Well, isn’t the whole point to post a copy of the full text as a basis for the discussion and joint development of it? (Why else?)
Well, that length is actually quite long, at least if itty bitty is any indication! (Note the url, ~4000 characters)
It’s to start a discussion – not necessarily re-invent the tiddler.
It turns out that the answer is browser-dependent. IE has a limit around 2000 characters. Android around 8000. But Chrome allows 32k characters and FF 64k characters.
So, we could impose an arbitrary limit, or limits based on what os/browser you’re using. Or just pretend that everyone’s using FF.
I’m not sure at the moment if carriage returns can be encoded, so that may be a problem too.
Actually, I guess the biggest tiddler on tiddlywiki.com sets the needed max limit 
Right, if we’re unlucky we hit the same problems as with the Share plugin which did something similar.
So another approach might be if tiddlywiki.com could know if the URL already exists or not. If it does already exist, then tiddlywiki.com would only show a regular external link to it and only if it does not exist would the link be the magic link.
So I guess the question is if tiddlywiki.com could send a ping to see if it exists?
I don’t think discourse will respond to a ping at a particular thread. That is, it’s only going to respond to the domain.
I think the easiest solution is to occasionally merge threads that are on the same topic. I’m thinking this will be a rarity.
I’ve added code so that it will grab all the text in the body and submit it. I’ve ignored size limitations for the moment. There are certain characters that require special handling, and I don’t know if I’ve managed to account for all of them. But it works for “HelloThere” and “Getting Started”.
Merging threads manually… rarity - OK, I guess the only way to know is to test.
Great! I tried your mirror site but the resulting editor here on talk looked like before, i.e a proper topic heading but no text. Where can I try it with the text? 
There’s a possibility that I didn’t hit “save”. There’s also the possibility that you didn’t reload. Yeah, let’s go with that one 
. Be sure to clear out your existing talk tab in case your browser is smart enough to navigate back to it instead of loading a new site.
You’ve replied to @twMat 3 times, did you know you could send them a personal message instead?
Edit: I’ve gotten rid of the preamble text and instead appended the date in plain numeric format. That should make it clear which version of tiddler we’re on, and also solve the 15 character problem. Keep in mind that there might be some tiddlers without modification datestamps.
LOL! 
I tested a few and it works quite well. I did hit the limit with the tiddler Code Blocks in WikiText but that is understandable.
…
As for checking broken (same as dead?) links a google search implies there may be js to do this. I’m too tired to read closer right now tho.
I found more characters that weren’t handled by encodeuri. I think if you reload and try again that Code Blocks will work for you now.
It’s better but not quite right: If you copy the generated text and paste it into an empty tiddler+ save and compare it to the original there is some error with something.
Ok, Another fix. It says “0 differences”.
@Mark_S - It work without any problems as far as I can tell. What characters have you found that don’t work? Or is that solved?
it definitely is! Unless we are able to -say- highlight a snippet of tiddler’s text and have a Discourse thread be created around just that snippet. But this would add more implementation difficulty and also scatter the discussion into distinct threads, each of probably too small granularity… don’t know if it is advisable.
Yes, but we are talking about a feedback discussion on the documentation pages for the sake of improving it, so having tiddler’s whole text at hand is key.
I think it’s solved. There’s characters that get special treatment because they’re part of the URI protocol. But they’re surprisingly poorly documented.
It appears to me that the vast majority of tiddlers could be commented using any browser. A handful would need desktop Chrome or FF.
A good test would be to find the largest non-system tiddler and test it. Unfortunately, all my attempts to find the largest tiddler get a stack range error. I’m thinking that changes after 5.2.0 are too resource intensive – I’m sure I did this exercise before 5.2.0 without problems.
What filter did you try?
This seems to work though I haven’t verified that it correctly identifies the longest tiddler:
[all[tiddlers]!is[system]!is[binary]!has[_canonical_uri]!tag[Icon]!tag[picture]] :sort:number[get[text]length[]] :and[last[]]
You could also try the “Alice in Wonderland” tiddler even though that isn’t strictly the longest tiddler (loaded via external text support).
That indicates that the largest tiddler is Release 5.1.23 . It errs out (this site can’t be reached) – from Discourse. I can see that the entire URL has been created and sent (i.e. the browser hasn’t blocked it). So Discourse might have some upper limit (to prevent buffer overload exploits, etc.)
However, a release note isn’t something anyone would typically be editing. So maybe should try some more reasonable tiddlers. Most of the tiddlers appear to be under a 1000 characters.
Not counting release notes (and images), there are only 3 document tiddlers that can’t be submitted due to size restrictions at Discourse (using Chromium desktop):
Demonstration: keyboard-driven-input Macro 8535
tiddlywiki.files Files 10781
Anchor Links using HTML 11038
Good hunting there.
Maybe a doc discussion to have, would be to split up those three tiddlers. Just an idea. Or maybe it is OK if a handful of tiddlers don’t work. It’s not like specific tiddlers are tweaked that often.
Next question – where on TT should new items go? I’m thinking of a sub-category called “tiddlers” to this category (Community Projects), assuming our Discourse implementation can handle sub-categories (@boris , @CodaCoder ) ?