I was pretty excited to see that DT 1.7.4 now has Webcore integration, and started moving all of my HTML manuals and e-books into it today. After getting about 2/3rds of the way through the process of importing all of my ebooks (appr. 50 large ebooks, all linked HTML files) I was moving some groups around in DT and DT suddenly quit on me.
When I restarted the application, DT is now telling me that I have 1 major file error, and that I need to delete all of the DT databases and reimport all of my files again.
Considering that it took me almost two hours to get this far, this is really frustrating. I started to reimport again, and about halfway through this time I had the same problem - suddenly quit, and then database error.
I’m importing each ebook separately, so it’s not like I’m importing a ton of files all at once. Still, I would think that DT’s database would be robust enough that it could handle a "suddenly quit’ scenario.
Is there anything I can do to either speed up the import process? Or am I basically stuck hoping that DT won’t quit on me?
sorry for the inconvenience. Could you provide us some crash logs (or even better - the data you’re trying to import)?
Basically it should (!) be impossible that DT crashs while updating its database files since v1.7.3 (as DT checks the memory structures BEFORE writing them back to disk). In addition, please verify your file system (the last two similar reports we’ve received were related to corrupted file systems).
On restarting DT said that it had a file error. It told me to export the lot, delete the database and import the lot (which I did; foolishly, because everything seemed ok in spite of the error message).
After importing I can now read the left pane (in Notepad view) as the entries there are now the exported folder and file names. Glossary has become Glossary.rtf and CatsPlaying has become xsry213.jpg - which is a right royal pain in the proverbial.
But worse, the database content is unrecognisable; i.e. entry Lecture 1.rtf should read:
Lecture 1: Phantoms in the Brain Until recently,[…]
it has become:
Lecre1: nineBrin\Unilrecenl,[…]
and:
Science can now offer empirical evidence for answering ancient philosophical questions about meaning and being.
has become:
Sciencecnnffereiricleiencenerinncieniliclqeineninein.
Did you backup the database (as advised) before exporting & deleting the old files? Then please send us the files (use our FTP server <ftp://ftp.devon-technologies.com/incoming>) and we’ll check this. But to fix the first problem, it’s only necessary to set the option “Preferences > Import > Titles” to “Strip extension”.
Probably the frozen computer caused the problem - e.g. if DEVONthink was updating the database at that moment or there was some critical data still in the disk cache of OS X but not yet on the disk. And as soon as DT recognizes more or less major issues, all database updates are disabled to prevent further problems (as it’s no good idea to modify an already damaged database).
Just send us the database (use our FTP server <ftp://ftp.devon-technologies.com/incoming>) and we’ll send you a fixed database within the next few hours if (!) possible. Another possibility would be to export all contents (select all first), then delete the DEVONthink folder (~/Library/Application Support/DEVONthink) and import the contents again. But please backup this folder first (e.g. create a copy using the Finder).
Or if there are backup files (with a ~ at the end) within this folder, then create a copy of the folder (see above), delete the four database files, rename the backups (just remove the ~ character), launch DT and choose "Verify & Repair" to check if everything is okay.
Finally, please verify your filesystem too - there are often a lot of errors after a forced reset.
Please have a look at the contents after reimporting them - if something should be missing or damaged, it’s probably better to send us the stuffed/compressed original DEVONthink folder.
By the way - the next release will introduce a new database format: More reliable (DEVONagent uses this format already for many months), smaller databases, using less memory and speeding up lots of things.
I might caution you to realize that importing a bunch of linked HTML files is a bit futile unless you are only interested in searching the text. DT only navigates among files external to the DB. For example, if you import all of the files to the DB, then delete the folder from your hard drive, the hyperlinks among the documents no longer function because they are referenced to a now missing folder.
I imported a huge hyperlinked archive of a set of books, the Pali Canon. When I clicked among links in the DT database, it opened the copies from my local hard drive and did not link internally to the database. What I discovered was that, apart from searching the text of those files, I needed to only import the start page from the set.
This will be definitely improved (probably in v1.9). In addition, a future version of DT will become a complete offline archive (automatically downloading/storing/referecing external material like style sheets, scripts and images too).