Posts
2798
Following
178
Followers
311
Linux kernel maintainer. Compilers and virtualization at Parity Technologies.

Jarkko Sakkinen

Tampere Sunday sightseeing
0
0
4

Jarkko Sakkinen

some funkier stuff produced some years ago https://analrecords.bandcamp.com/track/scum-cums #psytrance #suomisaundi
0
0
1

Jarkko Sakkinen

What I think of Biden/Trump? Pretty much the same feeling that I had when I did this track with my Slovakian-Finnish friend Vlado in 2020: https://globalfishmafia.bandcamp.com/track/i-dont-feel-anything

I was in Portland OR when 2016 elections happened, which was pretty cool I think or once in a lifetime historical experience, mind the result. Or actually, it is not my concern what I might mind the result. I only focus taking stand in the votes where I have a legit standing point in the first place.

E.g. I have a voting ticket for Finnish parliament, EU and LF deciding bodies. So I focus on those and deal with sometimes unfortunate reality.

#politics
0
1
0

Jarkko Sakkinen

songs that stick in your brains forever once you hear them for the first time https://www.youtube.com/watch?v=NlgmH5q9uNk
1
2
1

Jarkko Sakkinen

Binary parsing with #Rust and #goblin https://jarkko.codeberg.page/2024/07/13/binary-parsing-with-goblin.html

The next step is to use #Capstone bindings to disassemble the victim symbol.

#rustlang
0
1
0

Jarkko Sakkinen

For people concerned about #Mozilla #Firefox: #GNOME Web supports Firefox Sync out of the box.
2
0
2

Jarkko Sakkinen

Unfortunately many new open source projects have performance benchmarks that seem to be heritage of Apple marketing.

Zed editor is a great example of this. There's so many moving parts missing here that I could write a book just listing those one by one ;-) This is objectively marketing propaganda because absolute timing values without context and environment do not have any other meaning than giving preferred impression.

Examples of reasonable ways to analyze performance can be found from e.g. https://www.akkadia.org/drepper/cpumemory.pdf.
0
0
2

Jarkko Sakkinen

I guess I’ve activated to post random stories about #vim lately ;-) There’s so much hate for it, especially from some members #neovim community, so I just want to show that there is also other side of the story. In other words, there are people who pick “just” Vim purely based on technical advantages.

It is a bit over year since Bram Moolenaar died so I guess this is also good timing in that sense (RIP) ;-) Remembering that by migrating my vim files to vim9script 🍾

1
0
4

Jarkko Sakkinen

Just learned that #Servo is independent project these days because #Mozilla laid off all of its staff in 2020. #Rust was originally developed specifically for this project. So I guess we have now three competitive web engines instead of two: #Gecko, Servo and #WebKit. servo.org #rustlang
2
1
5

Jarkko Sakkinen

Edited 2 months ago

On #macOS, I’ve found that best way for me to manage #vim installation specifically is to just download the dmg and install pkg given that it auto-updates and contains all the command-line versions (instead of using homebrew or macports):

$ ls -1 /Applications/MacVim.app/Contents/bin
gview
gvim
gvimdiff
gvimtutor
mview
mvim
mvimdiff
mvimtutor
view
vim
vimdiff
vimtutor
xxd
0
0
1

Jarkko Sakkinen

When comparing micro-architectures in general so that it makes any possible sense, at minimum comparative analysis of advantages of disadvantages would be appropriate.

Not sure if I dare to say this but what a load of nonsense.

Just to name an example, the standard for RISC-V with MMU is unfinished and ambiguous to the point that how a SiFive CPU's do stuff is more of a guideline than the specs. The commercial weight if AFAIK more in the compute core area where you just "fork" the hardware description and make it your own.

