Publisher: Broken Index?

[Nathan Shirley]

I have a project that's close to 450 pages and have exported it to PDF without issue in the past. I updated to AP 1.9 so that I could add an index (I was disappointed to find AP doesn't currently support multiple indexes, but that's a different issue). After creating a fairly extensive index I tried to export to PDF but the export failed, giving this vague error: "An error occurred while exporting to"...

I noticed something that seems odd when I placed the index. It's a long index and so runs through multiple text frames, but when the end is reached, the index starts over again at the A's... so it's looped, seemingly infinitely. I adjusted the final text frame to end where the index should end, but of course the eye shows up that I've got overflow text. Not really a problem, but it seems to be a bug perhaps as I don't see it happening in the AP tutorial on indexing.

Also, when I check the preflight issues before exporting to PDF it says "the document index needs updating." It doesn't appear to need updating, as far as I can tell. Attempting to update results in the progress bar slowly working up to 100%, but then crashing the program. It crashes whether updating from preflight or from the index window.

I even tried deleting everything from the index along with the index text frames, but exporting to PDF still failed.

I found a thread talking about the same "An error occurred while exporting to" message, where it seems an unusual image might have caused the error and that exporting a PDF/X solves the problem. I have been able to export a PDF/X, but it removes all hyperlinking/indexing/etc (plus converts from RGB to CMYK I believe).

Any ideas what's going on? At this point I almost just want to remove the entire index and make the index manually with anchors and hyperlinks... since I had to do that for my second index anyway. But is there a way to completely delete an index once entries are made (and deleted)?

[Nathan Shirley]

Well by deleting the text frames for the index and then refreshing the index it seems to have fixed whatever problem there was (plus removed the looping). However I still get the same error when trying to export to PDF, so maybe the index issue was unrelated...

I also can't identify a specific page that seems to be causing the export to fail. For example, if I export only page 11, or, 12 they work individually, but if I export 11-12 it fails. I can export 1-10 or 13-15 without issue. I can't find anything unusual in page 11 or 12. There seem to be quite a few other groups of pages that won't export as well throughout the project, and I can't seem to find any link between them... it would take forever to find them all.

Prior to upgrading from 1.8 to 1.9 I didn't have any exporting issues, so perhaps moving to 1.9 is what caused this, rather than the index issue?

[walt.farrell]

 

[Nathan Shirley]

Okay thanks, well just in case it helps speed the process of finding a solution I went ahead and created a "pdflib.log" file. Attached. The end of it mentions something about a negative value that shouldn't be below zero... perhaps that has something to do with it?

pdflib.log

[walt.farrell]

Yes, I think that's the issue. Looks like a problem with a hyperlink to page 12, just after one to page 9. (I'm not sure if that will help you find it.)

[Nathan Shirley]

Hmm... I guess the question is what could possibly be wrong with the hyperlink? And maybe all the hyperlinks in this project have the same issue...?

[walt.farrell]

1 hour ago, Nathan Shirley said:

And maybe all the hyperlinks in this project have the same issue...?

Certainly not all of them, as there is at least one more in that log that worked.

Possibly it's associated with a particular index entry marker, but other than looking for index markers on page 12 and removing them, and re-adding, I'm not sure what to suggest. Perhaps one of the Serif staff, or another user, will have a better idea.

Serif is likely to want to see a copy of the broken document, so please save it before trying to fix it.

[Nathan Shirley]

Thanks for the suggestions. I saved a copy of the project, then reopened the original to try to hunt down issues, and got this message (hadn't seen this before):

"The file includes features from a later version of Affinity."

Pretty strange, since I've been on the latest version... never went backwards...

[Nathan Shirley]

One of the issues seems to be in the second index I made (manually since only one automated index is possible in AP). I made this index in a text frame with 4 columns, close to 200 hyperlinks. I have no idea how to figure out what bug I've hit here. My only guess is that the layout of the text which is hyperlinked might be messing up the hyperlink. It basically looks like this:

 

Some Word for the Index____________________________p. 259

Some Other Word for the Index______________________p. 259

 

And so each full line is a single hyperlink. It would take a really long time to break all those hyperlinks and add them back...

[Nathan Shirley]

I put off trying to fix this issue, hoping an update would fix the bug... unfortunately no such luck.

[walt.farrell]

58 minutes ago, Nathan Shirley said:

I put off trying to fix this issue, hoping an update would fix the bug... unfortunately no such luck.

It is possible, but in my opinion unlikely, that an update would "fix" a broken document. It would be more likely for it to simply prevent such breakage in future documents.

(One update did fix issues with getting an incorrect error stating "this document includes features from a later version", though.)

[Nathan Shirley]

I guess I was thinking the issue might have to do with hyperlinks including underscored lines (_________) and so an update that could handle them might fix the issue. But of course either or both of those assumptions could be wrong. This issue has caused several random-ish crashes, and once I actually got that "this document includes features from a later version" message, erroneously.

I did go ahead and write about this in the bug section since I'm pretty sure this is a bug.