https://bugs.kde.org/show_bug.cgi?id=347937
Bug ID: 347937 Summary: Collections added to "Removable Media" don't stay there Product: digikam Version: 4.10.0 Platform: MacPorts Packages OS: OS X Status: UNCONFIRMED Severity: major Priority: NOR Component: setup Assignee: [hidden email] Reporter: [hidden email] Adding collections to the "Removable Media" category on OS X works as a "backdoor" to adding them to the "Local Collections" category. Reproducible: Always Steps to Reproduce: 1. Open the Preferences dialog, and go to the Collections panel 2. Use the "Add Collection" button next to the "Removable Media" category to add a folder to that category 3. Close the Preferences dialog 4. Reopen the Preferences dialog and go back to the Collections panel Actual Results: The collection (album) just added shows up in the "Local Collections" category. Expected Results: The collection (album) just added shows up in the "Removable Collections" category. I have not attempted to verify, but I have a hunch the same would happen with the "Network Shares" category. This is based on my assumption that this bug has something to do with the way OS X mounts everything under /Volumes as opposed to how Linux organises things. I'm labelling this a Major issue as basically it disables all support for removable media. -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
https://bugs.kde.org/show_bug.cgi?id=347937
Gilles Caulier <[hidden email]> changed: What |Removed |Added ---------------------------------------------------------------------------- Summary|Collections added to |Collections added as |"Removable Media" don't |"Removable Media" change as |stay there |"Local Collections" CC| |[hidden email] --- Comment #1 from Gilles Caulier <[hidden email]> --- I think it's more relevant of Solid do not work well under OSX. In fact Solid API is used here to register collection in DB... Gilles Caulier -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
--- Comment #2 from RJVB <[hidden email]> --- Hmm, if you can point me to where/how this is done in digikam's sources, I can have a look at Solid. But I don't really see the point in providing the categories through the settings dialog if afterwards you're only going to use an external library to determine whether a collection is local, on a removable or on a network share. Why not store the information how a collection was added and use it as an additional property (or to override whatever Solid tells you on OS X and potentially other platforms where Solid dysfunctions)? -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
Gilles Caulier <[hidden email]> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |[hidden email] --- Comment #3 from Gilles Caulier <[hidden email]> --- Marcel, Can you clarify/guide RJVB according with previous comment. Thanks in advance Gilles -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
--- Comment #4 from RJVB <[hidden email]> --- (In reply to RJVB from comment #2) > Hmm, if you can point me to where/how this is done in digikam's sources, I > can have a look at Solid. I can try to see what goes wrong in Solid when I know how it is being used by digiKam. > share. Why not store the information how a collection was added and use it > as an additional property (or to override whatever Solid tells you on OS X > and potentially other platforms where Solid dysfunctions)? Collections are added as `Local`, `Removable` or `Network`, and treated accordingly. Apparently this property isn't stored; instead you rely on Solid to keep track of which collections are on local, removable or network storage. If Solid isn't reliable in telling you if a collection directory is on local, removable or network storage, you should store that property with the other, non-hardware related properties (like the collection name). This would probably also make it easier to work with temporary collections, "scratch" folders that are on local storage but not always present. The same could apply to collections on "fancy" filesystems: I have no idea how Solid would recognise a ZFS pool or dataset for instance. There's an underlying question of principles here: should you use technology to second-guess the user, or should you just take it for given when s/he tells you "this collection is on removable media, treat it as such"? -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
--- Comment #5 from Marcel Wiesweg <[hidden email]> --- Most of the relevant code is in libs/database/collectionmanager.cpp You can find the UI part in utilities/setup/setupcollectionview.cpp Network shares are indeed treated differently, they are stored by mount path. Solid does not care for network shares, nor did any other system library in KDE or Qt ar the time. I guess it's still the same. The fact that there's an extra button to add removable media is due to UI design decisions, we need an extra button for network collections and there's no good way to have a common button for normal storage volumes. Apart from that, to digikam there is no real difference if you remove your USB stick or screw out your harddisk, in either case the storage volume is not available. There are two flags in Solid::StorageDrive, isHotpluggable() and isRemovable(), which determine the "removable" property. -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
--- Comment #6 from RJVB <[hidden email]> --- Thanks, I'll have a look. Right now, I'm getting errors about all collections on non-mounted removables when I start digikam. Are you saying that would still be the case if it knew those missing collections were on removables instead of considering them to be on fixed storage? -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
--- Comment #7 from Marcel Wiesweg <[hidden email]> --- What kind of errors do you get? Normally, the collection should silently be marked as not available. It's not an error if a removable drive is removed. -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
