https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #295 from timo...@zoho.com ---
Sorry for the slow reply. A happy new year to you too :)
Unfortunately I don't currently have access to the computer I usually used to
reproduce this issue. I'm not sure when exactly I'll have access again. I'l
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #294 from caulier.gil...@gmail.com ---
timokau,
First i wish you an happy new year 2020.
I would to ask you if it's possible to check the current implementation of
digiKam with the Similarities feature (Find duplicates).
I perform currently
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #293 from caulier.gil...@gmail.com ---
Ok, i will plan the release next week to be ready for next week end.
Gilles
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #292 from Maik Qualmann ---
Hi Gilles,
I think if the Beta1 appears shortly before or on the 24th, that would be a
good date ((:-))
Maik
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #291 from caulier.gil...@gmail.com ---
Congratulations Maik for your effort.
So we can plan a first official 7.0.0-beta1 release before Christmas.
Let's me hear the best date to plan...
Gilles
--
You are receiving this mail because:
Yo
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #290 from timo...@zoho.com ---
My pleasure! Thank you for staying on this for so long Maik. I'm looking
forward to digikam 7.0 :)
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=399923
Maik Qualmann changed:
What|Removed |Added
Resolution|--- |FIXED
Version Fixed In|
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #288 from timo...@zoho.com ---
I think that just might have fixed it. I just completed the 10th (!) successful
scan with 95594f2438a4f6f80340e70b20964e3685b4e235. I don't think I ever had
more than 3 scans without a crash before.
Since testi
https://bugs.kde.org/show_bug.cgi?id=399923
Maik Qualmann changed:
What|Removed |Added
CC||r...@gmx.de
--- Comment #287 from Maik Qualmann
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #286 from Maik Qualmann ---
Git commit 95594f2438a4f6f80340e70b20964e3685b4e235 by Maik Qualmann.
Committed on 11/12/2019 at 11:41.
Pushed by mqualmann into branch 'master'.
try with a QMutex for DImg data
M +4-0core/libs/dimg/dim
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #285 from timo...@zoho.com ---
Again after 3 successful runs, with 8229ef53b49b0c62244a96cb8f7d06a7452b4ce3:
> digikam.dimg: "/tmp/tmp.x7EaxS04Zv/Pictures/p/Kamera-Uploads/2014-03-14
> 12.09.33-2.jpg" : "JPEG" file identified
> digikam.gene
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #284 from Maik Qualmann ---
Git commit 8229ef53b49b0c62244a96cb8f7d06a7452b4ce3 by Maik Qualmann.
Committed on 09/12/2019 at 21:59.
Pushed by mqualmann into branch 'master'.
various changes to the loading task classes
M +7-7core/d
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #283 from timo...@zoho.com ---
With 8dd3f5099466dede99fd300e55db84fbff6dbad0 (from 2019-12-07):
> digikam.general: Try to get preview from
> "/tmp/tmp.848xADW9lH/Pictures/p/some-path/DSCN1115.jpg"
> digikam.general: Preview quality: 1
> di
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #282 from Maik Qualmann ---
Git commit 58e483a0a3bddc9cb339ab4b5548623a2d18bc3d by Maik Qualmann.
Committed on 05/12/2019 at 22:29.
Pushed by mqualmann into branch 'master'.
small changes on the preview task
M +44 -52 core/libs/thread
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #281 from timo...@zoho.com ---
With 43ef007a130c058982aaf61a900a0c95d99fd748, crashed on the 4th run after 3
successful ones:
> digikam.general: Cancel Main Thread
> [Thread 0x7fff22ffd700 (LWP 17652) exited]
> digikam.general: One job is do
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #280 from Maik Qualmann ---
Git commit a0774df21a8f039e3355c8ad3ffa0ca20924dbbf by Maik Qualmann.
Committed on 01/12/2019 at 17:11.
Pushed by mqualmann into branch 'master'.
latest DImg change has a memory leak
M +1-3core/libs/thr
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #279 from timo...@zoho.com ---
With aacced0438943b163ffa95a3e1d82d74e4d065ae:
> digikam.general: Preview quality: 1
> digikam.dimg:
> "/tmp/tmp.HClqSMevcG/Pictures/p/Syncthing-3TB/Telegram-Android/Telegram
> Images/238228297_1801.jpg" : "
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #278 from Maik Qualmann ---
Git commit aacced0438943b163ffa95a3e1d82d74e4d065ae by Maik Qualmann.
