Proud to be Antifa, like my grandfather, who was killed fighting fascism in WWII, and was posthumously awarded the Air Medal and the Purple Heart.
Public Key
npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th Profile Code
nprofile1qqsgtr7wly5kyr6wyxme0xk3haf2rnprqkmpgylm2pqgpfxh0yu442cpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0dsl37vc6
Show more details
Published at
2026-05-28T19:57:07Z Event JSON
{
"id": "16b4859dfa500e9f74830835f23b900cd285414c482dfa057e1c23ea044f62a2" ,
"pubkey": "858fcef929620f4e21b7979ad1bf52a1cc2305b61413fb504080a4d779395aab" ,
"created_at": 1779998227 ,
"kind": 0 ,
"tags": [
[
"t",
"c"
],
[
"t",
"retrocomputing"
],
[
"t",
"sdr"
],
[
"t",
"firmware"
],
[
"t",
"fpga"
],
[
"t",
"verilog"
],
[
"t",
"vhdl"
],
[
"t",
"cplusplus"
],
[
"t",
"nixietube"
],
[
"t",
"calculator"
],
[
"t",
"rpn"
],
[
"t",
"microcontroller"
],
[
"t",
"microcode"
],
[
"t",
"nonpareil"
],
[
"proxy",
"https://mastodon.social/users/brouhaha",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://mastodon.social/users/brouhaha",
"pink.momostr"
],
[
"-"
]
],
"content": "{\"name\":\"🇺🇦 haxadecimal 🚫👑\",\"about\":\"Proud to be Antifa, like my grandfather, who was killed fighting fascism in WWII, and was posthumously awarded the Air Medal and the Purple Heart.\\n#microcontroller #firmware #C #CPlusPlus\\n#FPGA #VHDL #Verilog #SDR (SW Defined Radio)\\n#retrocomputing\\n#nonpareil HP calculator simulation at #microcode level\\nMaker (with John Doran) of #NixieTube #RPN #calculator\\nCHM PDP-1 Restoration Team\\ncis male\\natheist\\nDamned dirty ape\\nNot a tame programmer\\nSome assembly required\\nNo user serviceable parts inside\\n\\npronouns: he/him\\nage: old enough to know better\\nreligion: little endian, Emacs, Linux \u0026 BSD\\namateur radio call sign: N2ES\\n\",\"website\":\"https://mastodon.social/@brouhaha\",\"picture\":\"https://files.mastodon.social/accounts/avatars/000/027/729/original/95248bda6f9a51c7.jpg\",\"banner\":\"https://files.mastodon.social/accounts/headers/000/027/729/original/7fab8b351228fe8d.jpg\",\"nip05\":\"[email protected] \"}" ,
"sig": "ac0e26a185fe7ef689e71d692c33cc605fbc241044aad88e6621daa63ed8fbfd05555eb34eb2bef8a5ed5cf11b5e5d531fed679d2e26ffb313950940ee824ff3"
}
Last Notes npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 And that is ONLY the console and CE (maintenance) panels. The CPU (7101) was nineteen frames, of which were 20 inches wide, 48 inches deep, and 69 inches tall, except for one frame on the end, which was wider. The CPU weighed 26,290 lb. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 They marked the case as "resolved" and closed the chat. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 On chat with a Google "representative" (AI?) about why a device purchased new from Best Buy 11 days ago is not eligible for a Pixel Care+ protection plan. They/it keep telling me that Pixel Care+ is in a phased rollout and not yet available for my device, which is bullshit. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I got an EE at a place I worked a few years ago really angry with me, when I said my usual thing about "EE is trivial, you just wire together all the pins that have the same name." He must have known I was joking, but still reacted badly. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Starship Troopers npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 COBOL isn't particularly difficult. Just annoying. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Tektronix includes in the 11301/11302 manual, under "physical characteristics", BOTH the weight and mass of the scope, in appropriate units: weight in Newtons and pounds, and mass in kilograms and slugs. I don't recall seeing this fone in any other equipment manual; most such manuals incorrect conflate weight with mass. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Most of the time, my Rigol 350 MHz MSO5354 (upgraded 5074) will still be my workhorse, and of course it's much smaller, lighter, and more portable, and I can get digital screen captures from it. There are rare times that I've needed more than 350 MHz bandwidth, so now I have that covered. I'll have to get some 500 MHz probes. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Also tonight I picked up the Tektronix 11302 analog storage oscilloscope a friend offered me a while back. It's a 500 MHz mainframe with three slots for 11 series plugins. But it's not just any old analog storage scope. This has a microchannel plate (like the Tek 2467 portable 250 MHz), and has IEEE-488 and RS-232 ports for control. It also has an amazing self-calibration system, all the way to the display graticule. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Friends and I just watched Star Trek VI: The Undiscovered Country tonight. While I still consider The Wrath of Khan to be the best of the series, IMHO VI is very good, and generally underappreciated. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Someone asked me by email to make sure I package an item securely in "bubble warp". I didn't think warp technology was going to be invented until the 22nd century. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 So-called "AI" is just like the internet: npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Many RISC processors don't have a hardware stack, and depend on software to maintain one if desired, which of course it almost always is. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I'm not criticizing. I understand the fun of designing something minimal. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 No interrupts, I assume? npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Google appears to have a lot of money. Maybe we could convince them to fund an open-source privacy-respecting browser. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 That was the point of the poll. I am writing an assembler for an architecture for which there is no (publicly) documented assembly language. