paitkenhead wrote:
You can add me to the list of people having this problem. Brand new out of the box 24" iMac with 40gb photo library. Was working fine initially, but then found I was having the same problems as those mentioned above, i.e. spinning gear forever in "may also be in the photos below".
Couple of things I've found to date:
1) Seems to be a problem with either my daughter or I. Faces for others work fine with the "may also be" feature working 100%, but each and every time I try the same thing for either my daughter or I it goes into spinning gears forever mode.
2) After entering spinning gears forever mode, the same thing happens for all other faces in the library. Quitting iPhoto will resolve the issue (for all other faces except my daughter and I), except if I quit in the "may also be" window for either of us.
Have tried rebuilding thumbnails / db etc. and now creating new library. Will report on progress.
Created new library by copying files from 'Originals' folder inside the corrupted iPhoto container. After hours of copying the photos, followed by hours of waiting for it to finish finding all the faces, my first attempt to tag a face resulted in the 'may also be in the photos below' error.
I think the problem is in asking it to do so many cross-checks in one hit. With this in mind I went on to create (another) new iPhoto library - this time importing the library one directory at a time (from 10 to 1000 photos in each directory). As I bring each folder in I've allowed it to find the faces, and using a "FACE IS unnamed" smart album I identify all my John Does.
I then either tag the each face or delete it. The result is that each time you bring in a new directory iPhoto only needs to check the new photos (1 directory worth) for matches. So far so good. While it sounds like a very long and drawn out process I have been able to bring in about 8GB of photos with no problems at all, and iPhoto remains quick and responsive throughout. I actually think that although this process involves more steps that it will in fact be quicker as it's much easier to identify and confirm faces due to a higher successful ID rate (mainly looking at photos of the same people from the one event) - so, so far so good.
Will confirm if I'm able to bring the whole 40Gb library in without problems using this method.