Has anyone every put a DEVONThink Database as a repo in a Git system?
I have a repo in Git locally that is linked to my repo on Gitlab that is a Scrivener project. It works so that I can edit the original scrivener document on my mac and those changes can be pushed to the GitLab project. I use GitKraken to do this. This all works because Scrivener files are really a package containing many other files.
Similarly since DT databases are also packages with the files that have been imported and lots of other stuff, I was thinking that there might be a way to create the same sort of system so that there is a local DT database that is linked tothe same thing on GitLab.
However, as a complete newby in using Git (I can make it workk but I don’t really grok it) and not really understanding the details of how the Scrivener Git starter project got developed I’m lost. The DT forums seem not to understand why anyone would want to do that so no help there.
I do all my query development in DT because It’s faster for me to edit SQL queries there. I want to keep them updated on GitLab as I develop more standard queries and modify them to handle new DB structures etc. without too much remember to do this update separately type stuff.
My queries can be pretty complex and I tend to get them developed and then comment and clean them up and then I want to share them without copy and pasting into text files.
Good thought, So I pulled out the most current version of my clean queries and created a repo here.
It appears to be working as long as I close the database in DT before trying to update the GitLab repo.
However, I’d love it if anyone who has DEVONThink could take a look perhaps fork or clone it and see if it’s actually working to pull down and run as a DT database. There are only about 25 actual queries in there all as separate files plus one that is a test of me adding a note to see if it gets updated
A DEVONthink database is more than a collection of files. The client manages varies concordances contained in the database package as well as records of other metadata. So, revising the database package (e.g., in a repo) without using the client is probably not possible without somehow reverse engineering the client and discovering the rules that control the concordance and metadata parsing. (Which probably violates a license - but who’s counting? )