[Beta 020] rtf that will crash Scrivener compile

The enclosed file gives a hard crash of Compile to Print every time, if it is a segment file in Scrivener. I’ve cut it to two lines of one character each.

  1. I’m printing to Adobe Acrobat, for the compile. It gives a syntax error filled with garbage, at the crash.
  2. duplicating the segment, and trashing the original gives crashes on the dupllcate
  3. however, copying the formatted text out and pasting it into a newly created segment gives a working segment, looking as it should, with no crash.
  4. I do notice that in the copy-paste text scenario, all lines below the top are bulleted. I think this short file you have has bullets on both. But they don’t show in Scriv. A clue?
  5. I had to rename the segment rtf to txt, to please phpbb and get it uploaded
  6. There’s nothing much, and nothing odd in its accompanying two same-name files under Files. I can’t add more than one upload, so not sending them.
  7. As far as I know, this crash issue is fairly recent. That segment has been in and undisturbed for quite a while; however, I hadn’t compiled recently.
  8. Position of the segment in project doesn’t matter - was near beginning, is now at end for convenience.

Hope it helps

Clive
65.txt (620 Bytes)

some more detail, if it matters. I warn you it’s intricate.

  1. I said the offending segment actually carried, but does not show, bulleting on the lines.

  2. The specific is, if you cut-paste the lines to a fresh segment, trashing the old one, you still do not see the bullets, on screen or in the printed compile.

  3. To see the actual bullets, you have to paste the lines into another application. I’ve tried both Word and Evernote, and the lines show up as bulleted in both. Then you can copy-paste the visibly bulleted lines back to Scrivener’s segment, and now the bullets will show normally there.

  4. That fresh segment, with the ctrl-A-ctrl-C-ctrl-V paste of the originally bad segment, and not showing bullets. will sometimes crash a following compile to print. The original was 100%; this one seems about 50%, but not predictable. I think sometimes a copy-paste fresh one won’t crash compile at all.

  5. I don’t get crashes on a very similar segment written in 020 - one plain line followed by several bulleted ones. Thus the original rtf problem seems likely to have come from early betas. You and I would probably rather nothing of the sort could in any case crash a compile, but judgement on effort.

  6. Here’s one further to really make your day.

I thought I had a correlating issue for crashes, but it’s separate and may be of its own importance. Your first compile to print after closing Scrivener and reopening on the project is very swift - about what you’d expect from formatted text out into the pdf. However, a second try at the same thing right after will get a very long pause, well over 10 seconds, and enough to have Windows 7 printing …not responding… on the visible Scrivener windows. Then, finally, it picks up with a good print, and everything is ok. Sometimes, the next compile to print is fast. Sometimes, delay again.

  1. Wierdly, I’ve done this a number of times with Resource Monitor running, and there is no significant CPU or disk spike; nor does the disk light show any particular action. Just, a pause – which looks like the delays people have complained about elsewhere. Events or something? What could cause a timeout?

Ok, that’s your hour or two spent from me today, Lee - have fun anyway please.

Clive