Can't find MultiMarkdown executable on Linux 1.2.5 Beta

ma
magicfingers
Posts: 62
Joined: Thu May 17, 2012 8:29 pm
Platform: Windows

Wed Aug 01, 2012 12:20 pm Post

When I attempt to compile to a MMD target, I get a "can't find MultiMarkdown" popup error.

However I don't see any way to tell Scrivener where to find it. I presume there is or needs to be a setting analogous to the setting for locating kindlegen, but I don't see it anywhere. Have I overlooked it?

I have multimarkdown installed here:

Code: Select all

/usr/local/bin/multimarkdown


.../usr/local/bin is, of course, included in the system $PATH variable.

User avatar
garpu
Posts: 2110
Joined: Mon Oct 25, 2010 9:38 pm
Platform: Linux

Wed Aug 01, 2012 3:33 pm Post

Same with me. I'm using version 3.6 of MMD, too. (the fletcher-peg one, if it makes a difference.)
Slackware-current 64-bit, XFCE

ma
magicfingers
Posts: 62
Joined: Thu May 17, 2012 8:29 pm
Platform: Windows

Wed Aug 01, 2012 3:37 pm Post

garpu wrote:Same with me. I'm using version 3.6 of MMD, too. (the fletcher-peg one, if it makes a difference.)


Same here, it's the most recent one.

ma
magicfingers
Posts: 62
Joined: Thu May 17, 2012 8:29 pm
Platform: Windows

Wed Aug 01, 2012 5:55 pm Post

I've spent some time trying to sort this one out. On the windows version, the mmd.exe file is bundled into the "docformats" directory, which doesn't exist in the Linux version.

I copied my peg-multimarkdown executable 3.6 (the latest version; the windows release is bundled with 3.5. I can't imagine that this makes any difference) to the new directory "/usr/local/share/scrivener/bin/docformats/" as "mmd" since it's called "mmd.exe" in the windows version.

My test is to compile something to HTML. Even a plain text file should compile to a bunch of <p> tags.

I no longer get the "can't find multimarkdown" popup, and once in a while I get a compile progress bar, but a) I get no output and b) the compile dialog box doesn't close.

When I compile the tutorial, I get a directory with the spacewalk PDF file in it, that's all.

Thinking this was an architecture problem, I repeated this test on a 32-bit vm with a new 32-bit compile of multimarkdown from source. Same result.

So this appears to be one of those "it looks like it's there but really doesn't do anything" problems. That's a real disappointment.

At this point, a rigorous list of what is not expected to be working would certainly help, if there is anything that has not been mentioned already in the 1.2.5 beta announcement.