Headings changed to ordered outline when using Conductor and Marked plug for Obsidian

[AutonomyGaps]

I’ve finally gotten Conductor installed, with the default config file, but when I use the Marked plug-in for Obsidian, all the headings are converted to an ordered list. This doesn’t happen using the Marked Streaming action in Drafts, even with identical texts. In addition, if I have a mixed outline, with a top-level ordered list, with unordered bullet-points under each numbered item, they are all converted into a numbered outline. Again, this doesn’t happen with the Drafts action.

Any suggestions for how I can fix this? I love the idea of being able to activate Marked as an Obsidian plugin.

Also, smart quotes are converted to this: ���������

[ttscoff]

Oh my, that’s not great.

It sounds like Outline Mode is being triggered, which can be the result of a setting in Preferences. Do you have anything in this text field?

If it’s being triggered errantly for some reason, you should be able to open the gear menu on the preview and choose Outline Mode->None. Let me know if that does work and we can try to figure out why it’s being triggered. If that’s not the case, let me know and we can look elsewhere.

As far as the smart quotes, that’s a more interesting issue. Do the smart quotes exist in the original Obsidian note, or are they the result of the Markdown conversion in Marked? If you right click on the quote and “Inspect”, does the source shown in the inspector have the invalid symbols?

The most common cause of this is encoding issues. Try revealing the Obsidian note in Finder, copy it outside of the Obsidian vault (so it doesn’t trigger the Conductor track) and open it in Marked. Does the issue persist? If so, you can try opening Terminal and running file ~/path/to/file to see what encoding the document is being stored with. Let me know what you figure out from any of these and we can debug further.

[AutonomyGaps]

Hi Brett,

Thanks for taking this up!

Yes, I had “md” in the Outline Mode field (I think that had to do with an exchange we had years ago about using iThoughtsX with Marked.)

Removing it and restarting Marked hasn’t addressed the issue of flattening the mixed lists into only ordered (numbered). In the screenshot below, 2, 3, 5, 6, 8, 9, and 11 should be rendered as bullets.

As far as the smart quotes, that’s a more interesting issue. Do the smart quotes exist in the original Obsidian note, or are they the result of the Markdown conversion in Marked?

They are in the original note, copy-pasted from ChatGPT, I think.

If you right click on the quote and “Inspect”, does the source shown in the inspector have the invalid symbols?

That doesn’t seem to have the invalid symbols, if I’m doing it properly.

The most common cause of this is encoding issues. Try revealing the Obsidian note in Finder, copy it outside of the Obsidian vault (so it doesn’t trigger the Conductor track) and open it in Marked. Does the issue persist?

No, it doesn’t. The outline problem is still there, but not the invalid symbols.

If so, you can try opening Terminal and running file ~/path/to/file to see what encoding the document is being stored with. Let me know what you figure out from any of these and we can debug further.

I did this and got this result: Survey answers.md: Unicode text, UTF-8 text, with very long lines (658)

I’ve also switched the Appearance settings and the Style Settings plugin in Obsidian, to see if it has any impact. Nada.

Hope we can figure this out, because I really rely on Marked for generating clean RTF or Word or PDF documents from Obsidian. Pandoc doesn’t give me the same control.

Thanks!

[ttscoff]

@AutonomyGaps

Removing it and restarting Marked hasn’t addressed the issue of flattening the mixed lists into only ordered (numbered). In the screenshot below, 2, 3, 5, 6, 8, 9, and 11 should be rendered as bullets.

I’d need to see your original file, but you can’t mix bullet types in a list in Markdown. If it starts with a number, all succeeding list items will be numbered. If you want to nest lists, you need to make sure there are 4 spaces or one tab indenting. 2 spaces will work if your processor is set to GFM, but Markdown needs 4.

@AutonomyGaps

They are in the original note, copy-pasted from ChatGPT, I think.

That’s going to be the problem. Those characters might not be invalid in a text editor, but depending on the encoding they’re probably not recognized by the Markdown parser. I would recommend using a Service/Shortcut to straighten quotes, or ask ChatGPT to not use curly/smart quotes.

If you need a Service for cleaning up Smart Punctuation with a single keystroke, check out the Markdown Service Tools.