Committed on 01/12/2019 at 13:00.
Pushed by mqualmann into branch 'master'.
return new DImg from QCache
M +4-2core/libs/threadimagei
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #277 from timo...@zoho.com ---
With 330d3cea6749f1a50783bd0acd6ac57b95cf0519:
> digikam.dimg:
> "/tmp/tmp.CLpH7QxTNQ/Pictures/p/Android/Whatsapp/IMG-20140822-WA0033.jpg" :
> "JPEG" file identified
> digikam.general: Found 0 faces in "IMG-2
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #276 from Maik Qualmann ---
Git commit 330d3cea6749f1a50783bd0acd6ac57b95cf0519 by Maik Qualmann.
Committed on 30/11/2019 at 21:52.
Pushed by mqualmann into branch 'master'.
polish set loading status
M +7-3core/libs/threadimageio/
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #275 from timo...@zoho.com ---
Not sure how useful that debug information is, loading task status seems to
always be 0. With e69b8317e24ee6487b6a79a75052aab6e7024f43:
> digikam.general: Loading task status: 0
> digikam.general: Try to get p
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #274 from Maik Qualmann ---
Git commit 62ab5314c3b32a83c0b765b67c251ee2bd810b37 by Maik Qualmann.
Committed on 28/11/2019 at 21:53.
Pushed by mqualmann into branch 'master'.
add debug, we need the loading task status
M +2-0core/li
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #273 from Maik Qualmann ---
However, I see no error message that the memory could not be requested. Another
possibility is that the shared image data is changed by another thread at the
same time.
Maik
--
You are receiving this mail becau
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #272 from timo...@zoho.com ---
Created attachment 124172
--> https://bugs.kde.org/attachment.cgi?id=124172&action=edit
here's one additional trace taken during the last run
--
You are receiving this mail because:
You are watching all bug
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #271 from Maik Qualmann ---
The cause here might be that we did not get the requested memory as a block.
Hmm, yes we have to check if we got the memory, but what if not? Just do not
rotate? On the other hand, we have to look at where we lose
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #270 from timo...@zoho.com ---
I was using exiv2 0.27.1. I've updated to 0.27.2, disabled color management
(which was enabled in the other runs, since every run starts from a clean
digikam with default configuration). With
105ccfd47cc8ac73037
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #269 from caulier.gil...@gmail.com ---
It's clear. your last backtrace as nothing to do with face management.
The crash come from Exiv2 shared library when an ICC color profile is taken
from a JPEG image to render an image.
1/ did you have
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #268 from timo...@zoho.com ---
Created attachment 124143
--> https://bugs.kde.org/attachment.cgi?id=124143&action=edit
digikam resource usage with 105ccfd47cc8ac73037eac0d6873581a9e49c1aa
--
You are receiving this mail because:
You are wa
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #267 from timo...@zoho.com ---
With 105ccfd47cc8ac73037eac0d6873581a9e49c1aa:
> digikam.general: Try to get preview from
> "/tmp/tmp.CebnuTHvLs/Pictures/p/sth/P1020970.JPG"
> digikam.general: Preview quality: 1
> digikam.dimg: "/tmp/tmp.Ce
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #266 from Maik Qualmann ---
Git commit 33735b564c7cb1276d1a0588a518190300f83d31 by Maik Qualmann.
Committed on 26/11/2019 at 17:47.
Pushed by mqualmann into branch 'master'.
use QExplicitlySharedDataPointer for MetaEngineData
M +2-2
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #265 from Maik Qualmann ---
Yes, I think the crash is clear. We use a QSharedDataPointer, read access is
shared, write access calls detach(). I do not think he is able to correctly
clone the Exiv2::ExifData container. We have the getExifdata
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #264 from caulier.gil...@gmail.com ---
The backtrace from comment #262 is very interresting :
> #8 Digikam::MetaEngineData::Private::Private (this=0x7fff08086c80)
> at
> /tmp/tmp.m1S6vaoNjT/digikam/core/libs/metadataengine/engine/metae
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #263 from timo...@zoho.com ---
Created attachment 124119
--> https://bugs.kde.org/attachment.cgi?id=124119&action=edit
digikam resource usage with b50ee49d2455c936f02178d59389669e0796e506
--
You are receiving this mail because:
You are wa
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #262 from timo...@zoho.com ---
I tracked the memory usage with psrecord (attached). You can easily see the
point where it crashed, when the CPU usage dropped to 0 and the memory usage
flattened out.
