Unexpected behavior when installer already in DMG
epackorigan opened this issue · 5 comments
If you use https://github.com/munki/macadmin-scripts/blob/master/installinstallmacos.py to build DMGs for the installer, and mount those as the source for macinbox as /Volumes/Install macOS Mojave, the one created by macinbox ends up as /Volumes/Install macOS Mojave 1, and the first is the one that ends up being unmounted at the end of the process, leaving the other around...
Can the following features be implemented:
- Check to see if the installer is already in a DMG, and skip creating a new DMG to store the installer.
- Check where the new installer is mounted, and unmount that one.
Hi, @epackorigan! Thanks for letting me know about this issue.
I am already working on some unrelated changes that will address your second request; the wrapper image for the installer will be correctly unmounted using its device node instead of its supposed mountpoint.
I do like your first request; it would be much more efficient if the script detected that the installer was already in a DMG, and skipped creating the wrapper. I'll use this issue to track adding that behavior.
With the changes in 73aa46b the correct installer disk image should now be getting unmounted.
With the changes in 9623626, creation of the wrapper image is now skipped if the installer app is on a non-root volume.
I'll leave this issue open until I'm ready to release v3.2.0.
With the changes in 1c5115e I have implemented a --installer-dmg option which allows you to specify an unmounted installer disk image (e.g. one created by installinstallmacos.py). The macinbox script will automatically mount and unmount this disk image as needed.
v3.2.0 has been released to rubygems.org. Closing.