~sircmpwn

Amsterdam

https://drewdevault.com

I write code.

~sircmpwn/sr.ht-image-failures

Last active 6 hours ago

~sircmpwn/hare-dev

Last active 14 hours ago

~sircmpwn/sr.ht-dev

Last active a day ago

~sircmpwn/sr.ht-discuss

Last active a day ago

~sircmpwn/sr.ht-support

Last active a day ago

~sircmpwn/sr.ht-ops

Last active a day ago

~sircmpwn/hare-users

Last active a day ago

~sircmpwn/email-test-drive

Last active a day ago

~sircmpwn/public-inbox

Last active 4 days ago

~sircmpwn/hare-rfc

Last active 4 days ago
View more

Recent activity

Re: [PATCH meta.sr.ht] Billing/invoice: fix inconsistency a day ago

From Drew DeVault to ~sircmpwn/sr.ht-dev

Thanks!

To git@git.sr.ht:~sircmpwn/meta.sr.ht
   df4fe68..78b07b1  master -> master

Re: builds.sr.ht-images qemu docker build fails due to missing python distutils a day ago

From Drew DeVault to ~sircmpwn/sr.ht-discuss

Applied a patch which should address this issue. Thanks for the report!

Re: Question about polymorphism in QBE a day ago

From Drew DeVault to ~mpu/qbe

The following SSA works for me:

function $example(l %x) {
@start
	call %x()
	ret
}

Re: git.sr.ht patchset editor 2 days ago

From Drew DeVault to ~sircmpwn/sr.ht-discuss

Found the issue and deployed a fix -- thanks for the bug report!

Re: Passing a slice of strings to a variadic function 2 days ago

From Drew DeVault to ~sircmpwn/hare-users

Passing by to say that allocating a slice of formattables is one option,
but another option is to just enumerate the strings and call print in a
loop.

let x: []str = ...
for (let y .. x) {
	fmt::print(y)!;
};
fmt::println()!;

Avoids an allocation (but if the output is unbuffered it will bottleneck
on write syscalls).

[RFC PATCH harelang.org] Multiprocessing patterns for Hare 2 days ago

From Drew DeVault to ~sircmpwn/hare-dev

---
 ...06-11-Multiprocessing-patterns-for-Hare.md | 377 ++++++++++++++++++
 1 file changed, 377 insertions(+)
 create mode 100644 content/blog/2024-06-11-Multiprocessing-patterns-for-Hare.md

diff --git a/content/blog/2024-06-11-Multiprocessing-patterns-for-Hare.md b/content/blog/2024-06-11-Multiprocessing-patterns-for-Hare.md
new file mode 100644
index 0000000..e4d24a4
--- /dev/null
+++ b/content/blog/2024-06-11-Multiprocessing-patterns-for-Hare.md
@@ -0,0 +1,377 @@
---
title: Multiprocessing patterns for Hare
author: Drew DeVault
[message trimmed]

Re: Writing a Unix clone in about a month 4 days ago

From Drew DeVault to ~sircmpwn/public-inbox

I did consider busybox (and also toybox) but I'm not particularly fond
of the huge kitchen sink approach. They're also not as portable as they
seem, you need a fairly beefy POSIX implementation to get them to work.

Overall sbase is the ideal choice in technical terms for a small OS with
a limited POSIX footprint.

Re: [PATCH hare v2] os: add shared memory support 7 days ago

From Drew DeVault to ~sircmpwn/hare-dev

On Sat Jun 8, 2024 at 4:47 AM CEST, Mallory Adams wrote:
> I'm still working on this btw. Got it close to being finished tonight.

Nice, thank you for working on it!

Re: [RFC PATCH harec] Add an instrumentation system 9 days ago

From Drew DeVault to ~sircmpwn/hare-dev

Very nice!

On Wed Jun 5, 2024 at 5:32 PM CEST, Alexey Yerin wrote:
> The compiler can be instrumented by providing one or more -I options.
>
> No instrumentations are provided yet. Main candidates for implementation are
> AFL and AddressSanitizer (if it's still relevant with debug::).

ASan is still definitely relevant with debug::, though my earlier work
on this needs polishing up. Once this lands I can probably take a closer
look at it. One point of note is that it will probably be necessary to
accept additional options for ASan, like '-Iasan:disable=rt::*' or
something, or perhaps we can configure those sorts of things out of
band.