[LH3990] Markdown compile broken

In v17, if I compile to markdown, despite not having the smart to dumb punctuation option selected, ellipses get converted to three dots, em-dash to three hyphens, and italics simply go away (so no * around italicized text). Smart quotes are unaffected.

Random newlines are added or deleted after document breaks.

Also, the filetype selection filter is now missing in the Window’s dialog, instead the file type dropdown has “All Files (*)”.

I am also finding that compile to Markdown is removing random paragraphs. I only noticed upon comparing word counts.

My manuscript in Scrivener is 132452 words. When compiled to Markdown the word count is 115127.

Firstly, you should update from beta 17, I hadn’t realised that one still even allowed you to run it. The latest is 22.

  • The Transformations pane option isn’t wired up yet. You’ll note that despite it converting em-dashes to three hyphens, it does nothing about typographic quotes.
  • Italics—it’s never been intended for the software to convert direct formatting to semantic Markdown. If you want to have formatting in the editor turned into Markdown you have to tell Scrivener what you want. I would use the Emphasis style and set it up in the Styles pane to put asterisks around the styled text with the prefix/suffix fields.
  • “Random” newlines—I’d need specific examples to reproduce that. It may be random, but more likely there is a procedural reason. Check your Separators pane for added newlines. Not sure what would delete them though. I’d need a practical example. Sample projects are the best way to demonstrate this stuff, you can .zip it up and post it as a response.

Again, I’d need precise reproduction notes. I’ve never seen it “randomly” delete anything.

I would suspect that their being able to continue to use version 17 is because of the deactivation date being set to expire in September. The 15th if I remember which is also the same date as with future versions. Without it doing a ‘phone home’ scenario for updates, beta users will point out problems that may have already been addressed.

Yeah, I figured in retrospect, it’s just been a bit of a blur recently and it feels like 17 is from a long time ago. :mrgreen:

Hi, yes I should clarify that I am running Beta 22 (I hadn’t seen the version number in the OPs post).

I have a very simple reproduction case for my issue (it seems to be deleting lines that start with quotation marks). Could I send you the project file?

Just to put this in perspective, I hadn’t realized yet that Scrivener can compile to Markdown!

Features hidden in plain sight, and this one is going to be useful, may even get you some new customers from ‘here’…

Appreciating your blur, @AmberV

Please do! I have not been able to replicate this based on your earlier post, so there must be another component I am missing.

You can either attach it here or email it to win3beta@literatureandlatte.com . If you email it be sure to add a link to this thread. Thanks!

Thanks. I’ve attached the project for you to try and replicate it – also the sample MD output is in there as well.

On some further testing it looks like it is my use of straight quotes that is causing it. If I change these lines to smart quotes they are compiled correctly. I am using non-smart punctuation due to there being no option in the markdown compile to convert to these.
missing_lines.zip (368 KB)

Thanks! I’ve written it up.

By the way, what I said before about italics wasn’t meant to apply to the RTF conversion switch. With that enabled, of course italics should convert—indeed they do in my testing, including with your sample.

Not sure why the date of my post changed. I did a “bump this post”. The original post is from a while ago, the 2nd is the date I bumped it. Guess I won’t use the bump feature anymore, and simply bump the old-fashioned way by adding a reply. I’m currently running 22, not 17.

That’s absolutely ridiculous. Users should not have to mess with anything for bog-standard bold, italics, underline. It was working fine before, don’t treat a regression like a feature. At the very least make the standard behavior to do what you just said when using ctrl-i on text. Users shouldn’t have to go through a manuscript and tag every instance of an italicized word. Is that seriously how it works in the latest Mac version?

This is stuff (italics, dashes) that was working fine in earlier beta versions. I believe 16 and before.

Um, okay then. :smiley:

Just so we’re on the same page though, what you saw before was a bug. For a while the entire RTF to Markdown conversion engine was enabled no matter what, which is indeed not at all how Scrivener is meant to be used, unless of course you have the checkbox enabled to convert RTF to Markdown.

Scrivener is intended to be a platform within which one can write plain-text pure Markdown without any fear of having some conversion engine presume to convert dirty formatting into semantic markings. If for you, you use italic font variants to actually mean something semantic then okay, you can use the conversion engine if you please, that’s what it is there for.

Of course, and not only that, but it is how MultiMarkdown has worked in the Mac version since version 1.0.0. As I say, that is how it is meant to work and why what you saw was a bug.

And of course, you may also not be thinking of this from a forward-looking perspective. One is not “manually tagging every instance” in the sense I think you mean it. If that is how you write to begin with, then it’s all about the same, whichever shortcut you use while typing. Myself, I’ve always preferred the Shift+8 shortcut for italics, but hey that’s just me. :laughing: