If I look at the xml file produced inside a new .ooutline package (or just the flat-file xml), it’s completely garbled.
I understand that one shouldn’t normally mess around with the xml file within the .ooutline file, however, the messed up xml file prevents devonthink from being able to search inside the new ooutline files.
I suspect this is an encoding issue because if I try to open it with almost any text editor I can think of (or devonthink’s built-in viewer), I get a garbled mess. However, if I open it with textwrangler, it’s perfectly fine xml. So given that textwrangler is known for automatically recognizing/handling a wide range of encoding/LF issues, I am suspicious about an encoding issue with the new .ooutline files. Any thoughts on this would be greatly appreciated.
TextWrangler is possibly detecting that the XML text file is zipped (a common approach), and unzipping/rezipping it.
If you make a copy of a flat .ooutline file and give it the extension .zip, you should find that the Finder contextual menu will uncompress it for you.
In a more automated context you could use command line unzip and zip.
Interesting point. Thank you. I have tried both pulling the xml file out of the zip file (in the case of a flat file) and pulling the xml file out of the package. In both cases, there was no sign that it’s a compressed type of file. Its filetype is just xml. However, following your suggestion, I tried just using “more” and “cat” in a terminal window, and the bash shell thinks it is a binary file.
So, I don’t know what the heck is going on. An xml file that:
is not part of a package or a zip file
looks like a binary file according to the shell, most text readers, and devonthink
can be read by textwrangler.
Not sure what to do next. I’d love for OO to be compatible with devonthink again, but I’m not sure if this is omni’s problem or devonthink’s problem.
FWIW I’m having no difficulty here unzipping the flat files or inspecting the xml of the package files.
( PS you can recognise the binary content as zipped by the leading ‘PK’, at the start of the file, if you look at it in a plain text viewer or editor )
It may simply be that the QuickLook preview is not quite ready for the new file format – the Finder (like DevonThink) is not yet showing previews for these files.
Quick Look in Finder previews it w/o any problem. Devonthink can’t preview it or find anything inside it.
When I open the xml file with textwrangler (the only program I have that gives a non-garbled result), I don’t see any “PK” at the start of the file. It just begins with: <?xml version="1.0" encoding="UTF-8"?>.
When I look at the xml file with any other viewer, I just get garbage (garbage that doesn’t start w/ PK).
Right, that’s exactly what I find. The spacebar in Devonthink just invokes Finder’s quickview, so that the OS we’re talking to. But devonthink can’t read/find anything in the file. Also, if you look at the preview pane in devonthink, you’ll find that it’s just full of gibberish. I can’t figure out who’s to blame here. Is Omni not presenting a usable format to devonthink, or is devonthink dropping the ball on processing the file?
Though OO5 looks good, for some of us it will be a deal-breaker not to be able to have OO documents indexed in DEVONthink. And to have to look at that garbled text.
Yes, some word from an omni person would be helpful.
I did post about this over at devonthink, but there was mostly silence except for one person who misunderstood the issue. Maybe if someone else could followup on the thread, it might generate some interest from developers over there.
Here’s a link to the thread over there. Some more interest might spark someone to take a look over there as well http://forum.devontechnologies.com/viewtopic.php?f=4&t=22999 .
FWIW I’m having no difficulty here unzipping the flat files or inspecting the xml of the package files.
( PS you can recognise the binary content as zipped by the leading ‘PK’, at the start of the file, if you look at it in a plain text viewer or editor )
It may simply be that the QuickLook preview is not quite ready for the new file format – the Finder (like DevonThink) is not yet showing previews for these files.
This is a user forum, just occasionally watched by people from The Omni Group. If you would like help from them, you could simply send an e-mail to their support. They are very helpful.
The .ooutline file has two options, flat or package. The package option contains uncompressed xml so you should have no trouble reading it in a text editor. The flat option is simple the package zipped as has been noted by others. There’s no setting for .ooutline that just compressing the xml like there is for .oo3.
Quicklook was fixed recently, though does have one bug if you Quicklook multiple files in a row without dismissing the window. As for Devonthink, does it just use Quicklook and Spotlight? It looks like Spotlight indexing for the flat .ooutline format might be broken again, it was working before.
Thanks for joining the discussion @DerekM. I find that Quicklook works perfectly for me. Whether in Devonthink, in Finder, as a package, or as a flat file, Quicklook always works perfectly now.
The problem remains though that devonthink can not find words in OO files. So if I have a bunch of docs stored in devonthink and then search all docs for a particular word, no OO files will ever appear in the search results. This is consistent with the fact that devonthink can not preview the documents (note that devonthink’s preview pane is not the same thing as quicklook). In devonthink, flat files preview as gibberish, while packages show nothing at all in the preview pane. Neither type of file can be searched by devonthink. So for me, the question remains: Is the problem with how OO presents the file to DT, or is it in how DT processes the file?
@jgjones28,
Are you saying there’s a new beta of devonthink or of OO? I’m running the latest OO test build. I don’t know how to get DT betas.
