|
russ_becker
|
 |
« on: June 20, 2017, 01:02:48 PM » |
|
Has anyone else had these issues on Windows 10 x64 ENTERPRISE? Most of the time when I use QU there is either a lockup or a crash. It almost invariably crashes upon exit. I upgraded from my old QImage in order to get drag/drop to work with IMatch. That usually works, but if QImage points to a folder other than the one the dropped file is in, more than half the time I receive a pointer access violation error. Sometimes I can go on to print, but sometimes not. Invariably there will be a lockup.
This is the .1607 build of Windows 10, i.e. the Anniversary Update. 32Gb system memory.
|
|
|
|
|
Logged
|
|
|
|
|
Fred A
|
|
« Reply #1 on: June 21, 2017, 09:15:10 AM » |
|
Invariably there will be a lockup. Hi Russ, After the lockup, how do you close QU? Task Manager? If so, you then restart QU and you get a message asking whether you wish to restart normally, or in Safe Mode? What version of Qimage is currently running? Fred |
|
|
|
|
Logged
|
|
|
|
|
russ_becker
|
|
« Reply #2 on: June 21, 2017, 01:53:19 PM » |
|
Yes, I shut it down with the Task Manager several times. Upon restart, it asked for a settings reset which I declined. After several iterations of this, I launched my old QImage Pro from 2009 and it had no issues browsing folders or printing a few things. But drag/drop from IMatch doesn't work with that version.
I had thought there may be some interaction with Affinity Photo, which I had used to open/export/close a number of image files (I have run into memory leaks before with Affinity Photo). So I shut down Affinity and tried out QImage Ultimate again. Drag/dropped a few files, printed a couple, and then got the INVALID POINTER message. At that point I closed QImage Ultimate and it crashed on exit with the usual 'windows is searching for a solution' message.
|
|
|
|
|
Logged
|
|
|
|
|
russ_becker
|
|
« Reply #3 on: June 21, 2017, 02:56:48 PM » |
|
Follow up: just realized I actually running 2017.126; I had noticed the 2018.101 in the titlebar and assumed that was the version in use. I'll let you know what happens once I apply the update.
|
|
|
|
|
Logged
|
|
|
|
|
Fred A
|
|
« Reply #4 on: June 21, 2017, 05:01:31 PM » |
|
OK Russ, I actually running 2017.126; I had noticed the 2018.101 in the titlebar and assumed that was the version in use. I'll let you know what happens once I apply the update. Ok Russ, Let's sort some of this out. Let's talk about an ini file (initialization file) Programs use this file to remember and recall settings from the last time you used it. That having been said, older programs like Qimage Pro and Studio each have their own ini files. So if they were fine, all would be fine. When you open Ultimate, it uses qimage-u.ini. Next item.... suppose, that that qimage-u.ini file was corrupted. Everytime you opened Qimage Ultimate it would use that ini file. Therefore, it is likely that any corrupted data would be called upon every time you opened QU. So, based on that strong possibility, next time you open Qimage Ultimate to a message : NORMAL or SAFE MODE, (new message from 2018.101), Choose Safe mode. Rusn some work through it and if all seems OK, Close Qimage Ultimate. This will make a NEW ini file, and we can go on from there. Fred |
|
|
|
|
Logged
|
|
|
|
|
russ_becker
|
|
« Reply #5 on: June 29, 2017, 11:08:18 PM » |
|
Follow up. Restarted QImage Ultimate 2018.101 and chose SAFE mode. Application reset its user interface to the defaults, and chose the system default printer, which is not the photo printer. So I recalled the appropriate Printer Setup job and all was back to normal. Printed a few things.
Just today I drug an image file from IMatch onto my QImage Ultimate favorite, and it opened up to the image folder. I then drug that image file onto the PRINT page and received the dreaded INVALID POINTER error. I tried to close QImage, no response, and then eventually received the exception failure dialog and Windows terminated the process. So back to square zero; the old QImage Pro works and QImage Ultimate isn't reliably working.
|
|
|
|
|
Logged
|
|
|
|
|
admin
|
|
« Reply #6 on: June 30, 2017, 06:35:31 AM » |
|
Given that you are the only one having the problem, it sounds like we are dealing with some sort of image corruption here. If you can narrow it down to one particular image that always causes the problem when you try to use/print it in QU, it would help if you could provide a link to that image. QU reads more data from the image headers than the old Qimage Pro so it's possible that Qimage Pro just doesn't try to utilize the bad data in the suspect image(s).
Mike
|
|
|
|
|
Logged
|
|
|
|
|
russ_becker
|
|
« Reply #7 on: June 30, 2017, 11:44:12 AM » |
|
Well, I would like to be able to pinpoint the problem to one or more JPG files but it seems to occur with all of them, randomly.
Just this morning I both started QImage Ultimate by drag-drop from IMatch and by double-clicking the desktop icon, looking at different folders. As soon as I clicked on a thumbnail I received the INVALID POINTER error message, followed by a crash, i.e. no images were in the print queue. Then I went to a large folder (1700 images) of B/W scans. Used drag-drop to start QImage, waited for the thumbnails to rebuild, and had no problems, printed several B/W images (these are all to a Canon Pro 9000 MkII) for testing purposes. I then closed QImage, went to a folder of pictures from yesterday (all JPG processed out from RAW using PhotoNinja and then resized using FasttStone), switched to Red River Aurora Satin paper, chose a couple of images, sized for both on the same page, and printed them -- no problems. Went I exited QImage it crashed with only the standard Windows crash dialog.