With b50ee49d2455c936f02178d59389669e0796e5
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #261 from Maik Qualmann ---
Post of commits with CCBUGS seems broken. Please test this experiment:
https://invent.kde.org/kde/digikam/commit/7c50ae256ee7eb59e4c00dc101d7d7339a420ab9
Maik
--
You are receiving this mail because:
You are wa
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #260 from Maik Qualmann ---
It is only interested in what has changed in the memory usage.
Maik
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #259 from Maik Qualmann ---
Git commit c31a8b4e3650fe05392e01c2388f2f2847a191b4 by Maik Qualmann.
Committed on 21/11/2019 at 22:09.
Pushed by mqualmann into branch 'master'.
for a test we disable the QCache
M +2-2core/libs/threadi
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #258 from timo...@zoho.com ---
I kept a loose eye on my the ram usage during this run, seemed to hover around
~5-6G actual memory and 1-2G swap. The system definitely didn't OOM. I was
using it at the same time, and I should've noticed that.
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #257 from Maik Qualmann ---
Git commit ea529419b90d54ce5bc6052a14ec2c23c5f79b1f by Maik Qualmann.
Committed on 21/11/2019 at 16:58.
Pushed by mqualmann into branch 'master'.
various optimization for image loading
M +4-5core/libs/d
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #256 from timo...@zoho.com ---
Memory usage is pretty high. I have 8G of ram and swap space is still needed
for the scan to complete. It seems to me that the crash occurs more often when
I use the system otherwise at the same time, so it may
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #255 from Maik Qualmann ---
I think that in this case we still have 2 references to QSharedData. It crashes
in memcpy(). We do not check the memory allocation, whether we have received
it. Timokau, what is the memory usage during face detect
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #254 from caulier.gil...@gmail.com ---
Maik,
When i compile with Clang static ananlyzer 8.0 under Mageia 7.1 i can see that:
In file included from
/home/gilles/Documents/7.x/core/libs/dimg/dimg_data.cpp:26:
In file included from /home/gill
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #253 from timo...@zoho.com ---
With 6845671c7e6b60e337fae81b097d14bb32e0b6e8:
> digikam.general: Try to get preview from
> "/tmp/tmp.oF2HC2AYd3/Pictures/p/android_pics/DCIM/Camera/IMG_20171221_153250.jpg"
> digikam.general: Preview quality:
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #252 from Maik Qualmann ---
Git commit d235ea6eff1cff7a8240fd9ac1a67ce441aae940 by Maik Qualmann.
Committed on 20/11/2019 at 06:44.
Pushed by mqualmann into branch 'master'.
do not receive progress info in postProcess()
M +3-13 core
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #251 from timo...@zoho.com ---
With fe0020243fa27001b5daf575fb894663fdbe534b:
> digikam.general: Found 0 faces in "IMG_20171125_194502.jpg" QSize(3120, 4208)
> QSize(3120, 4208)
> digikam.general: Removing old entries ()
> digikam.general:
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #250 from Maik Qualmann ---
Git commit fe0020243fa27001b5daf575fb894663fdbe534b by Maik Qualmann.
Committed on 19/11/2019 at 17:35.
Pushed by mqualmann into branch 'master'.
check only for empty identities for face recognition
M +5-2
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #249 from Maik Qualmann ---
Yes, my mistake is with an empty face database. You would have to add a face by
name manually. I fix it...
Maik
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #248 from timo...@zoho.com ---
I cannot start face detection with ba2893d27dd4542034f90b56bc4d75f1cb727d42. It
gives me the error:
Face recognition is aborted, because there are no identities to recognize.
Please add new identities.
Even th
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #247 from Maik Qualmann ---
Git commit ba2893d27dd4542034f90b56bc4d75f1cb727d42 by Maik Qualmann.
Committed on 19/11/2019 at 07:10.
Pushed by mqualmann into branch 'master'.
fix wrong DImgLoaderObserver in postProcess()
M +3-3core
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #246 from Maik Qualmann ---
Git commit d7b1ea9ccdd65d9a7d85be0ed035b33f0f7d5d31 by Maik Qualmann.
Committed on 18/11/2019 at 19:49.
Pushed by mqualmann into branch 'master'.
prevent to access a QList simultaneously from multiple threads
M
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #245 from timo...@zoho.com ---
With b306f5aa14358dd04167b1d197de1fdc86e37bbc (on the 4th run):
> digikam.general: Try to get preview from
> "/tmp/tmp.sIBLW3QdU7/Pictures/p/Kamera-Uploads/2014-03-13 20.01.02.jpg"
> digikam.general: Preview q
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #244 from Maik Qualmann ---
Git commit f037b05d531c90f95d8f41b7f81d837ace6aeb5b by Maik Qualmann.
Committed on 17/11/2019 at 07:25.
Pushed by mqualmann into branch 'master'.
use for a test QExplicitlySharedDataPointer for DImg
M +2-2
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #243 from Maik Qualmann ---
Yes, I have already thought about that and switched in DImg to
QSharedDataPointer. I can apply the patch as a trial.
Maik
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #242 from caulier.gil...@gmail.com ---
Maik,
The problem from this file can be relevant of the long time problem located in
DSharedData container and never fixed.
Look this clang report:
https://www.digikam.org/reports/clang/master/
==>
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #241 from Maik Qualmann ---
Git commit da67056beb32c461251661e90d8f4c73f984ce26 by Maik Qualmann.
Committed on 16/11/2019 at 19:39.
Pushed by mqualmann into branch 'master'.
revert last commit, private DImg data is shared
M +6-6co
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #240 from Maik Qualmann ---
Git commit eb5a4a61fddaded66f07e57e44cd2d98862a1fa9 by Maik Qualmann.
Committed on 16/11/2019 at 19:12.
Pushed by mqualmann into branch 'master'.
fix ICC Manager, must use a DImg reference
M +9-8core/li
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #239 from timo...@zoho.com ---
With 2e3d03a2cea443f32d8a23c85393c345e8308e5b:
> digikam.dimg: "/tmp/tmp.7dg66TNVh9/Pictures/p/images/Telegram
> Images/223024373_562302.jpg" : "JPEG" file identified
> digikam.general: Found 0 faces in "22302
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #238 from Maik Qualmann ---
Git commit 6aaa6622d706972bf8a532c0a92ea721b11c7fa0 by Maik Qualmann.
Committed on 15/11/2019 at 21:41.
Pushed by mqualmann into branch 'master'.
do not add this task to the listeners here
M +3-3core/li
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #237 from timo...@zoho.com ---
With 47e8d093858421c1b95ea17acb396067ffccc8eb:
> digikam.general: Found 0 faces in "IMG_20150502_140220.jpg" QSize(3120, 4208)
> QSize(3120, 4208)
> digikam.general: Removing old entries ()
> digikam.metaengin
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #236 from Maik Qualmann ---
Git commit cab5995cacb2bfd97b304c2338b00e2ae176fdff by Maik Qualmann.
Committed on 14/11/2019 at 20:48.
Pushed by mqualmann into branch 'master'.
do not add this task to the listeners in the postprocess
M +2
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #235 from timo...@zoho.com ---
With 35339b4dbf2253dd1f60d3a374203d8717b05d86 (including Gilles' fix):
> digikam.dimg: "/tmp/tmp.0Tdwmd06eD/Pictures/p/Kamera-Uploads/2014-03-19
> 13.07.31.jpg" : "JPEG" file identified
> digikam.metaengine: E
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #234 from caulier.gil...@gmail.com ---
I commit this fix too :
https://invent.kde.org/kde/digikam/commit/ce0cac88500e3bcfc7219baa805dabb6d91688b1
WorkerObject class in used especially in face management pipeline for face
detection and recog
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #233 from Maik Qualmann ---
Ok, I think we are on the right way, because several attempts are needed to
reproduce a crash. Also this crash we have rarely seen here in this long
thread. I already have an idea...
Maik
--
You are receiving t
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #232 from timo...@zoho.com ---
This time (de875df40545d59cf038a0134c92377aed8d6187) two full scans succeeded.
The third one crashed:
> digikam.general: Try to get preview from
> "/tmp/tmp.pp4d0tIJXq/Pictures/p/Android/Camera/IMG_20140924_12
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #231 from Maik Qualmann ---
Git commit de875df40545d59cf038a0134c92377aed8d6187 by Maik Qualmann.
Committed on 13/11/2019 at 19:37.
