Notices by Vertigo (vertigo@mastodon.social), page 26

YES! Major milestone achieved with the Kestrel project. I managed to run my first C program on the Kestrel-2DX. https://hackaday.io/project/10035-kestrel-computer-project/log/66294-kestrel-2dx-runs-its-first-c-program

This does two things for me:

1) Finally have an approach to running C code, so I no longer have a need to port/invent a different programming language.

2) It vindicates the compatibility of my homebrew processor design against the RISC-V ISA specifications.

(I've since improved my build process since I posted the article.)

On the one hand, I want to use Shen to construct a compiler for writing ROM images for the Kestrel-2DX. On the other hand, it's just one more dependency a new-comer to the project would need to install to get things working.

Why can't Python just *come* with a parser in the standard library already? The value of parsers is known around the world. This should be a no-brainer.

I wish I didn't have to write my own systems programming language. I wish I could just think my firmware into existence.

@clacke Completely non sequitor, but I have to ask -- how did you manage to write a 2407 character toot? ;)

@clacke In times of dire need, those in fortunate positions should always help donate to recovery causes. But, I also sympathize with those who do not trust intermediating organizations, especially those with histories of mismanaged funds.

@clacke I'm going to play devil's advocate here, because I feel these need to be on record.

Donations are fundamentally different; you donate because it provides value to whom you are donating. You generally expect nothing in return; as such, you're basically an angel investor.

Also, there is a large body of historical evidence where charitable organizations were caught consuming greater than 80% of funds donated for non-charitable purposes (usually some form of overhead).

The Kestrel-2DX computer design is coming together. Just implemented the MGIA circuit, so it can now display 640x200 bitmapped graphics out of any one of three 16KB video pages. I also expose VSYNC pulse to the computer via the GPIA input port.

Once the KIA and SD card interfaces are implemented, the 2DX will have feature parity with the Kestrel-2 before it. It will then be time to write an operating system and/or resident programming language for it.

Well, this sucks. I want to post pictures from my mobile device, but Mastodons mobile web interface will not let me upload from it. I can take a new picture or I can film new video but I cannot upload anything from my gallery. :(

16K is used for video RAM, so that leaves only 8K to 32K for programs.

Block RAM is variable, since it depends on the FPGA. It's 48K for Nexys-2 board; but, probably just be 24K for Terasic DE-1.

I'm working on making yet another variant of my Kestrel-2 computer design.

* KCP53000 CPU (64-bit RISC-V) instead of the S16X4A (16-bit stack architecture)
* 32K of ROM
* 48K of FPGA block RAM
* GPIA for misc. I/O incl. SD card
* MGIA for 640*200 monochrome video
* KIA for PS/2 keyboard
* 16MB reserved for "RAM expansion" (where I can develop SRAM, SDRAM, et. al. modules)

Hoping to use it as a development platform for the Kestrel-3. Hence it's name: Kestrel-2/DX.

Attn @cmr - you might find this of interest! Factorial example in Rust successfully compiled for RISC-V for the first time.

https://groups.google.com/a/groups.riscv.org/d/msg/sw-dev/hMB8RKLFYrw/QiQvO-ScAAAJ

TIL: early C compilers did not have unique namespaces for structure members. This is why you needed to write stuff like:

struct Node {
struct Node *n_next, *n_prev;
};

instead of eliding the n_ prefixes.

This was important because the `.` and `->` operators only knew that its LHS was a pointer, not a pointer to a Node.

AROS would make a nice desktop environment for the Kestrel-3. I know I have always wanted a port of Plan 9 for it, but I dunno; AmigaOS just seems like it'd be substantially easier for people to learn how to use and write code for.

Don't mind me; just idly day dreaming as I'm getting ready for work.

Sometimes, I wonder if I should get back into the AROS community (FOSS AmigaOS clone).

@clacke You'll want to contact Bernd Paysan, IIRC, as he's the original author. I believe he hosts its source code on his website. Disclaimer: My info may be dated, though.

If I had to do it all over again, I probably would have gone with Fossil for the project.

Alas, due more to social reasons, I ended up placing my pages on Github (e.g., hosting my blogs as static content, as well as the project pages, etc). I had people saying they preferred git because everyone else was using it, etc.

Of course, except for a small number of people, nobody who said they'd help contribute actually has done anything at all, so go figure. :) Today, the continued use of Github is merely an artifact of inertia.

I remember back when I first was deciding whether or not to host Kestrel-3's pages on Github or not. #fossil was one choice on the table, and was actually a strong contender at the time.

Reason being, #fossil is relatively easy to port compared to its competitors. To port Mercurial, you need to first port Python and enough of its supporting libs to run it. To port Git, you need to basically clone an entire Linux distro's user-space.

Fossil would have let me eat my own dogfood quicker.

So far, I'm working on a side project of the Kestrel persuasion and using #fossil to maintain my project history. More as a means of getting used to working with it.

Good results so far; no worse than using Git except for a few gotchas (e.g., only tracking files and not dirs), but problems have been few in number and relatively minor in scope.