lone VS coreutils

I made something somewhat close to that: a freestanding lisp. It targets the Linux kernel directly. No libc.

https://github.com/lone-lang/lone

> register scanning isn't portable

Certainly not but it wasn't particularly hard to implement either. I just wrote some inline assembly for every architecture. Here's my programming language's x86_64 and aarch64 implementations:

https://github.com/lone-lang/lone/blob/master/architecture/x...

https://github.com/lone-lang/lone/blob/master/architecture/a...

Not too long ago, a project of mine was shared here on HN.

https://news.ycombinator.com/item?id=38126052

In that thread I wrote:

> I have this vision in my mind: embedding lone modules into sections of the lone ELF and shipping it out. Zero dependencies, self-contained.

I've been working on that since that day. Proud to say I've gotten it to work and thought I'd make it the subject of my first Show HN. Some free software projects gained features along the way too.

The link is to an article with a proper demonstration, technical details and what happened in the past few weeks.

The complete repository itself can be found here:

https://github.com/lone-lang/lone

I've completely reorganized it since the last thread. Would be very happy if you guys tried it out.

To test my programming language. It's a freestanding lisp interpreter that doesn't link to libc. I wrote the code that handles the environment variables and in order to test it I needed full control over the program's inputs including its environment. The env utility provides this control by emptying the environment and setting only the variables I specify, solving 90% of the problem. Only thing I still can't control is argv[0]. With this new feature upstreamed, my test suite will be complete.

Here's the code if you'd like to take a look:

https://github.com/lone-lang/lone#testing

https://github.com/lone-lang/lone/blob/master/scripts/test.b...

> only to be confronted with the notorious 'incompatible glibc version error'. It's super annoying.

I started making my own freestanding Linux Lisp because of this exact issue. It's nowhere near as performant as something like SBCL but it's small and once compiled has no dependencies and will literally run on any Linux.

https://github.com/lone-lang/lone

I'm taking a break from this project but I plan to add a feature where I can put a Lisp script into the ELF itself so I can just copy it with the scripts included.

The kernel just puts the data contiguously on the stack. Obtaining pointers to them can seem somewhat magical if you're writing a nolibc program but I wouldn't call it horrible.

I implemented it for my programming language with some rather simple assembly code:

https://github.com/lone-lang/lone/blob/master/arch/x86_64.c#...

https://github.com/lone-lang/lone/blob/master/arch/aarch64.c...

https://github.com/coreutils/coreutils/commit/fcfba90d0d27a1...

A summary of other changes just released in GNU coreutils 9.5 are:

* mv accepts --exchange to swap files

> some are simple like yes(1)

Not that simple: https://github.com/coreutils/coreutils/blob/master/src/yes.c

The -S / --split-string option[1] of /usr/bin/env is a relatively recent addition to GNU Coreutils. It's available starting from GNU Coreutils 8.30[2], released on 2018-07-01.

Beware of portability: it relies on a non-standard behavior from some operating systems. It only works for OS's that treat all the text after the first space as argument(s) to the shebanged executable; rather than just treating the whole string as an executable path (that can happen to contain spaces).

Fortunately this non-standard behavior is more the norm than the exception: it works at least on modern GNU/Linux, BSDs, and macOS.

[1] https://www.gnu.org/software/coreutils/manual/html_node/env-...

[2] https://github.com/coreutils/coreutils/blob/b09dc6306e7affaf...

> building a cat from scratch

> That would be an interesting project.

Here is the source code of the OpenBSD implementation of cat:

> https://github.com/openbsd/src/blob/master/bin/cat/cat.c

and here of the GNU coreutils implementation:

> https://github.com/coreutils/coreutils/blob/master/src/cat.c

Thus: I don't think building a cat from scratch or creating a tutorial about that topic is particularly hard (even though the HN audience would likely be interested in it). :-)

the yes command, writing to /dev/null, is making IO calls, which interfere with predictable scheduling.

If you look at the source code for yes, https://github.com/coreutils/coreutils/blob/master/src/yes.c

it builds a buffer of output and then writes that in a for loop

  while (full_write (STDOUT_FILENO, buf, bufused) == bufused)

Looking at the source of nohup, if the execvp() of the child happens then it _must_ have already done the signal (SIGHUP, SIG_IGN) so - WTF?

This got me wondering so I went and looked and it seems like lo and behold there was actually a commit to the GNU ls source just 2 weeks ago.

https://github.com/coreutils/coreutils/blob/master/src/ls.c

"maint: prefer char32_t to wchar_t"

even an empty file? Yes. so now it was a file with a copyright disclaimer and nothing else. And the koan-like question comes to mind is "Can you copyright nothing?" well AT&T sure tried.

Then somebody said our programs should be well defined and not depend on a fluke of unix, which at this point was probable a good idea. so it became "exit 0"

Then somebody said we should write our system utilities in C instead of shell so it runs faster. openbsd still has a good example of how this would look.

http://cvsweb.openbsd.org/cgi-bin/cvsweb/~checkout~/src/usr....

At some point gnu bureaucracy got involved and said all programs must support the '-h' flag. so that got added, then they said all programs must support locale so that got added. now days gnu true is an astonishing 80 lines long.

https://github.com/coreutils/coreutils/blob/master/src/true....

http://trillian.mit.edu/~jc/humor/ATT_Copyright_true.html

> Yes, ls is maintained. Although, maintained is a very strong word. It exists.

Why would it be a strong word? Here it is, in src/ls.c: https://github.com/coreutils/coreutils

It is then packaged by tens of operating system distributions, who themselves maintain extra patchsets, some of which are then upstreamed.

It is installed and used on millions (billions?) of devices, for 3 decades.

It's a very reliable and trusty "sharp stick of metal" :)