Carriage returns not imported

[gerv]

Compare
https://etherpad.mozilla.org/cabf-talking-points-cupertino
with
http://wiki-sandbox.allizom.org/CA:CAB_Forum:2015_Cupertino

In several places, I have used a new line to insert, er, a new line. But it seems that the importer only starts a new paragraph with two successive newlines, and assumes everything which is closer together must be all one runon paragraph. This is most obviously bad with the list of URLs at the end.

The newlines persist in the converted source but of course HTML ignores them. I think a single newline should be a
, and a double newline should be a

.

Gerv

[christi3k]

In mediawiki, paragraph breaks are indicated with two successive newlines. It's true that you can force a linebreak by entering a br tag, but doing so only creates the linebreak on the rendered page, not in the editor, making those pages hard to edit. You can see this here - mixed br and double newlines and here - br to newlines only.

In my experience, in etherpad folks generally use two line breaks to separate paragraphs because one does not provide a significant amount vertical whitespace to actually tell paragraphs apart. Where I see single linebreaks most often is for lists. In this case, the best thing to do to get a well-formatted wiki page that is also easy to edit is to use etherpad bullets. It's also easier in mediawiki to convert a "list" separate with single newlines than it is one separate by tags.

[christi3k]

I just realized that in my last comment I included links to my local machine which is entirely unhelpful for demonstrate purposes.

[gerv]

"but doing so only creates the linebreak on the rendered page, not in the editor, making those pages hard to edit."

That's not been my experience - you can put linebreaks into the markup and they persist. Or did you mean the visual editor?

I agree that your conversion will be much easier if people use appropriate Etherpad markup, but you can't always rely on that. If you think that's vital, then it might make sense to have some text on the import page, which tells people that they should make sure their lists are marked up appropriately before importing them.

[christi3k]

Adding br tags doesn't create a linebreak in WikiEditor unless you put line breaks around them.
I imported those sample pages to the sandbox wiki so you can see what I mean:
http://wiki-sandbox.allizom.org/Cabf_-_mixed_br_and_double_new_lines
http://wiki-sandbox.allizom.org/Cabf_-_br_to_new_line_only

If you edit the first link you'll see that the br tags are mixed in with the text. Now, I can insert a newline with those br tags when generating the content. But now if you want to have an actual list you have to remove all those br tags by hand. If they aren't there, you can just select the entire block of text and hit the bullet button in WikiEditor (you can try this by editing the second link).

I can see the value in both methods. I think I'd like some input from more users before deciding which way to go.

[gerv]

My thought is that the aim of the process should be to reproduce the formatting in the Etherpad - and that's what the first of those links does. It could be that people then want to edit it some more, but I suspect a lot of importing will be archival. If people do want to edit it some more, then they have to deal with their original formatting choices at that point.

"But now if you want to have an actual list you have to remove all those br tags by hand."

No - turns out, a list item with a <br> at the end renders the same as one without, according to my test. So if you put the newline in after the <br>, as you said you could, then I think we get the best of all worlds - formatting which matches the original, and the ability to convert to a list easily.