I updated Scrivener from 3.1.4.3 to 3.1.5.0 this morning, and when I compiled my 80% written book into a PDF, as I’ve been doing regularly, half the text is just not being outputted.
And it’s not whole sections missing, but large blocks of text across almost every section, with random paragraphs remaining in each.
My PDF that should have been over 300 pages, and was outputting just fine before I updated, is now only ~160 pages.
And if I compile to ODT or RTF, all the text is present.
I would try changing the setting for Print & PDF Resolution, in the Editing: Options tab of File ▸ Options.... That setting can have more of an impact than you’d think, for example greatly changing the quality of text layout.
Otherwise, if that does not work, I would suggest doing some testing in a copy of your project (using File ▸ Save As... to make a disposable copy). I’d approach this by moving everything out of the Draft and then selectively moving in sections that you’ve noted were missing text. If you can reproduce the problem with a very small amount of data, then trash everything else in the binder, empty the trash and use File ▸ Back Up ▸ Back Up To... and send us a copy of the zipped test project so we can take a look at it. Either a PM here or an email to tech support.
Many thanks. As noted in my response, this reproduced for us and should be able to help us locate the problem and fix it.
For anyone else running into this, if you spot blank areas in your PDF where text should appear (the blank areas seem to be roughly the same size as the missing text), then we currently do not know how to not make it happen and advise compiling to RTF, ODT or DOCX as needed, to generate the PDF from your word processor, instead.
There probably is a condition that causes it, as tests against other materials do not exhibit it. If we do figure out what the cause is, I’ll update this post. So click the bell icon to the right if you wish to receive notifications.
I am experiencing much the same problem. Compiling to PDF with a completely (from scratch) format. All folders and docs map to layouts.
It produces pages for the first part of this partially-written manual, about 29 items, about 10 pages, and then it stops mid-document and jumps over the next 28 binder items and resumes. I can’t tell the pattern.
And yes, an outline did pick up all the titles, and I do have all items included in my compile. I also am on 3.1.5.0, about a week ago.
Just a bit more. I selected DOCX as my output and cloned and modified the built-in Manuscript (Courier) format. It also chose to skip over parts of my selection and only export some of what I asked for.
I have not been able to reproduce this happening to text file output. So far all of my testing has been using a lightly edited “Default” with every Type assigned to “As-Is”, for the sake of simplicity. Initially I was unable to reproduce with that setup until I duplicated the format and removed all page break settings from folders, so the document was one long stream, making it easier to see gaps (with page breaks you’d get natural gaps to ignore).
With that setup I was able to see the problem using the built-in tutorial in PDF. But DOCX, RTF and ODT are fine. I then assigned everything to the “Text Section with Heading” Layout, to better see where malfunctions happened in the binder.
You mention duplicating Courier, presumably to make some changes to how it works. Could you describe them, or upload the .scrformat file? I would also try against the tutorial to see if the problem you are seeing may be widespread or perhaps a condition of project configuration.
Thank you for taking time for this important problem, Amber. I will attach the scrformat file I used, as your idea is useful.
I didn’t try this yet with a text file as the intended output, only PDF and DOCX (Word). If you want to recreate what I did, you would need to try it starting with Word (.docx). I really did have documents drop out that way.
Thanks, I’ll check it out! Which section layouts should I use?
Sorry for the confusion, I meant “text file” in the broader sense of the word, to include formatted text as well, not .txt alone. I’ve been primarily checking DOCX since that is what you reported, but it’s good to also check RTF since that is Scrivener’s native format and thus involves fewer variables. If that fails too, then we can assume any of the word processing file types will fail since that is where they all come from.
My project has a structure of Section, Part, Topic, as folders, and then content documents in the Topic folder.
I created a 100 item random tree with this structure, in a new project, item numbers in the titles, and imported the format from the other project. It compiled without any missing bits, leading me to conclude it must be my specific files that are the problem.
Now I am working more on trying to find the issue locally, but not yet resolved. I’ll remember the RTF idea, that’s helpful, thanks.