Re: [Shotwell] Bug 775196 - Duplicate detection broken with RAW files - Proposed patch for review

[Abel <akronix5 gmail com>]

Hi Amos!

Unfortunately, there are still a bunch of bugs and requests about importing photos in shotwell.

In particular, consider #730780 (same device, different filenames but similar timestamps) which I've experienced myself and it entails same data loss issues and unreliable behaviours like the ones you've stated. There, Shotwell seems to consider the timestamps to flag some pictures as duplicates

A fix comprising, at least, both bugs should be researched in order to have a reliable "import photos" feature.

Saludos,

                                                       Abel.

On 28 November 2016 at 00:02, Amos Brocco <amos brocco gmail com> wrote:

Hello,

I reported a bug (https://bugzilla.gnome.org/show_bug.cgi?id=775196) concerning duplicate detection of RAW files during import. I found it in version 0.22.0, but I think It also affects previous versions as well as the latest from git (judging from the source code). I don't know if this list gets automatically notified about new bug reports, so here it is (if someone could review the bug & the proposed patch):

Current duplicate detection for RAW files works by comparing the filename and file size. Unfortunately when importing RAW files from an sdcard I noticed that Shotwell wrongly assumed that many files were duplicated and already stored in the library. Upon further examination I discovered that the criteria used for duplicate detection with RAW files are based only on the basename and file size.

Now I have and use three cameras that are exactly the same make and model (Sony A55), and I frequently have pictures generated on these different cameras at different times that share the same filename (something like DSC0XXXXX.ARW) and the SAME SIZE.

This means that Shotwell will import the first copy without issues but will then detect files from other cameras as duplicates when there's a match (name + filesize).

Now if the option to hide already imported files is set, Shotwell won't show me some of the images and will not import them. When I import hundreds of photos at once the fact that some of the photos are missing will get easily unnoticed (I typically don't check every single file, I just select to import all images). Unfortunately this way I might lose some pictures, thinking that all photos have been safely copied on my computer: in particular, if I format my sdcard all photos that were mistakenly considered as duplicates by Shotwell will be unrecoverably lost.

Note: I marked this bug as major because it can cause data losses to an end-user (as I've already experienced myself)

In attachment you will find a patch that should fixe this bug by comparing not only the filename and filesize but also the exposure timestamp.

thank you

Amos

_______________________________________________
shotwell-list mailing list
shotwell-list gnome org
https://mail.gnome.org/mailman/listinfo/shotwell-list