Posts
4417
Following
315
Followers
471
Linux kernel hacker and maintainer etc.

OpenPGP: 3AB05486C7752FE1
Edited 11 months ago
When using #m2sync, compared to #mbsync, you need to use "maildirpp:" instead of maildir "maildir:".

Just though to mention since I was puzzling around this for few hours, and wondering why #notmuch finds everything, and aerc finds nothing.

EDIT: Had some issues with this but it seems interesting and relevant so I keep tracking this project. Now I at least got the mbsync bug fixed in Fedora.
0
0
1
@vathpela I just learned Shor's algorithm i.e. total newbie, and still have all kinds of crazy ideas (most of which are probably horribly wrong) ;-) Still excited that I could cope at least the basics.
0
0
0
@vathpela Yeah, right semiconductor CPU's used to be like 10-15 years of the current state of art.
1
0
0
To understand quantum, one needs to not think e.g. a cooking recipes, like in procedural programming.

Instead a closer ever day example for their programming model could be tuning a Swiss pocket watch, or understanding that it detail.
0
0
0
Near 0 K is required for super position states, which are distributions on a unit surface.

QPU's do not compute, they get results by deducing from multiple qubits in superposition. They work exactly and precisely as a natural phenomenon that distributes to the tuned environment.
1
0
0
I wonder if quantum computers were better off deployed to space.

I mean it is colder out the so it should take less energy to cool down to near 0 Kelvin.

All kinds of cargo and even space tourists are going to space anyway so that is not as high coast when amortized, as are the operating costs.

#quantum #computing
3
0
0

I found a modern actively updated alternative for #isync (or #mbsync): #m2sync

It is part of #vomit project: https://sr.ht/~bitfehler/vomit/

Compared to the complexity of ~/.mbsyncrc, my config is now dead simple:

cat .config/vomit/config.toml 
[kapsi]
local = "~/.cache/mail/kapsi"
remote = "mail.kapsi.fi:993"
user = "jjs"
pass-cmd = "pass show imap/mail.kapsi.fi"

I love vomit!

1
0
2
I realized that despite #mbsync crashing in #Fedora I could use #rsync as a backup because I have also SSH shell to my Maildir.

So I give a shot for that, i.e. ssh+rsync, but over longer period of time, I still would prefer to use mbsync. Anyway, it is a workaround.
0
0
0
systemd timers are factors better than cron for periodic stuff *during the session*, such as GIO mounts and mbsync. Outside the session the user timers simply are not there, which actually makes them also more secure than having cron jobs. My fav feature in systemd so far.
0
0
0
My complete systemd experience enlisted:

1. systemctl start/stop/restart/status
2. man journalctl && journalctl blabalbaa (always forget how it was used for some reason)
3. systemd timers (wrote this down so that I will forget it: https://jarkko.codeberg.page/2024/07/11/unprivileged-systemd-timers.html).
1
1
0
Edited 11 months ago
#coredumpctl is pretty cool! I tried it first time yesterday while debugging a SIGSEGV issue with mbsync (aka isync, I hate this name ambiguity) in #Fedora.

I'm surprised that mbsync segfaults with Fedora 40 given despite being niche it is a popular choice among kernel developers.

EDIT: s/journalctl/coredumpctl/ ;-)

#linux #gdb #systemd
1
1
2
@vbabka @melissa what about work less and emit stuff that people care about o_O

I'm with the Team Lazy ;-) Probably not best thing to post social media when you are actually seeking a new job...
0
0
2
Edited 11 months ago
Even if I was using Gmail as my primary email, I'd still create firstname.lastname.delete@gmail.com for subscriptions.
0
0
1
Edited 11 months ago
#Gmail is my #landfill.

I prefer to register web crap there because then I don't have to spend my #email time on unsubscribing the silly notifications that they keep sending.

Then I go there every six months or so and delete All Mail because garbage is all it contains.

#subscription #lifehack
1
0
0
Edited 11 months ago

The dedicated slab buckets series by @kees has been merged into 6.11.

This change allows putting chosen dynamically-sized slab allocations into separate caches instead of generic kmalloc ones.

Patch: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=b32801d1255be1da62ea8134df3ed9f3331fba12
@LWN's overview: https://lwn.net/Articles/965837/

So far, only msg_msg [1] and memdup [2] allocations got their own buckets.

1: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=734bbc1c97ea7e46e0e53b087de16c87c03bd65f
2: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=d73778e4b86755d527a0c6b249cde846770b2f66

0
4
2
@vncresolver BTW do you have screensaver version of this ;-)
0
0
0

@gromit @adamw I added couple:

  1. e70c300f7446 (“permit leading whitespace in INTERNALDATE strings”)
  2. isync-1.4.4-8.fc40.x86_64

Upstream was compiled with the usual autogen.sh && ./configure && make flow.

0
0
0
Show older