[digiKam-users] Segmentation fault when parsing entire collection
Dougie Nisbet
dougie at katsura.uk
Fri Jan 3 20:20:15 GMT 2020
Enjoying playing with Face Detection. I think I'm getting the hang of it
now.
If I try scanning my entire collection it faults at some point. e.g.
Digikam::DImg::load: "/store/media/images/2007/06/29/Jack Scott -
Houghall Arb -- Fri 29 Jun 2007 16-56-12 BST.jpg" : "JPEG" file
identified
Digikam::DNNFaceDetectorBase::selectBbox: Good rect = QRect(65,108
26x40) , conf = 0.858195
Digikam::DetectionWorker::process: Found 1 faces in "Jack Scott -
Houghall Arb -- Fri 29 Jun 2007 13-31-10 BST.jpg" QSize(2048, 1536)
QSize(2048, 1536)
Digikam::DMetadata::getIccProfile: Exif color-space tag is sRGB.
Using default sRGB ICC profile.
Digikam::DNNFaceDetectorBase::selectBbox: Good rect = QRect(160,116
56x82) , conf = 0.992359
Digikam::DNNFaceDetectorBase::selectBbox: Good rect = QRect(87,135
27x45) , conf = 0.988683
Digikam::DetectionWorker::process: Found 2 faces in "Jack Scott -
Houghall Arb -- Fri 29 Jun 2007 16-45-12 BST.jpg" QSize(2048, 1536)
QSize(2048, 1536)
Digikam::DatabaseWriter::process: Removing old entries ()
Digikam::FacePipeline::Private::checkFinished: Check for finish: 7
packages, 0 infos to filter, hasFinished() false
/tmp/.mount_digikadpkQOx/AppRun: line 151: 13672 Segmentation
fault digikam $@
I'm not too alarmed about this. My collection is 100K+ and is not
perfect, so quite possibly problems with my photos.
I've not noticed whether it faults on the same photo each time. I'll try
it again overnight tonight and find out.
The file: /tmp/.mount_digikadpkQOx/AppRun does not exist so no useful
information in there.
As I say this isn't a problem for me. Just part of being in beta. I've
just manually scanned the last photo mentioned above before the crash
and it was fine.
If I can provide more info please let me know.
Dougie
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/digikam-users/attachments/20200103/930b328e/attachment-0001.html>
More information about the Digikam-users
mailing list