--- Comment #8 from RJVB <[hidden email]> --- I agree it's not an error, but as I described above, on OS X: 1) collections on removable storage are classified among the ones on a fixed harddisk 2) as a result, an error is printed that they're not available when that's the case, because digikam doesn't know it's not an error because of the removable nature of their support. That's why I suggested that the easiest and most straightforward fix that would work in all situations where Solid doesn't work reliably, would be to store the isRemovable state provided by the user when s/he adds the collection, rather than relying on Solid to reproduce it. Alternatively, remove the choice completely and let isRemovable be determined by Solid even when the collection is added. I don't know what it means that "there's no good way to have a common button for normal storage volumes". Either you rely on Solid to guess the nature of a volume (and you only need a separate button for networked collections), or you rely on what the user indicates (and you need a button for removable collections if you want to support that notion). Again, I'm a proponent of listening to the user. If you support the notion of removable collections (however small the difference with non-removable collections) and let the user indicate which is what, you don't second-guess this choice. You could ask confirmation if Solid tells you otherwise on platforms where this is reliable, and you could use subsequent changes as an indication that a disk was for instance moved to an external enclosure (and ask confirmation if this change should be recorded). But in my book, if the user tells you to consider a collection as removable (or non removable), you conform as long as this instruction does not lead to conflicts or impossible situations. But then I try to pay attention to the little details that annoy me when using someone else's software :) -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
--- Comment #9 from Marcel Wiesweg <[hidden email]> --- In any way there is now no compatible fix to store the "isRemovable" bit. I meant to say that the fact that there are three buttons is more or less deliberate, in some other UI design there would be no separate buttons. It's like two doors leading into the same room. IMO, Solid should be fixed on OS X. We have a set of high-level platform libraries and should in our applications not add complex workarounds for bugs in the underlying library. -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
--- Comment #10 from RJVB <[hidden email]> --- What do you mean with "no compatible fix"? That everything about the defined collections is stored in a fixed-format database which doesn't provide the flexibility to add a parameter? I've wondered why the list of collections isn't stored in a standard rc file too. Solid is incomplete on OS X, and I got dissuaded from spending much time on it IIRC. Not that that was hard to do: I'm not familiar with IOKit, and getting more of it into Solid looks to be writing a considerable amount of doubly navel-staring (pardon, introspective) code ... Are you sure that Solid on Linux recognises all kinds of removable storage correctly? How does it handle ZFS pools/datasets, for instance? For example, I have datasets on my ZFS root pool that are not always mounted. They live on my internal harddisk, though. Come to think of it, there isn't even need to consider uncommon filesystems: one could keeps photo collections on partitions of the internal disk that aren't always mounted; the MS Windows partition would probably a good example. -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
--- Comment #11 from Gilles Caulier <[hidden email]> --- New digiKam 4.11.0 is available with official PKG installer for OSX. https://www.digikam.org/node/740 Can you reproduce the problem with this release ? Gilles Caulier -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
--- Comment #12 from RJVB <[hidden email]> --- Indeed, as can be expected from the exchange above. -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
--- Comment #13 from Gilles Caulier <[hidden email]> --- digiKam 4.12.0 is out. https://www.digikam.org/node/741 Problem still reproducible ? Gilles Caulier -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
--- Comment #14 from RJVB <[hidden email]> --- On my ToDo list for when temperatures drop ! -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
--- Comment #15 from [hidden email] --- digiKam 5.0.0-beta7 Installer for MacOSX is now available for testing... https://drive.google.com/open?id=0B7yq-xFihT0_ZjlIazRPb1lRTEE Gilles Caulier -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
--- Comment #16 from [hidden email] --- RJVB, Do you seen my previous comment ? Gilles Caulier -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
--- Comment #17 from RJVB <[hidden email]> --- Yes, but I suppose the detection of removable vs. local devices is still done via Solid? That library hasn't changed to my knowledge so I didn't run out to check if this issue is still present .... -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
[hidden email] changed: What |Removed |Added ---------------------------------------------------------------------------- Component|setup |Database-Setup -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
In reply to this post by René J.V. Bertin
https://bugs.kde.org/show_bug.cgi?id=347937
[hidden email] changed: What |Removed |Added ---------------------------------------------------------------------------- Summary|Collections added as |SETUP : Collections added |"Removable Media" change as |as "Removable Media" change |"Local Collections" |as "Local Collections" -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
Free forum by Nabble | Edit this page |