This is all on the .1703 build now of Windows 10 ENTERPRISE.
|
|
|
|
|
Logged
|
|
|
|
|
admin
|
|
« Reply #8 on: June 30, 2017, 11:59:49 AM » |
|
Just this morning I both started QImage Ultimate by drag-drop from IMatch and by double-clicking the desktop icon, looking at different folders. As soon as I clicked on a thumbnail I received the INVALID POINTER error message... Thanks for the info... If it's bad images, you may get a crash just from hovering the mouse cursor over a thumbnail because it reads the image headers as you roll the mouse cursor over thumbs (to display the EXIF info on the bottom status bar). I would suggest going back to that folder (above) and just moving your mouse over some images to see if one (or more) cause a crash. Clicking on/selecting a thumb does nothing but highlight the thumbnail: there are no pointer operations in that code. So I suspect what happened was that you selected a thumb and then moved your mouse over the offending (possibly corrupted) image which caused the crash. Mike |
|
|
|
|
Logged
|
|
|
|
|
russ_becker
|
|
« Reply #9 on: June 30, 2017, 01:12:29 PM » |
|
Did as instructed, found one image that consistently generated the INVALID POINTER error message when moving the pointer over it. This image, like the others, had been resized to final dimensions using FastStone 6.0 (1500x 2100). If you generate this message a sufficient number of times, you will get a lockup.
As an experiment, I opened this image with Affinity Photo, added a Vibrance layer, boosted the vibrance by 10 per cent just to change something, then exported it back as a JPG, with the same filename (not a real important picture). Opened up QImage to that folder, hovered the pointer over one image after another, no issues -- including the former culprit.
I did note a significant difference in file size between the exported Affinity Photo JPG and the FastStone JPG. I'll do some further tinkering with this, but I think you may be on to something. When I find another problem image I'll pass it along.
|
|
|
|
|
Logged
|
|
|
|
|
admin
|
|
« Reply #10 on: June 30, 2017, 05:43:19 PM » |
|
OK, that's progress. Let me know if you can email me or send me a link to a file that shows the error message when hovering the mouse cursor over it. I only need one image as an example: I suspect something you are using is creating images with corrupted headers so the ones that have this problem are all probably corrupted in the same way. I may be able to work around the problem if I have a sample image that causes the issue.
Thanks,
Mike
|
|
|
|
|
Logged
|
|
|
|
|
russ_becker
|
|
« Reply #11 on: June 30, 2017, 11:43:43 PM » |
|
I have been able to reproduce the issue. Went back to the original CR2 file, used PhotoNinja to create a new JPG using the current settings. That new image, before FastStone resizes it, causes the access violation. If I use FastStone to resize it to 2100x1500, I still get the access violation, so it must be an original metadata field/tag.
FWIW just set up a Canon Pixma Pro-100 on the network, modified the existing QImage printer jobs with the new printer driver and color profiles (other settings are basically the same as for the Pixma 9000 Mk2). Works flawlessly.
Where do you want me to email the problem file?
|
|
|
|
|
Logged
|
|
|
|
|
admin
|
|
« Reply #12 on: July 01, 2017, 01:59:42 AM » |
|
Where do you want me to email the problem file?
You can email it to mchaney@ddisoftware.com if it is under 15MB. If more than that, you could put it on Google Drive or use WeTransfer and just send me the link. Thanks, Mike |
|
|
|
|
Logged
|
|
|
|
|
admin
|
|
« Reply #13 on: July 01, 2017, 10:02:35 PM » |
|
Thanks for emailing the image. I was able to find the problem. It looks like Photo Ninja is inserting a very long string for "SubSecTimeOriginal". Typically that is either a two or three digit number (in string format) representing the hundredths (two digits) of a second or thousandths of a second (three digits) part of the EXIF shot date/time. Photo Ninja has a string value of "829999936" for SubSecTimeOriginal, meaning that the original date/time added by Photo Ninja for the image you sent would be 2017:06:28 07:22:31.829999936.
While that Original shot date of 2017:06:28 07:22:31.829999936 isn't technically "wrong" since "SubSecTimeOriginal" is an ASCII string, it's certainly odd to list subsecs in billionths of a second: the author of the third party code that I use for EXIF reading probably never ran into software that listed more than 2 or 3 digits for the decimal subsecs so he never noticed that his code had a bug where he should have been calling "Min" instead of "Max" when doing a memory move to a preformatted 3 character string. :-)
Qimage doesn't list date/time finer than the second but I fixed the code anyway which gets rid of the invalid pointer problem. I'll release the fix in 2018.103 soon.
I've attached a screen capture from ExifTool showing the subsecs down to the billionth of a second recorded by Photo Ninja.
Regards,
Mike
|
|
|
|
|
Logged
|
|
|
|
|
admin
|
|
« Reply #14 on: July 01, 2017, 10:52:04 PM » |
|
OK, the fix is out in 2018.103.
Mike
|
|
|
|
|
Logged
|
|
|
|
|
Qimage registration expired? New lifetime licenses are only $59.99!
