import from OO document

[PhilLAL]

When I import a document from OO into etherpad, I have some format issues :
-I get some empty bullets for a reason I do not understand.
-at random places, 8 spaces are inserted.
-I loose font size.

In the opposite way, after importing a document into etherpad, and then exporting it to a doc, it creates a tab which was not in the initial document.

It is probably an issue since refactoring the imported document needs a lot of time.
Is there any trick to avoid this ? Does this depend on the used format? I actually tried with doc, docx, odt and html, all with the same behaviour.

[ldidry]

-I get some empty bullets for a reason I do not understand.

I think I do understand: when importing an OO file (or whatever format that is not .html or .txt), the file is first converted to HTML, using Abiword or LibreOffice.

Unfortunately, with Libreoffice (Abiword not tested) bullets points are converted to

<ul><li><p style="margin-bottom: 0.28cm; line-height: 108%">Bullet 1</p>
	<li><p style="margin-bottom: 0.28cm; line-height: 108%">Bullet 2</p>
</ul>

You can see two new lines, that are unfortunately interpreted by Etherpad as Enter key… that creates an empty bullet

Manually converting the OO file with

soffice --headless --convert-to html pad.odt

Then changing the list to

<ul><li><p style="margin-bottom: 0.28cm; line-height: 108%">Bullet 1</p><li><p style="margin-bottom: 0.28cm; line-height: 108%">Bullet 2</p></ul>

makes the import working as expected.

I don't know how to deal with this problem:

• removing all new lines in the html file before calling importHtml.setPadHTML in src/node/handler/ImportHandler.js?
• modifying importHtml.setPadHTML to deal with that?

[PhilLAL]

Thank you @ldidry for investiging this point, and for the suggestions. Hope this will help the developers to solve this issue. :)
Regards.

[JohnMcLear]

I have made some progress.

"c:\Program Files\LibreOffice\program\soffice.exe" --headless --convert-to "html:XHTML Writer File:UTF8" testList.odt

--convert-to "html:XHTML Writer File:UTF8" is doing some stuff, things are better, but not right. weird bullet formatting..

It's defo part of the fix but the other part is working contentcollector to drop spans with specific classes.

Bullet_20_Symbols and Numbering_20_Symbols should be entirely discarded.

I also had to modify <title> removal to be more catch all <title in ImportHandler.js

[JohnMcLear]

Fix ready for testing / review.

[PhilLAL]

I'm not sure to understand:

HTML formating

the difference between

<ul><li><p style="margin-bottom: 0.28cm; line-height: 108%">Bullet 1</p>
	<li><p style="margin-bottom: 0.28cm; line-height: 108%">Bullet 2</p>
</ul>

and
<ul><li><p style="margin-bottom: 0.28cm; line-height: 108%">Bullet 1</p><li><p style="margin-bottom: 0.28cm; line-height: 108%">Bullet 2</p></ul>
seems only the breaklines so the rendering should be exactly the same.

import document into etherpad

I did not succeed converting document using soffice command.
However, when importing a document containing a list with two bullets, in the result, each bullet was preceeded and followed by a bullet, resulting in a five bullets list.

Then it means the etherpad adds additional bullets which should not be there. I am not sure how the HTML breaklines could help.

[JohnMcLear]

Try my pr

[JohnMcLear]

sorry about that fast quick response and so the changes uses a different rendering engine for the soffice output.

please let me know how your test goes

[PhilLAL]

I used a private instance of framapad.org in which I imported an ODT containing

test
- bullet1
- bullet2

The result is a pad containing the following

test
- 
- bullet1
- 
- bullet2
- 

Is this the information you need? If not, which one do you need?

[JohnMcLear]

@PhilLAL No. It's okay, I think I have the issue resolved.

[muxator]

I was able to replicate @PhilLAL's bug on a recent development version (6fd2bf4), from 5 days ago.

The result is a pad containing the following

test
- 
- bullet1
- 
- bullet2
- 

Applying #3783 did not solve the issue for me, as the import crashed with no logs.

The ODT file is small (9 KB). I can paste it here if needed.