How cool is @grunfink@comam.es's Snac2 over NetSurf?
The CSS has been created by @anzu@items.minimals.org (inspired by paper-dark and somewhat hacked by me). The only issue were CSS variables that libcss does not support (yet).
Unfortunately Snac relies on detail/summary HTML tag that is not yet properly handled by NetSurf. It would be nice to have a server.json option to open all forms in dedicated pages, so that a social network programmed in C in could be easily used through a browser programmed in C too!
#Snac2 #NetSurf #FreeSoftware #PermaComputing #KISS
Lack of support for the details / summery tags is also why the Dillo browser (@dillo@fosstodon.org) does not provide a completely great experience when connecting to snac.
Unfortunately, changing snac's code to provide an alternative way of managing foldable threads and dialogs is not a trivial change.
CC: @anzu@items.minimals.org
@grunfink@comam.es @giacomo@snac.tesio.it @anzu@items.minimals.org the approach of bloat is to link to an expanded page view when clicking for more details, which seems to work fine in any browser (even without CSS).
Notice that it is also unsupported on Links, Lynx, W3M and Netsurf. I would recommend not depending on the support of the summary tag without a fallback.
snac is essentially a single-page application, opening additional pages to reveal parts would require major changes to the generative code, it would also distort the minimalistic concept underlying snac. It's obvious that opening a new page with text all visible is supported by anything that claims to be an HTML browser. :)
I believe the only valid approach is (as already proposed) to replace details/summary (as configuration level) with a more compatible, fully expanded structure inside page (perhaps tagged with classes), then everyone can apply or not apply any CSS.
CC: @grunfink@comam.es @giacomo@snac.tesio.it
snac is essentially a single-page application
Not much sure I agree.
After all I see a few different URI handled by my snac instance:
https://snac.tesio.it/giacomo
https://snac.tesio.it/giacomo/admin
https://snac.tesio.it/giacomo/notifications
https://snac.tesio.it/giacomo/people
https://snac.tesio.it/giacomo/instanceI can even focus on a single post through search (as I'm doing right now) and interact with it.
So in abstract I don't think it's something that would degenerate snac's nature: it might serve a few more single task views/page/uri, such as
https://snac.tesio.it/giacomo/admin/newpost
https://snac.tesio.it/giacomo/admin/operations
https://snac.tesio.it/giacomo/admin/user-settings
https://snac.tesio.it/giacomo/admin/follower-hashtags
https://snac.tesio.it/giacomo/admin/blocked-hashtagsAnd the Reply button might be turned to a link to search of the post one wanted to reply to.
With such a setup, I guess both #NetSurf, #Dillo and even #Links and #Lynx would be able to interact with a snac server.
Would I make it the default? well, who knows? in an ideal world, since the change would impact the authenticated user that infact also choose the browser to use, I'd make it something people could configure in a user setting.
But I know I'm asking for work I have no time to do myself these days... for something that few people would appreciate. But I guess those few people would appreciate it a lot! π
CC: @dillo@fosstodon.org @grunfink@comam.es
Yes, there are other pages, its so obvious. I use snac and, coincidentally, I've seen.... I'm referring to posts threads with replies. How many pages do you intend to resolve? One per reply? One every three replies? Four replies? How complex do you want to make the generative code for page?
Have you ever seen the rendering code? Since I modify part of it for my own use, I have.
As I said, the best alternative is to eliminate details/summary with a couple of classified divs (classified, so if I want to close them, I can) that remain open without CSS support. After all, there is already a thread collapse setting.
However, since @grunfink@comam.es does a good job and seems to be on its own, I honestly think that snac is fine as it is. Considering that it doesn't use non-standard features, I don't think the problem lies with it, but rather with incomplete browsers. I would say, that you could help them instead of insisting on changing snac code for your own use. I do it because I wanted more markdown and some other goodies, so you can do the same if you want your own features.
CC: @dillo@fosstodon.org @grunfink@comam.es
Sorry for the delay, just seen your reply.
Sure, one might argue that the problem lay on browsers that do not implement the full set of Web standards.
But since more then 20 years such standards are designed by #Google (and geek-friendly #PR dept known as #Mozilla) to reinforce its own hegemony. This means that no real independent group can reasonably hope to build a "standard compliant" browser from scratch.
Now, obviously I gave a deep look to the code of #Netsurf before asking @grunfink@comam.es to modify Snac2.
And unfortunately, the problem is not much handling the details/summary tags (despite them being more complex then one might think at first, because of the name attribute management), but the whole rendering that should be refactored.
The #NetSurf team told me they are analyzing the issue of dynamic rendering, but a solution seem still far to be implemented.
Till then, there is no point into customizing the tags in libdom and enabling a custom rendering in netsurf's layout engine.
Yet you are right that I didn't look specifically into the snac's page rendering code.
And I'm obviously fine to leave it as it is as long as I can't hack on it myself.
But anyway, I supposed that @grunfink@comam.es might be interested in the idea and were curious to discuss it with you.
CC: @dillo@fosstodon.org
Giacomo Tesio
