Package: libguestfs0t64
Version: 1:1.54.1-2+b1
Severity: wishlist

Hi,

This package has a large array of dependencies:

Depends: libacl1 (>= 2.2.23), libaugeas0 (>= 1.8.0), libc6 (>= 2.38),
         libcap2 (>= 1:2.10), libfuse2t64 (>= 2.8), libhivex0 (>= 1.3.14),
         libjansson4 (>= 2.14), libpcre2-8-0 (>= 10.22), librpm10 (>=
         4.14.0+dfsg1), libselinux1 (>= 3.1~), libsystemd0, libtirpc3t64 (>=
         1.0.2), libtsk19t64 (>= 4.12.1+dfsg), libvirt0 (>= 1.2.7~rc2),
         libxml2 (>= 2.8.0), libyara10 (>= 4.0.0~), supermin (>= 5.2.2-4~),
         qemu-system-x86, qemu-utils, db-util, icoutils, netpbm, osinfo-db,
         acl, attr, binutils, bsdextrautils, btrfs-progs, bzip2, cpio,
         cryptsetup-bin, dash, dhcpcd-base | isc-dhcp-client, diffutils,
         dosfstools, e2fsprogs, exfatprogs, extlinux, f2fs-tools, fdisk |
         util-linux (<< 2.29.2), file, gawk, gdisk, grub2-common, iproute2 |
         iproute, kmod | module-init-tools, ldmtool, less, lsscsi, lvm2,
         lzop, mdadm, mount, mtools, ntfs-3g, openssh-client, parted,
         pciutils, procps, psmisc, scrub, sleuthkit, squashfs-tools,
         syslinux, systemd-sysv | sysvinit-core, udev, util-linux-extra,
         uuid-runtime, xz-utils, zerofree, zstd
Recommends: libguestfs-hfsplus, libguestfs-reiserfs, libguestfs-xfs

Some of these are obvious and mundane, but it's less obvious why a lot of
these things would have to be strict dependencies.

For example, if a user needs virt-customize to set up simpler VMs that only
use something default like MBR partition tables, Linux, ext4, bridged
networking, and never inspects these visually, it seems like a lot of this
stuff to support completely different sorts of VMs, or allows rendering
image icons etc, is entirely redundant.

And it's not just redundant, but enlarges the security footprint of
the host machines for something that might not actually be used.

Could some of these dependencies be made optional on run-time, and moved
to Recommends at least?

TIA.

-- 
Josip Rodin

Reply via email to