I tried to write IPTC metadata to JPG files and I got
No original files were modified
Aperture was not able to modify one or more original files because they have a format that does not permit modification.
i found some old posts saying this was broken in Lion.
Anyone have any ideas?
Which version of Aperture are you using?
Are your JEGS referenced? You can receive the same error message, if the referenced originals cannot be found. Try to export the original file to check, if it is available.
The very latest version, they are aperture managed (not referenced).
The JPGs were produced by ScanSnap (bulk print scanning), they may be peculiar.
On the other hand after I exported them as versions and tried writing metadata Aperture crashed on me (Yosemite). i think some data was being written, so that's probably just Aperture being buggy. I suspect Aperture's software was in a very bad state when Apple killed the product, it smells like it had been given to interns years ago.
See my edits above. If Aperture is losing track of the originals (for managed or referenced images) you will get the error panel above, when you try to write the meta data to the originals. SO check, if the originals are available by trying to export or edit them.
And Aperture will give the same error message for some scanned images, if you want to write EXIF tags that are missing in the scans. Some scanned images do not have a capture date tag, for example. And then it is not possible to write the capture date tag to the original. I am usually getting the same error message for BMP files, if they are scans.
I exported them and imported into another Aperture database as new JPEG versions. There the IPTC write seemed to initiate, but Aperture crashed reliably (even after system restart, full power cycle peripherals).
Crashed Thread: 0 Dispatch queue: com.apple.main-thread
Exception Type: EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x000000000ca286c0
I'd rebuilt the database on that repository.
Lots of ugliness in both Yosemite and Aperture; I'm afraid Apple delegated both to interns. i do think the original error message was due to the Fujitsu ScanSnap JPGs missing the necessary EXIF headers.
PS. Most reliable crasher I've ever seen with Aperture, and I've seen quite a few. I am going to run diagnostics on the drive.
Which version of Yosemite are you running? I did not have any Aperture crashes after the last Yosemite update.
Very latest Yosemite. I was able to get IPTC writes to work in some other projects, so i'm going to try partitioning this project into albums of 20 images and see if it's one particular image causing the crash. My Library is 42,000 photos and 325GB, wonder if i'm drifting into data store limits.
My Library is 42,000 photos and 325GB, wonder if i'm drifting into data store limits.
You can rule that one out anyway. My last Aperture Library was one third as big again and at least one regular poster on here had a library with ten times the number of items. I think the issue is most likely related to the source of the images. many scan simply do not have the metadata space.
Good to know about size.
Scan JPGs caused the original write failure, but the Aperture crashes with IPTC writes is something else.
I got tired of crashing my primary library so I exported the problem project as a new Library and tried IPTC write. This time I got "One or more original files were not modified .."
That's curious indeed. At least now that I have a standalone Library version I can play with the bug safely.
have you tried to repair the permissions on your test library?
Yes, though repairing permissions on a package is weird. You have open package, move everything into a folder, set permissions to propagate, then move them back out.
It's not permissions though. This is really weird bug. I have 266 images in my test library. I tested on multiple drives and with both kinds of library repair. I partitioned into sets of 20-40 images.
Sometimes I could write successfully except for files that couldn't be written to. Then I'd repeat and get "one or more original ... locked". Then repeat and get the hard crash. Then repeat and get another crash.
Next up -- trying to do this from a different user account on the same machine.
Yes, though repairing permissions on a package is weird. You have open package, move everything into a folder, set permissions to propagate, then move them back out.
Don't use the Finder to repair permissions. Aperture has its own tool to repair the permission. Using the Finder will only make it worse.
Try this, even if you already repaired the database or rebuilt it. For some weird reason repairing the permissions is not automatically included when you are using the other repair options.
In all my years of Aperture use I must have become so accustomed to this dialog I just read past the permissions part. Thanks for picking that up. I'll give it a try and retest on the toy Library. I'm glad I have good Aperture backups -- though sometimes I'm not sure that's even possible. (With apps that use databases a restore is very hard if you miss that something has gone wrong and keep using the app.)
Repairing permissions and rebuild database didn't help -- still the unpredictable crashers as noted above.
Only thing left to try is another machine -- just for the heck of it.
It's very weird bug.
Moving to another machine didn't change anything, the test Library crashes with IPTC writes. So now I don't need to worry about some other app running and contributing to the crash. Now I can just play with the test library. I'm going to start inspecting the package components. What I really need is to put some kind of debugger on the process but that's beyond my skill set. Apple isn't doing any Aperture fixes, but it would be good to understand the crash.
My hunch is that there's a bug with date handling. These images have a date in 1986, the date was written into them when I exported versions from the original. I suspect the IPTC writing code has a memory access bug triggered by the dates. The unpredictable behaviors suggests the bug is writing to random bits of memory -- sometimes crashing Aperture, sometimes triggering misleading error messages, occasionally completing.
Unable to write IPTC metadata to JPG files