Pushed by mqualmann into branch 'master'.
move static functions to DImg for more possible thread safe
M +1
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #230 from timo...@zoho.com ---
The scan came very close to finishing, then this (with
816d4e61b9cf9ddbaa41381c99e67e0c60ecae75, which came after your latest fix):
> digikam.general: Found 1 faces in "IMG_20171228_134932_1.jpg" QSize(1944,
>
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #229 from caulier.gil...@gmail.com ---
Note : i don't yet tested your last commit
#66346eddf9faafc0d8a16c603385e4b450a96968
Test was processed with this one : #c3476af439fa8351eee23f4477aaed8b2c00f56c
Gilles
--
You are receiving this mail
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #228 from caulier.gil...@gmail.com ---
With the mutex, it crash after a long time to detect faces (completed at 80%):
digikam.dimg: "/mnt/data/photos/TESTS/JPEG/Horizontal/cicada.jpg" : "JPEG" file
identified
digikam.general: Using 8 CPU
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #227 from Maik Qualmann ---
Git commit 66346eddf9faafc0d8a16c603385e4b450a96968 by Maik Qualmann.
Committed on 12/11/2019 at 21:46.
Pushed by mqualmann into branch 'master'.
first step to move static function to DImg
next step move static u
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #226 from Maik Qualmann ---
We'll wait and see what the QMutex changes. We have access to the
InfoProvider() in the static exifOrientation() function. This can not be easily
moved to DImg.
Maik
--
You are receiving this mail because:
You
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #225 from Maik Qualmann ---
Okay, then the best way to move these functions to the DImg class.
Maik
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #224 from caulier.gil...@gmail.com ---
Inter thread static function uses are definitively the hell. Protect with a
mutex, or make this function not static, included in DImg class for ex.
My crash is difficult to capture. Even is a core dumpe
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #223 from Maik Qualmann ---
Git commit c3476af439fa8351eee23f4477aaed8b2c00f56c by Maik Qualmann.
Committed on 12/11/2019 at 17:51.
Pushed by mqualmann into branch 'master'.
protect ExifRotate with a QMutex
M +2-0core/libs/threadi
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #222 from Maik Qualmann ---
This crash is one of many here in the thread. It crashes in a static function
to the Exif rotation, writing the DImg attribute into the QMap. The DImg is
alone at this moment and not shared with another thread. Pr
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #221 from caulier.gil...@gmail.com ---
Timokau,
I tested DNN face detection while this week end on my huge collection. It crash
also few time, and after restart it complete the job.
I investigate to found the bug on my computer, but i'm not
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #220 from timo...@zoho.com ---
That is definitely a lot faster! The scan finished once with a specificity of
1%. Then I set it back to 80% and it crashed as follows:
> digikam.general: Cancel Main Thread
> [Thread 0x7fff237fe700 (LWP 9711) e
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #219 from caulier.gil...@gmail.com ---
Git commit 85af2a46ebcd7b19ec3b6809c85f587dbbd14f2e by Gilles Caulier.
Committed on 11/11/2019 at 14:11.
Pushed by cgilles into branch 'master'.
Face Detection Neural Network optimizations:
Switch from
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #218 from timo...@zoho.com ---
Unfortunately turning down the accuracy doesn't impact the runtime very
significantly. With 2% instead of 100% it still didn't get very far (to 12%
scanned) in ~6h. It didn't crash either though, so maybe the c
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #216 from caulier.gil...@gmail.com ---
Timokau,
The long time to scan for face detection is reproducible here and is not a
dysfunction.
In fact it's just a setting to adjust (the level of detection). Currently, the
value is too low and here
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #217 from caulier.gil...@gmail.com ---
Maik,
about crash, i recommend to check new cppcheck reports around threadimageio,
dimg, and faceengine.
https://www.digikam.org/reports/cppcheck/master/
Look well the virtual calls in constructor war
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #215 from timo...@zoho.com ---
On a902eddf0d84b1561e7c7ee75164e58fdeaf8672 face detection takes unreasonably
long again, I assume because the new face recognition branch was merged in. I
left it running for some 10 hours and got to 25%, and i
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #214 from Maik Qualmann ---
Git commit a902eddf0d84b1561e7c7ee75164e58fdeaf8672 by Maik Qualmann.
Committed on 31/10/2019 at 10:51.
Pushed by mqualmann into branch 'master'.
use DImg as pointer in static functions
M +2-2core/dplug
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #213 from timo...@zoho.com ---
With 707891ca729fa69c1cbcad6c133f0c5585773747:
> digikam.metaengine: Exif color-space tag is sRGB. Using default sRGB ICC
> profile.
