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.
Do you have a very old computer hooked up to the internet, and want to browse the web like it's 1996 again? Normally this is practically impossible, because a 1996 browser will choke on modern html, css, and javascript. And because most everything is encrypted with TLS 1.3.
Enter the http proxy server that solves both problems: Macproxy Classic
Well, this is in fact no news. The original Macproxy was created by Tyler Hicks-Wright in 2013. I created my fork in 2021. Another super cool fork, Macproxy Plus, emerged in 2024.
What is new today is that I combined the best of all forks, touched it up with some bug fixes and improvements, and polished it to a shine.
Viewing the PDF version of an html file (sans images) in #Zathura, my favorite keyboard-friendly #PDF reader: 157.5 MiB RAM used
Viewing the same original html file (with images) in #dillo: 40.0 MiB RAM used
in #NetSurf: 74.6 MiB
in #GnomeWeb / #Epiphany: 397.0 MiB (wow, kinda lean!!)
in #Falkon: 541.1 MiB
The same file in #luakit: 623.1 MiB RAM
in #firefox / #LibreWolf: 1.31 GiB (YEP)
I can't believe it myself, but yes, I've made #Dillo my default browser on my personal laptops. It's never been my default, even though I've used it occasionally, off-and-on, for 25 years. XD
I still fire up #LibreWolf (#firefox fork) occasionally on those machines, but #DilloBrowser fits in this neat space between terminal browsers and "full-fat" browsers like Librewolf and #luakit.
I'm just wishing it had a "follow mode" for following links from the keyboard, and wondering if there was some way to make it use the clipboard by default, instead of primary selection. I don't quite understand why classic X11 programs use primary selection so much. XD
@rl_dane nice! I also find myself these days running
@dillo as my first browser choice for anything "plain-browsing" until either I can't load a given page or I need to use a webapp (banking, etc). And if I'm using a modest computer, Dillo is the only browser I use.
The amazing thing is that sometimes #dillo is the better choice because it will show me the content of pages that would not otherwise load in Firefox with JS disabled!
I just wish more search engines would accept it. Besides DDG lite, it seems that most #Searx instances outright block its user-agent, making searching for things besides text a lot harder with it :\
@dillo changing the user-agent is a 50-50 thing. I've done this in the past and while doing this did do the trick for #Searx, as soon as I did it a whole bunch of websites stopped loading.
My guess is that after sniffing the user-agent and, seeing something like modern Chrome/Firefox, these sites double down on recognition by sniffing other features that such browsers would present that #Dillo cannot provide. This mismatch is then perceived as a bot of malicious intentions and the request is blocked. Thus it was better for me to keep the real dillo UA if only the search engines were adverse to it (instead of lots of endpoint sites).
But I agree, this is more of a Searx issue rather than Dillo.
Just normal conversations that happen in my discord server, Ritchie is doing accessibility development for the KDE e.V,
@qdot works on the open source sex toy control project called buttplug io
Ritchie says ritchie @qdot can you explain to mr why it is it's easier to find documentation on how to program a buttplug than it is for me to find documentation on how to make a screen reader be able to read my UI because I'm in hell rn
qdot says Priorities do be weird like that
szaszm says xd number of people who enjoy buttplugs > number of devs who care about screen readers 😭
Like the #LKML is kinda shit in terms of UX but at least there are searchable indexes for it so one can just point at a specific message & comment wthout having to be granted access to it.
Not to mention that discord is #ableist af by literally preventing the use of #Screenreader-friendly browsers such as #Lynx and #dillo to work, blocking
@torproject / #Tor users and explicitly banning the use of it's #API to build better clients.
Not even a #paid-for exception is offered, thus making integration into even paid-for tools like #HootSuitenot legally possible!
I started to develop a minimal version of FosseryWeb 1-2 weeks ago, with the aim of Dillo and NetSurf compatibility, I just "finished" the site (in the sense that all articles and cheatsheets are available there).
In the beginning, it was a bit challenging to find ways to implement the planned design in a way that it isn't broken in the above mentioned browsers, but actually it was a fun challenge, it was a bit like a "don't use modern CSS features" challenge lol. I had to make some changes to the originally planned design to make it work, but hope you like it.
Besides the compatibility thing, I also made the site as lightweight as possible, by including only the necessary CSS styles for each page (mostly), and only include images where it's necessary (e.g. in a few cheatsheets).
If you noticed a bug or have any recommendation, feel free to drop a comment.
I'm aware of some article/cheatsheet download issues mentioned on the Known issues section of the About page (https://fosseryweb-min.codeberg.page/about.html), but I couldn't find a solution for those yet (some of those are browser-related, so I can't really do anything about them I guess). I could solve the Dillo issue by including Markdown files in zip, but even then, unzipping it is one extra step, so it wouldn't be any better than right clicking on the link imo.
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.