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 preferred assembly language operand order npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 [@mos_8502](https://studio8502.ca/@mos_8502 ) npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 In my ongoing research and development of computer peripherals, I have some improved types of printers, for which I am preparing patent applications: npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Money sent! npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 The spikes are essentially a deciding race. I'll study it more. In the mean time, try inverting the clock into the FFs, but not the AND gates, and see if that's better. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Ignore the top two FF in the '362 diagram; they're an extra divide-by-four for its crystal oscillator, unrelated to the divide-by-four used for the four-phase clock generator. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 In fact, the same thing can be done with the '74 FF. See the TI datsheet for the 74LS362(TIM9904), functional block diagram. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Here's an untested idea for a three-chip TTL solution with no startup issues and needing no power-on reset, using a '109 dual J-/K edge triggered FF, and two '11 three-input AND gates. Being an off-the-cuff idea, it definitely needs to be simulated. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 On further study, it looks like the design is prone to startup problems in general, because flip-flops power up in random states in real life, but not in simulation. But even in simulation, inverting /Q should have the same result as using Q. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 The obvious drawback is that the PIC needs to be programmed, but iit's one chip instead of five. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 To a casual glance, that looks perfectly fine, though I wonder whether you really need the inverter from /Q of the right flop rather than just using Q. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 And all bets are off with virttual inheritance. Even P0149r3 disclaims that. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I found and fixed my most recent C++ issue. It wasn't in the code I was working on at all; that just exposed an extremely stupid thing I'd done elsewhere. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 But I like MATE! I don't want to switch! npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Bit flip errors are the primary reason for use of ECC for on-die caches on high-end CPUs (i.e., x86, ARM 64-bit, but not small microcontrollers), and in DRAM memory systems in servers. Bit flips can occur in DRAM surprisingly frequently, because bits are stored by very few electrons. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 However, radiation of various sorts can cause register and RAM bit flips, or signal errors, without causing CMOS latchup. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 As I understand it, CMOS latchup is caused by triggering parasitic SCR structures (PNPN), and that standard CMOS process design rules make it extremely unlikely to be internally triggered under normal operating conditions. An SEU can inject enough energy to trigger a parasitic SCR. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 That a policy or plan can't possibly result in their stated objectives has never stopped the Republican Party, or their supporters, from anything, no matter how misguided. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 My oscilloscope seems to be getting closer to the edge of the bench every time I use it. I don't want it to fall, so I moved it back in a few inches. Maybe the bench isn't quite level? I've not before had firsthand experience of scope creep. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 "Soon AI agents will be writing better, cleaner code than any mere human can, just like compilers can write better assembly." npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 SSDs are rated to retain data unpowered for six months for consumer grade devices, and for three months for enterprise grade. They need to stay powered up so that they can scrub (refresh) memory cells that lose their charge over time. This is because each data bit is actually stored by a surprisingly small number of electrons in modern NAND flash devices. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Some of us care that our browser is open source, in other words, not loaded up with vendor spyware/malware. It might be reasonable to run the open-source Chromium, but not actual closed-source Google Chrome. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Sure, I don't mind the occasional tech question. Unfortunately, It's hard to know what's gone wrong. Since it's both HDMI and network, normally unrelated functions, that suggests that it might be a software failure, and that reinstalling the device drivers, or the whole OS, might fix it. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Feel better! npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Something I've had to wait three months for: npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Is there a memory expansion header? I think 512MiB is fine, if a daughterboard can add more. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Germany is now advising people to stock up on cheese and sausages. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 The PDP-11/20 could have been installed in racks with other equipment, but the PDP-11/20 processor and main memory (no peripherals) is a box 10.5" x 19" x 24" and weighing under 100 lb. The IBM 1401 processor and main memory (no peripherals) was 61" x 58" x 31" and weighed 2199 lb, so it is 23x the volume of the PDP-11/20 and around 20x the weight. