I should also add a thought or two more.

After working my way through this and then creating some sort of help files, I see, Lucas, that you suggested this be done on a small file first. I see that you added it later. I always try to do experiments that way at any rate. But thanks for the thought.

I will try it later on a larger file, though only as an experiment, just to see how it works.

I will also need to learn to put more material in the containers and less in the titles. Scrivener will like that. But so long as the export only worked with titles, they were longish.

John

Further to Paul's points, it is also not a requirement to use <tag/> for a tag with no children, <tag></tag> can be used as well (indeed that is what his example does). But, as anyone who's tried debugging broken OPML can attest, using indented code makes it much easier to see the closures.

I think the real underlying problem for @John was an unclosed <outline> tag since when we've been going round in circles. After a morning spent tinkering with OPML I've found some other things (e.g. here re checkbox info) but I think they aren't the issue. I now want to get a code test that I know works verified by John so we can move forward. As a result, I've updated my previous "OPML-for-Scrvener" TBX (right click link and Save as). I've also uploaded the OPML output that it exports: NewOPMLTest.opml (right click link and Save as). The TBX looks like this:



The file's set up so only "New OPML Test" exports, so it's just a mater of opening the file and using File -> Export as HTML. The use of accents, angle brackets and such us deliberate - they're there to test encoding works correctly. I've also added as an alternate, a new pair of templates based on Paul's suggestion further above although please be aware that to use the latter you'll need to make the template assignment changes described in the TBX's help note "The 'Simple' templates".

John, could you please try the above (specifically the files linked to in this post not earlier ones) and tell me if it works for you? I'd welcome another Scrivener user giving a report on whether this works.

I want to ensure we've got export working for John and then , in my admin 'hat' I'll prune or possibly simply delete this thread and write up the successful result. In doing so, I'm not dismissing the help of various forum members here but I'm just conscious that subsequent readers need to work through a page and a half of what doesn't work to get to a solution.  Thought I'd flag the latter as coming - I don't want anyone thinking they're in some way censored.

Jean, what flavor (pro, not pro) and version of OO are you using?  Does it give an error message?  If you export the OPML from Tinderbox, and then try to open it as a new Tinderbox document, does it work?

I'd agree that I'd rather see this thread left intact, and maybe the final answer posted on a locked, sticky thread - or the Cookbook (is the Cookbook actually maintained these days?).

A difficulty with some threads (I've not been following this one closely) is that they fill up with confusing, irrelevant, or obsolete information.  The result is that people today decide not to use Tinderbox because of something that someone erroneously claimed was impossible in 2007.

As I believe I said earlier in the thread:

(a) OPML export from Tinderbox is should not be difficult
(b) We're looking into even better export tailored to Scrivener

Many thanks. Just hopefully get a yes from John and I can sort this thread out.

P.S. A polite request from your moderator that others than John do not add to this thread now as I'll be re-factoring it once I know it he's got the correct code working.

I'm back later with mixed results, Mark. The export file you linked in your text imported directly into Scrivener just fine. No use of Omnioutliner. The children showed fine and the text in the containers as well. Good work.

Not so good with the second material, the full TBX file that included your code. I did an export with it via html, as you suggested. Scrivener didn't like it as a file import. Error message was:

Some files could not be imported because they were of an unsupported file type or contained data that could not be read.

If, however, I exported that file as text using the Simple Wrapper template, it worked just fine. Perhaps you mistyped those instructions.

So, in summary, the code works fine. So long as you use the "text" export button on the templates associated with it. But doesn't work with the html export option.

Thanks again for all the attention to this.

John Kendrick

(Later edit) I continue to want to vote to keep this thread as is. It's a terrific illustration of good customer support and, as I said, a vibrant user community.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Admin note: Thread resolved and locked. After a trip around the houses, a working OPML export example is here.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~