Not slandering RISC-V per se, I actually like many thing in that ISA, just saying that it is not *even* comparative to x86 and ARM at this point of time. E.g. OpenMIPS would have been (if there had been any analysis of other ISA's than x86).

https://hackaday.com/2024/03/21/why-x86-needs-to-die/
1
2
0

Jarkko Sakkinen

Edited 2 months ago

Interactive external commands is also one reason that made me return back to regular #vim, in addition to working remotes out of the box and the fact that I use only a few plugins, which carry out somewhat trivial tasks where e.g. performance is not a factor.

Interactive external commands is a useful feature from time to time because it allows to leverage privileges for an external command.

For example, this will result an error in #neovim instead of a password prompt: :r !sudo ls.

1
1
1

The next time you're worried that you're not doing great at your job, just remember that someone probably made six figures to change HBO to HBO Max to Max.

0
4
2

Just published the schedule! Lots of good stuff, and at least one terrible talk that nobody should attend.

Early bird tickets are also still available - but not for long - go grab them while they last!

https://all-systems-go.io/

1
8
1
Edited 2 months ago

Another step forward towards making the compiler omniscient:
https://discourse.llvm.org/t/rfc-overflow-idiom-exclusion/80093
(Excluding "test for overflow" code patterns when using the unexpected overflow checker in Clang.)

0
3
0

Jarkko Sakkinen

Edited 2 months ago

Built a site with a PDF resume (see /about) using Jekyll and Typst:

https://jarkko.codeberg.page/

Woodpecker CI puts everything together for every push building both site and the resume:

https://codeberg.org/jarkko/pages/src/branch/main/.woodpecker.yml

And resume made with Typst is a separate project, which can be updated independently while publication is taken care by the site project: https://codeberg.org/jarkko/resume

This is convenient…

#codeberg #woodpecker

0
1
1

Jarkko Sakkinen

Edited 2 months ago

I just realized that #vim has these in the distribution:

❯ ls -1 /usr/share/vim/vim91/pack/dist/opt/
cfilter
comment
dvorak
editexisting
editorconfig
justify
matchit
shellmenu
swapmouse
termdebug

Thus, I can shrink my plugin list a bit:

diff --git a/.vimrc b/.vimrc
index a66c03b..383ea80 100644
--- a/.vimrc
+++ b/.vimrc
@@ -35,13 +35,14 @@ nnoremap <silent> <C-l> :nohl<C-R>=has('diff')?'<Bar>diffupdate':''<CR><CR><C-L>
 nnoremap <silent> <leader>lcd :lcd %:p:h<CR>:pwd<CR>
 nnoremap <silent> <leader>n :set number!<CR>
 
+packadd! comment
+packadd! editorconfig
+
 if !empty(globpath(&rtp, 'autoload/plug.vim'))
   call plug#begin()
   Plug 'ap/vim-buftabline', { 'as': 'buftabline' }
   Plug 'catppuccin/vim', { 'as': 'catppuccin' }
-  Plug 'editorconfig/editorconfig-vim', { 'as': 'editorconfig' }
   Plug 'kaarmu/typst.vim', { 'as': 'typst' }
-  Plug 'tpope/vim-commentary', { 'as': 'commentary' }
   Plug 'vim-scripts/git_patch_tags.vim', { 'as': 'git_patch_tags' }
   call plug#end()
 endif

Leaving only:

if !empty(globpath(&rtp, 'autoload/plug.vim'))
  call plug#begin()
  Plug 'ap/vim-buftabline', { 'as': 'buftabline' }
  Plug 'catppuccin/vim', { 'as': 'catppuccin' }
  Plug 'kaarmu/typst.vim', { 'as': 'typst' }
  Plug 'vim-scripts/git_patch_tags.vim', { 'as': 'git_patch_tags' }
  call plug#end()
endif

#holiday activities

0
0
4

Jarkko Sakkinen

I think #hmac #encryption series for #TPM went well, and also think that enabling it by default only for x86-64 at first was the right call. The recent issue on power architecture proved it. It is better increase scope per request by architectures instead. #linux #kernel
0
1
0

Jarkko Sakkinen

holiday activities
0
0
0
Show older