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 No, DHGR _almost_ fell out automatically from the basic design for 80 column text, just as double lores did. Apparently they realized very soon after the start of production, and before first customer shipment, that a very minor motherboard change would enable it. I am 99.9% certain that no ROM change was required, because the ROM doesn't even know whether it's in double mode. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 If it works on an enhanced IIe, it should work on unenhanced IIe. Nothing in the ROMs is required to make DHGR mode work. All IIe variants with 128K RAM can support DHGR, except for a very small number of unenhanced units that shipped with a very early revision motherboard. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I'm thinking about putting everything on a bagel. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Was common on "big computers", and very handy. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Humans (Homo Sapiens) are considered to be in the taxonomic order of primates, and family of hominids (great apes). npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Give a man a gun, and he'll rob a bank. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 That worked because it wasn't running any graphical interface. Just serial terminals. A current Linux kernel, even if you configured out absolutely everything not needed for that, probably wouldn't fit in 1 MiB, though I haven't tried. However, even stripped to a minimum configuration, a current Linux kernel provides a lot of features that BSD 4.2 did not. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 BSD 4.2 ran fairly well in 1 MiB on the VAX 11/780. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I was suspended from Twitter early in the pandemic for replying to someone else's tweet that gave misinformation about hydroxychloroquine as a COVID treatment. I wrote something like: "Hydroxychloroquine is just as effective a COVID treatment as ivermectin. In other words, neither are effective at all." Even on appeal, Twitter refused to reinstate my account until I agreed to delete the tweet. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 That's why my hobby is #retrocomputing. Sure, the old computers from the 1960s and 1970s were thousands of times slower, and had a millionth as much memory, but they were, and still are, fun! npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 And then deleted? I tried to reply to the edited post, but my reply failed,. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I don't have an SR-71, but I do have an SR-52 and an SR-56, both fully operational. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Alchemy shouldn't do absolutely nothing. It should consume a bunch of something valuable, and produce a little of something useless. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Isn't the point of having some smaller ones like 0.1, 0.047, or 0.022 that the impedance vs frequency curve is better? Or do larger values like 4.7 in 0201 have impedance curves more like the 0.1 in 0603? npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 My computer doesn't have a halo, and probably should have horns and a forked tail. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I haven't lost my mind. I have it backed up on floppies somewhere. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Breaking: npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Fun Fact: npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 My friend @npub17pj…xm9e has posted two articles about RISC-V assembly language techniques to his substack: npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I'm writing a new simulator that won't run x86 software, so I'm naming it DONTSBOX. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 My first Linux distribution was Yggdrasil. When that went away, I tried Debian, but back then the Debian install process was like playing 20 questions, except that it was a lot more questions, many of which totally stumped me, despite my having been a Unix user and programmer since 1984. (Clearly Debian fixed that at some later date.) Red Hat Linux was easy to install, which continued with Fedora. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I was worried when Red Hat announced that their non-enterprise distribution was going to become the community-driven Fedora, but it's worked out amazingly well. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 There are many things that have occurred in my life, that I remember as if they were yesterday. Sadly, yesterday is no longer one of them. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I Know What You Did Last Week npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Agreed, but modern programming languages, like C++ but also some others, have made that much more difficult to achieve in the general case. I haven't spent a great deal of time pondering it, but my gut feeling is that raising the semantic level of programming languages might inevitably have this effect. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I should look into whether any of the open source static analysis tools would have detected this. Valgrind detects it at runtime, but I didn't understand what was causing the issue. Poking at it with gdb didn't help me much either, because I was looking in the wrong places. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 It's another case where C++ gave me enough rope to shoot myself in the foot. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 When the vector grows past its previous capacity, it has to be reallocated, which invalidates the iterator, but the loop is still trying to use it. In this case, I think changing the range-based for loop to an ordinary for loop with a numeric index will solve the problem. