Improved my #email workflow a bit with #isync.
So the basic config in ~/.mbsyncrc
is:
IMAPAccount kapsi
Host mail.kapsi.fi
User jjs
PassCmd "pass show imap/mail.kapsi.fi"
SSLType IMAPS
IMAPStore kapsi-remote
Account kapsi
MaildirStore kapsi-local
SubFolders Verbatim
Path ~/.mail/kapsi/
Inbox ~/.mail/kapsi/INBOX
Channel kapsi
Far :kapsi-remote:
Near :kapsi-local:
Patterns *
Create Both
Expunge Both
SyncState *
Essentially only other change I had to do was repeal and replace sources in ~/.config/aerc/accounts.conf
(I was already using msmtp since many many years ago):
[personal]
source = maildir://~/.mail/kapsi
outgoing = msmtp -a iki.fi
default = INBOX
from = Jarkko Sakkinen <jarkko.sakkinen@iki.fi>
copy-to = Sent
[kernel]
source = maildir://~/.mail/kapsi
outgoing = msmtp -a iki.fi
default = INBOX
from = Jarkko Sakkinen <jarkko@kernel.org>
copy-to = Sent
folder-map = ~/.config/aerc/folder-maps/kernel.map
For the record kernel.map (also existed previously unrelated to this change) is just:
personal = INBOX
INBOX = kernel
I use this magic given that @kernel.org routes to a sub-address of the same account.
Have been planning to level up here for a long while so thought now is the time given tomorrow is first workday after holidays.
#Bitwig 5.2 is best update since I started because it replaces so many external plugins with stock options:
It is the first version at least where I’m actually able to draft a track without VSTs. Especially lack of proper bus compressors has been an itch when migrating from #Ableton Live.
Cool got some response from rust-vmm/vm-memory
people, asking to give more details, which I will provide next week.
#Enarx is a kind of hybrid #VMM, and I think this is exactly where I think the overall rust-vmm umbrella has the best opportunities. I.e. instead of just hosting guests with QEMU you might want to do a tailored VMM below to get the optimal characteristics in performance, security and whatnot for your use case.
Enarx hosts #WebAssembly payloads in confidential computing environment. To realize those it:
/dev/sgx
to partition address space for boostrapping a wasm run-time inside an #Intel SGX enclave.The first one is very surgically constructed VM with all access outside it disable, and the second is not actually VM at all.
My take is here that any commercially viable memory API should first of all scale to all possible types on VM’s, and also address stuff “on the edges of universum”, e.g. SGX enclaves, which in many ways resemble VM’s (at least to some level).