Notices by Dillo browser (dillo@fosstodon.org)

Sadly, the SVG library Nanosvg is now slop.

We embed a sane version in #Dillo to read Wikipedia math which will be kept LLM-free. Human-made changes will continue to be cherry-picked from upstream.

@fanta la URL del repo ya no es esa, nos hemos ido de GitHub y no subiremos más commits ahí: https://dillo-browser.org/news/migration-from-github/

El nuevo repo de git es: https://git.dillo-browser.org/dillo/

Y la última versión es la 3.3.0: https://dillo-browser.org/release/3.3.0/

"If you are a normal visitor and you expected actual content instead of this page, then use a normal, up-to-date browser. It's that simple. [...] You'll have to use another, normal browser. I'm sorry for the inconvenience."

Is 2026 the year of accurate bot protection?

Introducing page actions.

Custom commands can now be executed from the page context menu. Combined with the new dilloc tool, it allows users to arbitrarily manipulate the current page.

Here are some examples:

- Using curl impersonate to bypass browser detection.
- Replace current page by the output of rdrview (reader mode).
- Per-page custom fixes by URL.

Migrating #Dillo from GitHub to our self-hosted server. :blobcatcookienom:

I wrote a post about the current situation with #GitHub and how we ended up self-hosting our own infrastructure to be robust against data loss. We now store all important data (including issues) in git repositories which are replicated accross #Codeberg and #Sourcehut.

https://dillo-browser.org/news/migration-from-github/

Looking for a lightweight CLI fedi/mastodon client while I fix brutaldon OAuth on #Dillo.

Preferably with _very low_ RAM usage.

For reference, toot(1) (written in Python) uses more than three times as much memory as Dillo.

All bug trackers seem to be too complicated, so I made one myself in less than 300 lines. I imported the issues from GitHub and it seems to work well in #Dillo and #links2.

https://bug.dillo-browser.org/

Issues are stored in plain Markdown files with some special headers in a git repository. They get rendered to HTML when pushing a new commit to the bugtracker repo.

The whole thing is less than 300 lines of shell script and awk. It uses no javascript or databases.

Added AddressSanitizer to our automatic CI pipeline, which also runs the memory leak detector.

We are passing the test suite without any leaks on #Dillo code. 😎

#PreciousBytes

@catsalad now put "AI" on the rotating thing and "me" on the cat and welcome to my daily experience on the web.

Go to the kichen and take a scale if you have one. Place under your mouse and turn the scale on, so it starts at 0.

Slowly press the wheel button and see what weight it reaches before it clicks. Same for the left and right buttons. You'll need to repeat it a few times to make sure.

For me, 200 g (!) for wheel and 60 g for the left and right buttons.

This big difference is why in #Dillo the _default_ button to close tabs is set to the right button, rather than the middle one.

Added @Mojeek search engine to #Dillo.

Tip: Press Ctrl+L to select the location bar, then type your search query preceding it with "mj", like so:

mj hdd acoustic attack

I'm testing cgit on #Dillo and after some small CSS fixes it renders really well :blobcatcool:

#SelfHosting

Playing with a new experimental tool "dilloc" to control #Dillo from the command line via a UNIX socket. Here is a simple example to open the current page in the reader view plugin (in the same tab):

% dilloc open rdrview:$(dilloc url)

The "dilloc url" command prints the current tab url to stdout, which can then be easily manipulated. Here is another example with sed, to redirect the current tab to the old reddit interface:

% dilloc open $(dilloc url | sed 's|www.reddit.com|old.reddit.com|')

Hi @Mojeek, can you fix your search engine so that it works without JavaScript?

To reproduce I guess you'll need to visit the search engine with a fresh IP on any of #Dillo, Links, Netsurf or W3M. Any search term causes the error for me. You won't be able to reproduce it if you already did a search on Firefox (and possibly other modern browsers).

I'll assume this is yet another case of collateral damage caused by LLM scraping protection.

For those in the #OldComputerChallenge in need for a lightweight #web browser, #Dillo can be built from source in under 10 minutes from old CPUs (let us know otherwise).

We always welcome new pictures for the gallery: https://dillo-browser.github.io/gallery/

#Dillo release size over time, constrained by the size of a single #floppy disk. We still have some room left.

I suspect the big increase in 3.1.0 is mostly caused by adding a screenshot in the readme (136 KiB).

Not sure if we need to say this but... No, we won't put LLM or any other AI bullshit in the browser.

Fetch a document. Read a document. That's it.

@gettie I currently use brutaldon with some CSS fixes, see:

https://gitlab.com/brutaldon/brutaldon/-/merge_requests/18

@byte @silverfish thanks for the heads up. I'm evaluating self-hosting with snac2 so I can use Dillo directly to load the frontend, but it would take a while.

In the short term you can use RSS: https://fosstodon.org/@dillo.rss

There is no electricity in #Spain today. And yes, I’m sure I don’t need your 15 MiB JS crap to overload the cell tower in order to read the text-based news.