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I've been trying to track down a nasty memory corruption problem in my C++ program, which manifests right after I do an emplace_back() to a std::vector. My debugging efforts this far have mostly focused on the constructors involved in creating the element for the vector, but I haven't found anything wrong there. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I really like depositing checks by phone. I can hardly wait until they roll out the new feature to deposit cash by phone npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 It appears that the 65C816 ABORT was intended to support virtual memory, but didn't quite get there. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 This is one reason why it would be difficult, if not impossible, to implement virtual memory in a fully general way on the 65C816. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 The WDC dataset only says that processor registers are unchanged if the instruction is aborted, and does not claim that nemory is unchanged. In table 5-7 "Instruction Operation", there are multiple references to note 3, giving exception cases where the registers actually are modified, even on ABORT. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 For instance, suppose that, in 16-bit data mode, you execute an INC of memory. For a 16-bit read/modify/write instruction, the 65C816 reads the low byte of the operand, reads the high byte, does the increment internally, writes the high byte of the result, then writes the low byte. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I just noticed that there's something wrong on the internet! npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Sound output has mysteriously stopped working on my desktop Linux system (running Fedora 42). npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 For NMOS, I'm using the Rockwell R6502 family datasheet. At some point in the past, I carefully compared the table data to a MOS Technology datasheet. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I've now added instruction- and cycle-counting to Ibex, for both 6502 (NMOS) and 65C02, though I think I'm counting one cycle off for read-modify-write instructions in the absolute indexed by X addressing mode on the 65C02. On the NMOS part, these always take the same number of cycles, but on CMOS, they take one less cycle unless there is a page crossing. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 The hate is that they're often used for way too much, or _everything_, e.g. what Firefox does by default except on MacOS. There's no reason Firefox on a desktop OS needs a hamburger menu at all, since normal menus work fine in the environment. Hamburger menus should be provided for little-used commands on platforms with where display space is at too much of a premium for use of a menu bar. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Making a simulator pass all BCD tests is a huge hassle. On the 6502, the resulting N, V, and Z flag values are unspecified, though deterministic. The results with invalid operands are unspecified, though deterministic, but differ between the 6502, 65C02, and 65C816. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 The ibex 6502 & (subset) Apex simulator has now been updated to support the 65C02, including the Rockwell bit instructions that weren't in all vendors' 65C02 parts, but not the WDC STP and WAI instructions. The default is still NMOS 6502, but CMOS can be selected with the "-c" command line option. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 My new probably-uaeful-to-no-one 6502 simulator is ibex: npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 degenerative AI npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I computed the flags wrong for the BIT instruction. I think I made that same mistake when I last wrote 6502 emulation code, 30-some years ago. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 And if I wrote a new 6502 cross-assembler that the world doesn't need, then I ought to write a new 6502 simulator that the world also doesn't need. Right? npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 My new 6502 cross assembler (for PAL65 syntax) is impala: npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 As far as I know, their is no surviving copy of the actual PAL65 cross-assembler, which ran on the DECsystem-10, though the asm65 native assembler was by design compatible with it. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 The bigger question is, what assembler's set of psedo-ops should the PAL65 psedo-ops be translated into? Pseudo-ops for 6502 assemblers seem to have a great deal of variation. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 In last night's 6502 Group meeting, Wayne Wall, primary original author of ASM65, FOCAL, and other 6502 Group software, mentioned the idea of writing a PAL65-to-MOS-Technology syntax conversion utility. And I thought, gee, I've just been writing a cross assembler using a PEG parser. I could almost trivially build such a utility. But since it was already an assembler, it was easiest to first turn it into a PAL65-compatible cross-assembler. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 This was easier to parse, and could be implemented as macros in another assembler (assuming that # and @ characters are allowed in macro names). npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 Examples: npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 I've written yet another 6502 cross-assembler, named impala, based on my work-in-progress IBM System/360 cross-assembler. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 We may be witnessing the end of history, because in the future, it will become impossible to distinguish "real" records from AI slop. npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th 🇺🇦 haxadecimal 🚫👑 That's because Wayland etc. are new and improved.
