I just published kotama, an experimental GOOS=tamago compiler branch that allows execution of a rich example in barely 6MB of RAM, running on an rv64imfc CPU (yes, no d or a extension).
This paves the way for supporting aifoundry.org Minion cores!
github.com/usbarmory/ko...
Really Cursed go.dev/issue/77930
TIL there exists such thing as syscall conflicts between Linux and vendor forks of Linux.
3. the best one, on the correct endianness a pointer to a number is also a pointer to all of it's truncations.
So Go gives a *int64 to Linux, Linux loads it as a *int32 so it only loads the lower half.
But this is fine because to convert an int64 to int32 you truncate it to the lower half anyway.
This one is pretty funny go.dev/issue/77934.
Turns out a struct used for Go ←→ Linux syscalls use int64 when Linux expects int32.
But by chance this can't cause a bug because:
1. it is the last field of the struct
2. the size of the struct doesn't change since C aligns int64 to 8 bytes on 32 bits
This could be a devious hacking challenge.
Make a program that has a vulnerability only if you cross compile it on amd64 yet run it on arm64.
I assume it depends what architecture the program is compiled on.
If you compile it on an amd64 machine you get a stupid-ish answer.
Otherwise you get a more sensible answer.
groups.google.com/g/golang-dev...
What sample did you used to fine tune the translation model ?
I cold compiled Firefox today.
Not in artifact mode, a complete compile.
And it finished in 35 minutes !
35 !!!
Linux takes a bit more than an hour.
I was ready to come back tomorrow.
How is it so fast ?
In my threat model hackers have arms and USB cables.
github.com/meshtastic/f...
Screenshot of video game children wizards from the game YAPYAP.
TELEPORTANA
ANAPORTTELE
github.com/golang/go/is...
After a bit of research the algorithm currently used is un-optimimal and a simple fix would be kicking down the worst case later.
I don't know if I want to commit to rewriting it ?
Code comment: partition can grow in the loop. By not using a range loop here, we process new additions as they arrive, avoiding O(n^2) behavior.
HUMMMMMMM
This doesn't look to be working as intended.
Plot of numbers of lines of code in a function versus the compile time by go build. Shows an O(n*n) trend with 2 to the 15 at 50 seconds.
Hum that compile time versus number of lines of code in the function looks O(n*n) suspicious.
A pick and place machine in front, with an industrial storage shelf in the back with three 3D printers.
Playing factorio in real life:
The factory is growing !
2nd Pick-and-Place and 4 3D printers print-farm (4th one is bigger and doesn't fit).
Commit 7b7abfe in torvalds/linux. From November 10 2005, shows a completely empty description and title.
Feeling like that time someone merged a completely empty commit description and title to the Linux kernel.
In fact one already supports doing so, but the UX makes it unsuitable for whole home backup.
I guess we will see more DC charger spoofing options, where an inverter basically directly pull energy from the battery pack through the DC charging pins (which skips the car's on board power modules).
An other interesting idea is the household already owns 140kWh of battery storage !
However it is in the cars, and difficult to access due to proprietary Vehicle to Grid solutions.
Ideally we can use the on board car AC charger reversed as an inverter.
Conclusion after 7100€ / 3300€/y = ~2.15 years I could pay off a completely offgrid photovoltaic + battery system for a home system.
Now think what utilities can do with economies of scale. (also maybe like -20% consumption if I swap the heatpump but renter moment)
Adding the cost of batteries I currently use ~30kWh during the "night".
Remove ~20kWh of that by charging the EVs during the day (currently doing night to use reduced rate nuclear power).
Right now I could buy a second hand EV 17kWh battery for 4k€, maybe add an other 1k€ for the charger.
Given my electrical provider does net metering (aka the meters run backward if I provide power to the grid).
Overproducing during the day would cover night consumption.
Note that this is toxic behavior, since you are a net monetary loss to the power network, utilities are catching on.
Ok so I did the math, results are crazy.
My household is currently paying 0.19€/kWh this is 3312€ a year, yes it's a lot, reasons include a broken heatpump and 2EVs.
For 1k€ I can buy 5kw of solar panels, a 6kw for 1.1k€ more.
As you might have noticed, if I install them myself, 2100€ < 3312€.
NVM rm is multiple orders of magnitude faster if you don't use -v
Never before I thought the performance of rm needed to be improved.
I need to remove a single directory with a few hundred millions of files.
I know it looks insane but it's on a modern file system (btrfs) so it works fine.
Anyone got multithreaded faster alternatives to suggest ?
> ssh root@192.168.1.147 WARNING: connection is not using a post-quantum key exchange algorithm. This session may be vulnerable to "store now, decrypt later" attacks. The server may need to be upgraded. See https://openssh.com/pq.html
TIL openssh-client now warns you with a very helpful error message if the server does not support post-quantum cryptography.
Gotta update my robot vacuum I guess.
Recent golang/go commit history shows many commits by @Jorropo (myself).
Enjoying the end of the winter freeze and you ?
I need to make a story inspired on Roko's Basilisk but using instrumental convergence as a premise.
It has a monotonic component, for all relative times materialized from `time.Now` and not stripped of their relative component (imagine serializing it and reparsing it), .Sub does the difference between two points since the start of the program.
Counter point:
> Mon Jan _2 15:04:05 2006
Didn't have "the Mercator projection will cause the end of the world" on my bingo chart
