Obviously the installer cannot do very much if mkfs.ext4 never completes.
And I intended to write:
So the bug must be in e2fsprogs (or even in the kernel)
However, thinking one more time I don't think either of those statements is necessarily true.
The output of mkfs is somehow redirected to the installer GUI (ncurses in my case).
So I guess this redirection could cause a deadlock, such that mkfs waits forever
to write some output to the installer.
I wrote:
Obviously the installer cannot do very much if mkfs.ext4 never completes.
And I intended to write:
So the bug must be in e2fsprogs (or even in the kernel)
However, thinking one more time I don't think either of those statements is necessarily true.
The output of mkfs is somehow redirected to the installer GUI (ncurses in my case).
So I guess this redirection could cause a deadlock, such that mkfs waits forever
to write some output to the installer.