senknvd

I don't think RFC 4180 differentiates between an empty string and a null value. As long as you add a check that all string columns are free of empty values before writing you should be good.

I think in polars it's

    df.filter(pl.col(pl.Utf8).str.len_bytes() == 0).shape[0] == 0

although there's probably a better way to write this.

There is a --word-diff flag in git diff. It can also be customized using --word-diff-regex to possibly match sentences.

Why would you compare (mostly) military aid to the GDP? Aid from the US has comprised around 15% of Israel's defense budget in recent years and the current bill under deliberation would almost quintuple that sum.

It is the largest recipient of American foreign aid since WW2

> when throttled to similar wattage (see the section about 35W)

I find that Flatpaks[1] work really well for getting the latest version of GUI-only apps. For CLI tools and libraries I haven't found a great solution but I make do with an Arch Linux distrobox[2] container.

[1]: https://flathub.org/setup/Debian [2]: https://packages.debian.org/bookworm/distrobox

From what I've read[1], vim9script was pushed and developed almost exclusively by Bram. With him, a lot of knowledge about its internals and vision for its future dies.

[1]: https://github.com/vim/vim/discussions/12736#discussioncomme...

> There are also these new image lines, I can't recall the funky name for them, that are even smaller.

You might be thinking of the chiselled images. An interesting idea but very much incomplete[1].

[1]: https://github.com/canonical/chisel-releases/issues/34

The only models in the X1 lineup that even offer 8GB RAM are the previous generation. The latest X1 Carbon starts at 16.

According to Notebookcheck[1], if you put a power limit on the 7940HS, it will achieve 15625 points on Cinebench multicore with an average of 66 watts. This puts it above the M2 Max in efficiency and right behind the M2 Pro. Far ahead of any Intel chip.

If you don't want to mess with software power limiting, you'll have to wait for benchmarks of the 7840U.

[1]: https://www.notebookcheck.net/AMD-Ryzen-9-7940HS-analysis-Ze...

Which Chromebook did you go with? And what tools (IDEs, debuggers, etc.) do you use for development?

The M2 can't match the 7840HS in CPU performance though[1]. It's half as fast in multi-core benchmarks and a bit slower in single-core.

[1]: https://www.notebookcheck.net/M2-vs-R7-7840HS_14521_14948.24...

This seems to be true. However, looking at https://rockylinux.org/news/rocky-linux-9-2-ga-release:

9.2 Release for PowerPC (ppc64le) architecture held back

During testing, we discovered an architecture-specific issue on ppc64le systems with the bundled version of Python 3.9. This issue not only prevents installing, but may break existing installations.

This issue is reproducible on CentOS Stream 9 and RHEL 9.2. We have opened a bug report upstream at rhbz#2203919 and are working to fix the issue.

Seems like they hold back releases to do additional testing. In this case, they avoided a bug that was present on RHEL (and presumably AlmaLinux).

I think the entry Pynchon for people on here is Bleeding Edge. Set during the dot com boom with plenty of references to late 90s tech.

I first learned about this in high school from "A Basic Just-In-Time Compiler"[1]. It was mind-blowing that you could just cast some data storing machine code to a function pointer and execute it. Trying the program now it seems GCC no longer accepts it with -pedantic-errors because casts between object and function pointers are, according to the standard, implementation-defined.

[1]: https://nullprogram.com/blog/2015/03/19/

You can do the same with Meson. I just added the warning, debugging, and sanitizer options in-line for clarity. The benefit (I'm not sure if Bazel has this too) is you only have to run the commands once to set up the build directories. After that, you can just use the ultra-fast ninja and almost ignore Meson completely.

If you're using a decent build system[1], it's trivial to test your code with Clang, GCC, and more:

    export CC_LD=lld
    CC=clang meson setup -Dbuildtype=debug -Dwarning_level=3 build-debug-clang
    CC=gcc meson setup -Dbuildtype=debug -Dwarning_level=3 build-debug-gcc
    CC=clang meson setup -Dbuildtype=debugoptimized -Dwarning_level=3 -Db_lundef=false -Db_sanitize=address build-asan

Now you can just

    ninja -C <name of build folder>

to build any and all of these without affecting your source directory. I usually wrap all the useful configurations in a Makefile-like shell script.

[1]: https://mesonbuild.com/

> This seems like a spectacularly bad idea to me, but presumably it made somebody's awful hack work at one point.

but elsewhere you say, "You are welcome to build a type which has this property and declares that it is Eq". This is exactly the difference between double and Double in Java. double has the `==` operator that isn't reflexive and works as required by IEEE-754 and Double has Double.equals, which is documented[1] to be a reflexive, transitive, symmetric, consistent, and substitutable relation.

[1]: https://docs.oracle.com/en/java/javase/17/docs/api/java.base...

the iPad Pro gets to around 9ms perceived latency by using a very interesting trick. Essentially, it predicts the direction the pencil is moving and generates a short straight line in that direction. After the pencil has moved on, it corrects the line to follow the actual path the pencil took. I wonder if the e-ink companies could do that?

Unfortunately, Alpine patched their version of Python in a way that is incompatible with musllinux wheels so we are back to where we started until Alpine Linux 3.16 in 2022[1].

EDIT: I spoke too soon. Apparently, they have decided to amend the patch in a way that is both backwards-compatible and compatible with musllinux wheels[2]. Hopefully, it will roll out soon.

[1]: https://gitlab.alpinelinux.org/alpine/aports/-/issues/13227 [2]: https://bugs.python.org/msg406939

Funnily enough, the situation on FreeBSD is reversed: cut(1) indexes by Unicode codepoints and awk(1) indexes by bytes.

But, more importantly, just indexing by codepoints isn't enough. Grapheme clusters are the real unit of user-perceived characters [1][2]. Technically and practically (just replace the anime music with Indian classical music and you're sure to find some artists with names that can't be split into codepoints), both implementations of cut(1) are broken.

[1]: https://manishearth.github.io/blog/2017/01/14/stop-ascribing... [2]: https://unicode.org/reports/tr29/