💾 Archived View for blog.schmidhuberj.de › 2022 › 01 › 21 › html2gmi-mature captured on 2023-04-26 at 13:10:39. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2022-03-01)
-=-=-=-=-=-=-
Posted on 2022-01-21
I just pushed another update to my html2gmi converter, probably making it usable for most cases.
The conversion now correctly converts links without paragraphs and also lists, this now means that all official gemtext-syntax is now supported (I believe). This improvement is now already available on the default template, meaning every website will be supported. This also means that before mentioned gmi2html2gmi now produces correct results.
There are still a ton of sites that need JavaScript to load correctly, these are obviously not supported and will most likely never be supported.
Another major setback was Cloudflare. Some sites use them to protect themself, and Cloudflare seems to think my project attacks their clients servers. The weird thing is, that the protection-site blocking me from accessing their clients websites do not seem to appear using a browser with JavaScript and Cookies disabled, so I have no clue how they figured out I am not a browser, I even spoof user agent in my program. Any help is appreciated on this front.
There are still some other things that might be worth implementing, for example images or tables. Images most likely only need changes to the default template, tables might need some more work. But as tables are not supported by the gemtext-standard, I probably will not implement it until I notice I really want them.
Many gemini sites link to some website on the web (including this one). I proxy those requests using html2gmi-proxy to stay on my favorite gemini client, amfora.
gemini://makeworld.space/amfora-wiki/
I occasionally use this to initiate searches from aforementioned gemini client, but only if I do not have my normal browser open.