What is the crowd's opinion of Lynx?

From 3.5 billion comments

I'm not sure that it would really be possible even then unless everything was condensed into one single program that could view webpages. If there's an open standard to view something, you can make anything to view it.

There are things like w3m and Lynx are even able to view web pages in just a text based environment instead of fully rendered webpage.

I can't tell if you think Lynx just lets you open tons and tons of tabs, or you looked into Lynx and genuinely want a text based web browser. I used it a lot back in the day to download drivers for Linux, especially the nvidia graphics driver. Their site was designed really well for text based browsers.

I still use it these days if I need to download something quickly on my Linux server but I don't want to use VNC and I don't have a direct link for wget or curl to use. You can check out a windows binary here if you're interested. Or you can download Knoppix and get the full Linux experience without having to install anything.

Do we need to collect all of the implementations of this idea and have a showdown? (Maybe someone already did this?).

That said, I've mostly been using https://crates.io/crates/uq for this, and I've never needed to care about its performance. The killer feature of this algorithm is that it is line-buffered and streaming, so you can feed it back into itself with a bit of tail -f | while read hackery:

$ (set -eo pipefail ; shopt -s lastpipe ; echo 'https://lynx.invisible-island.net/' > links.txt; tail -n 10000 -f links.txt | grep --line-buffered --only-matching 'https://lynx.invisible-island.net/.*' | uq | (while read -r -t 1 url; do lynx -dump -listonly | grep --line-buffered --only-matching '[.] .*' | tee -a links.txt ; done; echo done; exit 1)) | uq

A few here and there are fine. It tends to start conversations that generally wouldn't be touched. If you want to be as purist, use lynx instead, and as a bonus, no more ads :)

the version I got there https://lynx.invisible-island.net/lynx2.8.9/features.html doesn't, or I gave up trying ot fuss with it to get it to work to be honest.

Is the site server rendered? Like if you "view source" (ctrl-u on Windows or Linux), is there text content there?

Do you have a text browser installed, like lynx? You can see what the pages look like to a bot by using a terminal command like this:

lynx https://example.com/

Not sure what your level of expertise is, but have you tried running your things in virtual machines? Though depending on what you mean by "Chrome can handle 68 tabs" that might just make your problem worse.

What program are you using to drive chrome? You could look into website automated testing tools, selenium is one I've heard of, though this isn't really my field of expertise.

I don't know what kind of scripting capabilities chrome has, but if lynx can handle the website you want to do things on it might be able to do what you want?

In what way doesn't the API work? Are you just missing authentication in postman? if you use the chrome dev tools to repeat the API call does that work? If you take the same action twice what is different between the API calls it makes?

Just to put it on your radar: There are many more text based browsers you can use in a pinch, if other browsers aren't working for some reason.

lynx is pretty good, but I personally like w3m

Nice, thank you, I did not know about uq. If so I might just have provided a patch. You can do the exact same thing with anewer. and since it spits out the lines to stdout and into your file you can make your line simpler (remove the last uq and substitute the tee -a):

$ (set -eo pipefail ; shopt -s lastpipe ; echo 'https://lynx.invisible-island.net/' > links.txt; tail -n 10000 -f links.txt | grep --line-buffered --only-matching 'https://lynx.invisible-island.net/.*' | anewer | (while read -r -t 1 url; do lynx -dump -listonly | grep --line-buffered --only-matching '[.] .*' | anewer links.txt ; done; echo done; exit 1))

I just did a quick speed comparision with my testfiles and anewer outperformed uq clearly and was around twice as fast. I have to say that I used a testfile that was in favor of uq's internal hash algorithm FxHash. It outperforms anewers aHash on short strings. Since anewer still was faster than uq it means that it might be even a magnitude faster on input with longer lines.

pipe the output of that to a .htm file and open it in your browser. Might look rather ugly though depending on whether it needs to load assets.

You could also try the lynx browser: https://lynx.invisible-island.net/. It will allow you to just ignore that cert error.

Mu understanding is that lynx should be able to connect to SSL ports just using:

nntp://host:port/*

For example:

lynx nntp://news.eternal-september.org:563/*

as per docs