File Export only writes RTF files

I have one folder in my binder with files I need to export to the doc format. I choose all files, click on File/Export. These are all text files. Regardless of the format I choose (DOC, DOCX, ODT) the exporter always produces RTF. (I know RTFs are almost universal but the people I need to send them to are not computer savvy, they’ll freak out.)

I don’t see anything suspicious in the Console. About at the same time I exported the first batch I see a coreservicesd message about a sandboxed application. It’s only once, I suspect I should have found one for each instance I tried to export the files. Here it is:

11/1/13 9:20:26.907 AM coreservicesd[72]: SFLCanProceedWithRequest: Read access denied for sandboxed application to recents belonging to another application.

No Scrivener messages in the Console.

Thanks!

I can’t think of anything that would cause this. What if you try to export to plain text or FDX (as a test)? Also, what if you try creating a new folder in the Finder and export into that?

I did some more testing, so I hope these observations help:

  • I can choose any format if I’m exporting one file at a time
  • I can export multiple files using the FDX format
  • I cannot export multiple files (a selection of documents or a folder) using the doc/docx/odt format
  • The location of the folder doesn’t change anything. (If I create a folder in the Finder, then Scrivener creates its own subfolder to place multiple files.)
  • The problem does not depend on the scrivener project in the different projects I tried.

I imagine it’s an issue with the Java converter? It seems I have everything set up alright and I’m running Scrivener 2.5.

Thanks!

It might pertain to the Java converter, yes. What sort of Mac are you and what version of OS X? If you turn off the enhanced converters via the Import/Export preferences, do .doc and .docx files get output?

Turning off the enhanced converter creates the files I want. They are of a poorer quality than the enhanced converter but for the time being it works.

I have an iMac running Mavericks.

Hmm, very strange. I think the best thing we can do is provide you with a custom build that outputs some extra information to the Console to find out where things are failing. Could you please drop us a line at mac.support(at)literatureandlatte.com and mark the subject line for my attention? I’ll look out for it on Monday and put together a build so that we can troubleshoot what is going on.

Actually, one thing that occurs to me first: do you have anything on your machine that might prevent Java from being able to run? For instance anti-virus software or suchlike?

It must be that the Java converter doesn’t kick in regardless of the IMport preferences because even the importing is wrong. This wasn’t the case in the previous version (2.4.x?). I don’t have an antivirus in this computer and I reinstalled Java just in case. I’m not sure how to test if Java is working. I’ll send an email right away.

Hi Keith.
I’m issuing the same problem while compiling to .docx. If compiling to .doc, everything seems ok. If to .docx, I have an all-rtf file without tabulation, nor footnotes.
The last time I compiled part of the same project (my thesis :slight_smile:) was on Scrivener 2.4 just before upgrading to 2.5, but already Mavericks and the result was still ok.
I should send some pages to my supervisor… and I can’t find any change in the compile options that could explain.
Thanks to chantum, I had the idea to re-do it console opened.
There is the result:
“15/11/2013 18:10:04,714 Scrivener[820]: *** Failed to create file using Java Aspose converters; falling back on standard NSText exporters.”
Should I change something on my Macbook 7.1 (Mavericks) to have converters work?
Thanks a lot for help!
Jean-Louis

edit 18:41
I created a new project and copied the documents to be compiled to. Then tried to compile: worked fine! Thus, something might have been corrupted in the original project…

To me, it sounds as though something in your document caused the Aspose converters (Aspose being the third party component we use) to fail, which would then cause Scrivener to fall back on the only thing it could do at that point. To be sure, please download this version of Scrivener:

literatureandlatte.com/dlbet … ForAEC.zip

Try it with your problematic project. It should fail just the same, but it should print out much more information on the Console that will give me more of a clue as to where it’s failing. Please provide me with the Console report again.

The main difference between Scrivener 2.4.x and 2.5 is that 2.5 comes with a version of Java 1.7 bundled and always uses that. So it might be that the your document is causing a problem with Aspose on Java 1.7 for some reason. I’d therefore be grateful if you could also zip up and send us the problematic project for testing (mac.support AT literatureandlatte.com, marked for my attention).

Many thanks,
Keith

Thanks for answering so quickly.
But I can’t manage to have the “scrivener reporter” work.
“Normal Scrivener” quitted, “scriv. rep.” copied in app folder, then launched, it crasches :blush:
Console:


