Le 2018-10-10 à 15:08, Willy Williams a écrit :
I have three laptops on which Darktable is installed; two Windows 10
and one Ubuntu Linux. I generally sync the two Windows computers'
photos folders, including JPEGs, RAW files and XMP sidecar files.
I've begun to wonder if I'm shooting myself in the foot by doing so.
The question is this - are the sidecar files closely tied to an
internal Darktable database on the computer where the work was done?
Am I making the dog's breakfast of things by syncing files and in
doing so, inadvertently getting the sidecar files out of sync with the
unique database associated with each computer?
I think it should go without any problem. I would probably make sure
darktable or the sync program is fully closed before running the other,
just in case.
What sync program do you use ?
## Closely tied ? No fear, darktable has a sane model and reconcile
capabilities
Darktable maintains a copy of the data in its library from the XMP files
to allow finding pictures from the criteria in the XMP without
re-reading all the XMP.
Besides, darktable updates XMP files whenever you end an edit (switching
to another image, or switching from dark room back to light table), so
they are up-to-date with respect to the library which can be regarded as
a cache.
Still, darktable notices when XMP files are newer.
I've seen darktable ask me a number of times whether to import new
information from the XMP files. On virtually all occurrences, I clicked
on "check all" (or similar) and confirmed because XMP were indeed newer.
You may have a similar experience. I've seen no important ill-effect.
Still you might be interested in a similar case, read on.
## One solution to avoid any sync issue
Actually, my collection is so big that most of it is not mounted at the
time I run darktable. It is split among several hard drives.
And even part of it is mounted, the path to the storage change with
time. For example, a disk is sometimes mounted via USB and sometimes via
network, thus the mount point are different.
As a result, the per-image information in darktable library is full of
stale/unsynced copies from various mount points.
This is actually not much of a problem. It mostly causes any view of the
library to show duplicate entries and lots of skulls in lighttable.
Anyway, since the per-image information in darktable library is actually
useless to me, for about a year I've run darktable with --library
:memory: option.
This way, per-image information in darktable library is empty at each
start. I can't search per album of whatever but I don't care, I just
open darktable with the directory I'm interested in as argument. No
duplicate, no skulls, only the directory I'm working on.
This way, darktable never asks me about which version to keep (because
the "memory" library is just empty so never conflicting), and things go
smooth.
Here is the executable script sitting in my ~/bin/darktable :
#!/bin/bash
exec /usr/bin/darktable --library :memory: "$@"
Open to any comment, observation, similar experience or other solutions.
--
Stéphane Gourichon
____________________________________________________________________________
darktable user mailing list
to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
