when you return 0 from the bootblock, the bootstrap code will just continue booting at the code pointed by A0. That is then set up by the shown code, which read sectors 2..9 (offset $400-$1600) to be executed there. So they just needed a longer boot code it seems

chexum, 15 days ago

@craiggrannell this sounds somewhat similar to when ml/tf regulations mark a transaction suspicious and can end up taking several weeks to resolve “on its own”, as if by magic. When this happens, your(!) bank will keep all parties in the dark as they are required to do so.

Though it’s usually resulting in frozen accounts, so may not be the case - but unusually large amounts are often a trigger.

https://www.which.co.uk/news/article/why-banks-are-freezing-accounts-and-what-to-do-if-it-happens-to-you-aPXrh7F1YCx1

chexum, 15 days ago

@craiggrannell I wouldn’t be surprised if the bank would not admit it though :( If this happens often, it would explain why the publisher is not helping more (they know they have no influence over how the process is handled).

Otherwise the publisher should be aware that only they have the details that could help you confirm (or trace) the payment, so it’s annoying that they are not on your side…. I hope it will be resolved soon!

chexum, 3 months ago

@hyc I have a foggy memory of a radio program about someone explaining that generally all the magic numbers in tales are the “many” of their respective era, like the 7-headed dragons, earlier the 3-headed dog of Hades, so they are basically an imprint of an age when we realised we can’t count to that number, so they practically date the tale.

I couldn’t find any other mention of this so far before your post…

chexum, 6 months ago

@stux This was the tv “programme” that got me on the right path to computers that i kept reading about. I have also got the book, from which I learned the basics of well.. BASIC. It was such a revelation what a FOR NEXT loop actually does. Am I old or what. #TheComputerProgramme

chexum, 7 months ago

@therightarticle I don't get the popularity (or at least the mindshare) of the party that is clearly for the billionaires, by the billionaires...

chexum, 7 months ago

@Wen @therightarticle @simon_brooke I sort of get that for the US folks with their American dream that they can be a (m/) billionaire, they are just not smart enough/don't work hard enough. In the UK it's that's largely stamped out, but apparently what took over is the belief that identifying a Tory is all they need to get as rich as a rich Tory. Don't they see the mass of competition? It makes more sense to support your own peers, and with it, yourself, instead of... these.

chexum, 7 months ago

An iPad goes a bit higher speed but triggers just the same a bit quicker. When I switch off all other APs for the same SSID, occasionally the speedtest completes (600-800 Mbits/s), but still triggered just as easily. The errors just show people trying to find why their (other type of) travel router can't be connected. Finally I got a hint from a message with a different iPhone claiming it can't join the SSID because the used channel is busy with other APs or something. Only 5GHz/WPA3.

chexum, 7 months ago

I turned to a dlink DAP-X1860 with #OpenWRT and it worked all right, at least for a while, which made me realise that the WAX220 has now an easy way of installing the latest 23.05 too (no serial port soldering needed). Then it turned out this WAX220 does the same disconnection with OpenWRT too! (Whaa....????) After a few hours I found out it's indeed the iPhone deciding that the channels are too busy and 80/160MHz can't be used reliably on 5GHz, and that dlink can make it fail just as well.

chexum, 7 months ago

I'm in the EU, so I have only channel 36-48 and 149-161 for reliable no-DFS 80MHz operation, with the dlink not supporting channels over 144. The area has businesses nearby, so that means WiFi 5GHz is indeed not capable of 80 MHz in general, and... that's it. Sometimes it's working for hours, sometimes days. With DFS, it's sometimes minutes, sometimes a week or two until either a legitimate radar or a spurious noise shuts down the AP. So that's wifi in 2023.

chexum, 7 months ago

If it weren't for OpenWRT, I wouldn't be able to update the software - these days, I only buy WiFi infra when I see at least traces of OpenWRT and rightly so - I never had much luck of official AP firmware being released at all after I bought the device. This project-based development seems to mean vendors stop caring after they get the hw for sale. Netgear stopped updating WAX220 fw after May. (I purchased mine in June).

chexum, 8 months ago

@vsaw @itwars phoronix’s review has quite some details about the expected temperatures and the performance with/without cooling. Also, the PSU wattage increase seems to be mostly so that they can provide more power for the USB ports (or, potentially, to M.2/PCI).

chexum, 9 months ago

@ahnlak @Edent 330MBps (bytes) sounds a good effort for a 5Gbps (bits) USB3 cable with USB overhead shenanigans.

chexum, 9 months ago

@ahnlak @Edent yep, that was my point as well, that many older (or cheaper) cables are not 10/20 Gbps capable indeed. But I think actual video data indeed needs alt-mode usb-c, which I’m not sure plain usb-c whichever gen is enough for, or you need specific alt-mode cables (for mere mortals the easiest flag for that is tb3/tb4)

chexum, 9 months ago

@baldur that made me think how much of the perception of intelligence depends on using correct and proper language, which puts foreigners at a great disadvantage.

But language, speaking and writing came very late in the evolution. “AI” does nothing good, except that: writing proper English. Then suddenly it’s the bees knees. That’s seen as a way to displace people, without even displaying any other aspect of thinking?

Instead it’s should assist to empower and level up people!

chexum, 9 months ago

@apps I’m thinking the idea is even simpler than to let the bullies keep their voice. It’s probably because blocking was a great way to remove advertisers from your timeline.

chexum, 10 months ago

@partim according to cpubenchmark.net - the Xeon is about 4.4% faster, but consumes 15W more, otherwise they are in many ways the same (caches, threads, max clockspeed). I'm somewhat disappointed by server class CPUs in general - they are usually much behind current developments. However, at a similar performance point, I would go for them. Except, if we consider consumption as our responsibility...