> digikam.general: Try to get preview from "some_image.JPG"
> digikam.genera
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #212 from caulier.gil...@gmail.com ---
Git commit 707891ca729fa69c1cbcad6c133f0c5585773747 by Gilles Caulier.
Committed on 30/10/2019 at 09:56.
Pushed by cgilles into branch 'master'.
Cppcheck is right here : we have 2 declarations of "send"
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #211 from caulier.gil...@gmail.com ---
Git commit dbcaeafec2ef2627dbf61095b94d7385ec9107f3 by Gilles Caulier.
Committed on 30/10/2019 at 09:47.
Pushed by cgilles into branch 'master'.
Cppcheck is right here : we have 2 decalration of i shado
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #210 from caulier.gil...@gmail.com ---
Git commit a3614106b068d39ed820af7bc06f3036e8799d60 by Gilles Caulier.
Committed on 30/10/2019 at 09:20.
Pushed by cgilles into branch 'master'.
fix out of bound index reported by cppcheck
M +5-2
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #209 from timo...@zoho.com ---
My reproduction script is as follows:
https://gist.github.com/timokau/a50d1033bce78f56804406bb8585b607
The first couple of lines are to set up an clean environment in a tmpdir for
the build. The `nix-shell` li
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #208 from Maik Qualmann ---
Ok, here is the crash that should never occur, we are in the wrong overriden
function. The crash in the static function is already strange. Although static
functions with threads are considered critical, no proble
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #207 from timo...@zoho.com ---
One full scan actually went through, but on the second try:
> digikam.general: Found 2 faces in "2012-08-31 18.24.29.jpg" QSize(640, 480)
> QSize(640, 480)
> digikam.general: Removing old entries (FaceTagsIfac
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #206 from Maik Qualmann ---
Git commit 6a77ab5528a85bdf1bb0d5f30bb70f928d99c7b7 by Maik Qualmann.
Committed on 28/10/2019 at 17:32.
Pushed by mqualmann into branch 'master'.
move these two static functions for a test in PreviewLoadingTask
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #205 from timo...@zoho.com ---
Okay, setting that environment variable appropriately fixed that issue. So
here's the latest crash (with 27f663c1155498c03c174ce93d7c94e69d77489a):
> digikam.facesengine: detectMultiScale: image size ( 52 , 52
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #204 from Maik Qualmann ---
OK, JPEG is an unknown format. Now the log would be good from the beginning.
With digiKam-6.4.0, the image loaders will now also be plugins. These plugins
are probably not found in your installation. The search is
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #203 from timo...@zoho.com ---
I doesn't look very helpful to me, but here's the full log of the scan
(completed in 7 seconds): http://sprunge.us/fVfY1j
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #202 from Maik Qualmann ---
In principle, it should be a clean startup configuration. It is difficult to
say why the images are not processed. Maybe it helps to post the console log.
Maik
--
You are receiving this mail because:
You are wa
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #201 from timo...@zoho.com ---
I set $HOME to a tmpdir for each test to avoid carrying over configuration and
I deleted all databases in the picture dir. Is there anything else that could
cause this?
--
You are receiving this mail because:
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #200 from Maik Qualmann ---
Presumably it's because you tested the developer branch with the new DNN face
recognition. Check the facial recognition settings and use a new
recognition.db.
Maik
--
You are receiving this mail because:
You ar
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #199 from timo...@zoho.com ---
With 27f663c1155498c03c174ce93d7c94e69d77489a the scan finishes in just a few
seconds, seemingly doing nothing.
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #198 from Maik Qualmann ---
Git commit 27f663c1155498c03c174ce93d7c94e69d77489a by Maik Qualmann.
Committed on 19/10/2019 at 16:28.
Pushed by mqualmann into branch 'master'.
review loading tasks
Related: bug 406019
M +42 -37 core/libs
https://bugs.kde.org/show_bug.cgi?id=399923
--- Comment #197 from timo...@zoho.com ---
Sorry for the delay. The test is still running.
One thing I can say for sure is that with
04179402304ee8d9d813d5761dabf3ea8c769048 the performance has degraded a lot. It
takes up to 10G of ram at times (4G on a
https://bugs.kde.org/show_bug.cgi?id=399923
Maik Qualmann changed:
What|Removed |Added
CC||gen...@moin.fi
--- Comment #196 from Maik Qualm
1 - 100 of 296 matches
Mail list logo
