DB Diagnosis (1 Warning)

[Darius1968]

My database diagnosis (attached log file) came up with one warning, and when I search for w>, there is no warning to be found.  So, what am I doing wrong that I don't land on w>, and what is the warning associated with this database diagnosis? 

[Mario]

Search for warning. The diagnosis log does not use W>.
IMatch logged: Warning: 3 files marked as pending for write-back which have no data to write-back. Fixed.

Not sure how you've created 3 files marked as pending but without anything to write.
Do you recall what you did in the IMatch sessions since the last diagnosis run?

[Darius1968]

No.  That one totally beats me! 

[sinus]

If you edit a field in the metadata-tags, jumps to another field, goes back and changes the edite field again to the previous value, you have finally nothing changed.
But the pencil will be displayed.

This happens from time to time here, because I change e.g. a mistake in a word, go to the image, then I realise, that it was indeed correct before and edit again to the previous value.
So for me all is fine ... but I have an image tagged as write-back.

Could this be the case?

[Mario]

I tried that.

Change the Title tag from A to B.
Then changed it back to A again and then saved the changes in the MD panel.

The file is marked as pending and the pen tooltip shows that the Title tag needs to be written.
Which is the expected result.

This is correct, since IMatch does not keep a history of each  tag to tell if the user really made a change when editing the tag. It just marks the tag as modified in the MD panel when the user edits it.
And this marks the tag as modified when the user confirms the change in the MD panel.
And this sets the write-back flag.

Tip: Don't confirm your change in that case (click X) or use the Edit > Undo afterwards to undo your changes and reset the write-back state.

But all this is not related to the warnings the OP reported.
The file was pending but without any tag marked as modified so the "nothing to write" warning was logged.

This is not a problem, but I wonder how to create this situation so I can prevent it.
I have not seen this, so I asked if the OP remembered what he did to produce this.
Which steps to perform to mark a file as pending write-back, but without any metadata changed?

[sinus]

I see, then one, what I described, was the only situation, what I could imagine.

Like you, I have not idea, how can this happens otherwise.

[Darius1968]

Thanks for pointing that out, Markus. 

Yes, in this case, it could be applicable.  But, more often than not, when I do DB diagnosis, I get these metabase errors, and there have been no occurrences of what you describe. 

This is not a problem, but I wonder how to create this situation so I can prevent it.  I have not seen this, so I asked if the OP remembered what he did to produce this.  Which steps to perform to mark a file as pending write-back, but without any metadata changed?

I can't put my finger on it, but I'll keep an eye on anything I think could be the culprit. 

[ubacher]

I do get this error in the diagnostic file frequently. So much so that I ignore looking at the diagnostic file when there is an error reported. (That is why I, in a comment to another post, I suggested that the error lines be displayed immediately.)

I also get the case, when moving the pointer to the yellow pencil, it does not show what needs to be written back.
I ignored this since it seems a minor nuisance.

[Mario]

How to you produce this situation?

I do get this error in the diagnostic file frequently. So much so that I ignore looking at the diagnostic file when there is an error reported. (That is why I, in a comment to another post, I suggested that the error lines be displayed immediately.)

This is a warning, not an error.
When IMatch logs an error, you see a corresponding message when IMatch closes.