Captures deleted when output to /tmp - by: TallFurryMan

I stumbled upon this issue while testing my changes to the scheduler.
I created a simple sequence file, with a single 1-second luminance capture, configured to store to /tmp/kstars_tests.
I observed that every time the sequence was executed, a capture would me made, and the previous capture would be immediately deleted.
Running KStars with a syscall filter gave:

tallfurryman@LubuntuEkos:~/dev/kstars_build$ strace kstars 2>&1 >/dev/null | grep '/tmp/kstars_test'
inotify_add_watch(26, "/tmp/kstars_tests", IN_MODIFY|IN_ATTRIB|IN_CLOSE_WRITE|IN_MOVED_FROM|IN_MOVED_TO|IN_CREATE|IN_DELETE|IN_DELETE_SELF|IN_MOVE_SELF|IN_UNMOUNT|IN_ONLYDIR) = 5
openat(AT_FDCWD, "/tmp/kstars_tests/1x1s_Lum.esq", O_RDONLY|O_CLOEXEC) = 41
mkdir("/tmp/kstars_tests/Light/Luminance", 0777) = -1 EEXIST (File exists)
stat("/tmp/kstars_tests/Light/Luminance", {st_mode=S_IFDIR|0775, st_size=4096, ...}) = 0
open("/tmp/kstars_tests/Light/Luminance", O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC) = 41
stat("/tmp/kstars_tests/Light/Luminance", {st_mode=S_IFDIR|0775, st_size=4096, ...}) = 0
openat(AT_FDCWD, "/tmp/kstars_tests/Light/Luminance/Light_006.fits", O_WRONLY|O_CREAT|O_TRUNC|O_CLOEXEC, 0666) = 41
openat(AT_FDCWD, "/tmp/kstars_tests/Light/Luminance/Light_006.fits", O_RDONLY) = 47
openat(AT_FDCWD, "/tmp/kstars_tests/Light/Luminance/Light_006.fits", O_RDONLY) = 47
openat(AT_FDCWD, "/tmp/kstars_tests/Light/Luminance/Light_006.fits", O_RDONLY) = 48
openat(AT_FDCWD, "/tmp/kstars_tests/Light/Luminance/Light_006.fits", O_RDONLY) = 48
mkdir("/tmp/kstars_tests/Light/Luminance", 0777) = -1 EEXIST (File exists)
stat("/tmp/kstars_tests/Light/Luminance", {st_mode=S_IFDIR|0775, st_size=4096, ...}) = 0
open("/tmp/kstars_tests/Light/Luminance", O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC) = 54
stat("/tmp/kstars_tests/Light/Luminance", {st_mode=S_IFDIR|0775, st_size=4096, ...}) = 0
openat(AT_FDCWD, "/tmp/kstars_tests/Light/Luminance/Light_007.fits", O_WRONLY|O_CREAT|O_TRUNC|O_CLOEXEC, 0666) = 54
unlink("/tmp/kstars_tests/Light/Luminance/Light_006.fits") = 0
openat(AT_FDCWD, "/tmp/kstars_tests/Light/Luminance/Light_007.fits", O_RDONLY) = 47
openat(AT_FDCWD, "/tmp/kstars_tests/Light/Luminance/Light_007.fits", O_RDONLY) = 47
unlink("/tmp/kstars_tests/Light/Luminance/Light_006.fits") = -1 ENOENT (No such file or directory)
openat(AT_FDCWD, "/tmp/kstars_tests/Light/Luminance/Light_007.fits", O_RDONLY) = 48
openat(AT_FDCWD, "/tmp/kstars_tests/Light/Luminance/Light_007.fits", O_RDONLY) = 48

See how "Light_006.fits" is getting deleted at the instant "Light_007.fits" is stored? I'm still looking for the part of the code that is doing the "unlink" on the last capture file.
This is annoying to me because (1) it proves that I did not run the test vectors I committed at the exact location their contents suggest and (2) I'd like to use /tmp as a generic location for automated tests. Any ideas on the reason? Thanks!

-Eric