15/11/2013 21:52:33,347 Scrivener[5567]: *** Assertion failure in -[NSMenu removeItem:], /SourceCache/AppKit/AppKit-1265/Menus.subproj/NSMenu.m:750
15/11/2013 21:52:33,350 Scrivener[5567]: An uncaught exception was raised
15/11/2013 21:52:33,350 Scrivener[5567]: Invalid parameter not satisfying: item != nil
15/11/2013 21:52:33,350 Scrivener[5567]: (
0 CoreFoundation 0x00007fff9060741c __exceptionPreprocess + 172
1 libobjc.A.dylib 0x00007fff9a8b7e75 objc_exception_throw + 43
2 CoreFoundation 0x00007fff906071f8 +[NSException raise:format:arguments:] + 104
3 Foundation 0x00007fff923b0c61 -[NSAssertionHandler handleFailureInMethod:object:file:lineNumber:description:] + 189
4 AppKit 0x00007fff8f92ed11 -[NSMenu removeItem:] + 135
5 Scrivener 0x00000001000482b0 Scrivener + 295600
6 CoreFoundation 0x00007fff90527709 -[NSSet makeObjectsPerformSelector:] + 201
7 AppKit 0x00007fff8f914b2d -[NSIBObjectData nibInstantiateWithOwner:topLevelObjects:] + 1157
8 AppKit 0x00007fff8f9095fe loadNib + 327
9 AppKit 0x00007fff8f908b9c +[NSBundle(NSNibLoading) _loadNibFile:nameTable:withZone:ownerBundle:] + 288
10 AppKit 0x00007fff8f90898c -[NSBundle(NSNibLoading) loadNibNamed:owner:topLevelObjects:] + 197
11 AppKit 0x00007fff8f90877b +[NSBundle(NSNibLoading) loadNibNamed:owner:] + 357
12 AppKit 0x00007fff8f904617 NSApplicationMain + 448
13 Scrivener 0x0000000100002284 Scrivener + 8836
)
15/11/2013 21:52:33,351 Scrivener[5567]: *** Terminating app due to uncaught exception ‘NSInternalInconsistencyException’, reason: ‘Invalid parameter not satisfying: item != nil’
*** First throw call stack:
(
0 CoreFoundation 0x00007fff9060741c __exceptionPreprocess + 172
1 libobjc.A.dylib 0x00007fff9a8b7e75 objc_exception_throw + 43
2 CoreFoundation 0x00007fff906071f8 +[NSException raise:format:arguments:] + 104
3 Foundation 0x00007fff923b0c61 -[NSAssertionHandler handleFailureInMethod:object:file:lineNumber:description:] + 189
4 AppKit 0x00007fff8f92ed11 -[NSMenu removeItem:] + 135
5 Scrivener 0x00000001000482b0 Scrivener + 295600
6 CoreFoundation 0x00007fff90527709 -[NSSet makeObjectsPerformSelector:] + 201
7 AppKit 0x00007fff8f914b2d -[NSIBObjectData nibInstantiateWithOwner:topLevelObjects:] + 1157
8 AppKit 0x00007fff8f9095fe loadNib + 327
9 AppKit 0x00007fff8f908b9c +[NSBundle(NSNibLoading) _loadNibFile:nameTable:withZone:ownerBundle:] + 288
10 AppKit 0x00007fff8f90898c -[NSBundle(NSNibLoading) loadNibNamed:owner:topLevelObjects:] + 197
11 AppKit 0x00007fff8f90877b +[NSBundle(NSNibLoading) loadNibNamed:owner:] + 357
12 AppKit 0x00007fff8f904617 NSApplicationMain + 448
13 Scrivener 0x0000000100002284 Scrivener + 8836
)
15/11/2013 21:52:34,054 com.apple.launchd.peruser.503[162]: ([0x0-0xa30a3].com.literatureandlatte.scrivener3[5567]) Job appears to have crashed: Abort trap: 6


I tried twice, downloading twice to make sure.
Sorry!
But don’t worry about me. Since I can copy all documents in another project, and compile rightly, it’s not worth spending time on debugging my file. I might have done something wrong while using Onyx or something alike.
Your Scrivener is almost fantastic and helps me so much to write my thesis.

Strange, it works fine on my machine. Could you please send the problematic project to me anyway (zipped up)? I’d really love to get to the bottom of what is happening, as I may need to send a bug report to Aspose.

Thanks!

Done

Great, thanks! Was it compiling to .docx format that you had the problem with, or exporting via File > Export. I just tried compiling to .docx from your project and it worked for me, typically. Are you on the Mac App Store version or the version downloaded from us directly?

Thanks!

!!!
Compiling to .docx, not all, but “Le rôle spécifique…” and doc linked below.
Scriv. downloaded from L&L

Hi Keith.
Here are some good news.
I disccovered today a folder created beside my usual one, but with e?? instead of French accent é in the name. We use accents a lot in French.
The folder contained a good .docx copy of the compiled draft, with an accent-removed name of file (me??moire instead of mémoire), while the « official folder » only contained the poor Aspose-failed compilation.
In short : all at once, a good compilation in the bad folder under a bad file name, and a bad compilation in the good folder under the good file name
It seems then that Aspose simply does not like French accents. What a shame !
To make sure,
I compiled the same documents from the same project,

  • to a file with name without accent, in a folder without : good result
  • to a file with name with accent, or in a folder with : double result (two files in same folder; or two folders with one file each, only one being good)
    Hope this time it will help to debug,
    Cheers!

Sorry for the late reply. But a-ha! That may well be the issue. It might not be Aspose but the Java script I use to batch convert the files. Either way, it should be possible to give the names a temporary name internally during the conversion to work around the issue. Now that I have something to go on, I can run some proper tests - thanks!

Just an update on this: it turns out that, for some reason, paths with special characters in them get scrambled when passed through the bundled version of Java 1.7. They work fine when passed through the system version of Java 1.6 (if installed), which is why there wasn’t a problem with Scrivener 2.4.1. I’m not sure why this should be so, but it may be down to the system version of Java 1.6 having been installed and modified by Apple whereas Java 1.7 is straight from Sun Microsystems with no Apple modifications. (Unfortunately, Apple is not developing its own versions of Java going forward, which is why we have to bundle Java 1.7 - bundling is Apple’s recommended route for the future.)

Anyway, I’ve found a fix for this for the next update - Scrivener will look for special characters in the file names and if any are present write to temporary paths internally, thus avoiding problematic paths during the conversion process (all of which will be invisible to the user, of course). In the meantime, there is a way that users encountering this problem can force Scrivener 2.5 to use the system version of Java instead. Here’s how:

  1. Ctrl-click on Scrivener.app and select “Show Package Contents”.

  2. Drill down into /Contents/Resources/FileConverters

  3. Move the file entitled “jreloc” to the Trash.

The internal “jreloc” file tells Scrivener to use the bundled version of Java. Removing it will cause Scrivener to look for an installed version of Java instead.

All the best,
Keith