AFAICS, this dep chain (syslinux->mtools->glibc-gconv-extra) is
the only reason glibc-gconv-extra is in the installer env. It's
quite large (8M). syslinux's dep on mtools seems to be due to
(one of) its installer(s) using mtools, but I don't think we
ever run that from the installer env; we only pull syslinux into
it to set up bootloader stuff for the installer image itself,
which happens in x86.tmpl and doesn't involve actually running
a syslinux installer, just copying files around.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
As best I can tell, guile22 is only in the installer env because
we include gdb and gdb depends on the libraries from it. gdb's
dep is for guile scripting support, AIUI. I don't think any kind
of libreport or manual gdb debugging we'd do in the installer env
would need that support, and even if it does, the ccache does
seem to be a cache - docs say if the files are not present,
they'll be generated on the fly from the .scm format sources in
/usr/share/guile/2.2. So I think it should be safe to ditch this
cache, which is large (it takes up ~38M uncompressed).
Signed-off-by: Adam Williamson <awilliam@redhat.com>
This addresses most "no files matched!" and "no files to remove!"
errors in current F36/Rawhide. They mostly relate to packages
that are no longer pulled in at all, or whose layout has changed.
The entries for audit haven't been fixed for the /usr merge years
ago. In linux-firmware, all files have been xz-compressed at some
point, so the entries were not matching any more. The usbdux/
subdir contains only sources for the actual firmwares that are a
level above, and we don't include those in the package any more;
the actual firmwares are useless in the installer env, so this
removes them. util-linux was split into util-linux and
util-linux-core, so we have to add an entry for util-linux-core
and move the relevant excludes to that entry.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
Prior to bc46724, we dropped several variants of DejaVu to save
space, including DejaVuSans-Oblique.ttf, to which I think this
is the equivalent. I believe the idea is that it's not worth
half a megabyte just to get tuned rendering of italics in the
installer environment (where they're rarely used anyway); AIUI
in the absence of a specific oblique/italic font face, freetype
will produce one by slanting the regular face.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
The dropped google-noto packages contain fonts of scripts for
languages that the installer is not translated into. Most are
obvious, but for the record, "lao" is for the script and
language also called Lao; "thaana" is for the script Thaana,
used for the language Maldivian.
sil-abyssinica-fonts was an older set of Ethiopic fonts used for
e.g. Amharic; we now prefer google-noto-sans-ethiopic-vf-fonts,
the installer environment does not need both.
Similarly, sil-scheherazade-fonts covers a similar range to
google-noto-sans-arabic-vf-fonts and paktype-naskh-basic-fonts,
which are now preferred.
xorg-x11-fonts-misc contains bitmap fonts for some non-Latin
scripts. I'm fairly sure nothing in the installer environment
should need bitmap fonts any more.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
scp in openssh 8.7 will change to use sftp protocol. See
https://github.com/openssh/openssh-portable/pull/194 for details.
This enables the sshd internal-sftp implementation so that newer scp
versions will continue to work as expected. Note that the sshd service
is only running during the installation if inst.sshd is passed on the
kernel cmdline.
Resolves: rhbz#2040770
rngd should no long be needed, recent kernels have enough entropy out of
the box. This leaves it in the boot.iso but only enables it if inst.rngd
is passed on the kernel cmdline.
Resolves: rhbz#2028720
Since we're leaving pipewire-libs in, it'll still pull in liblilv
from 0.3.41 onwards. Currently liblilv is in the same package as
some binaries which we don't need, and one actually requires the
removed libsndfile, so we need to trim it.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
We aim to remove all sound support from the installer root, but
this was never updated for Pipewire. This just started causing
compose failures because of a dep chain from pipewire to lilv.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
Marvell Prestera firmware has been split into its own subpackage,
so instead of stripping the files from linux-firmware, exclude
the package from the globed install command.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
As Peter Robinson explains here:
https://bugzilla.redhat.com/show_bug.cgi?id=2011615#c3
these are not useful, as the devices they're for do not support
netinst-style deployment.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
These firmwares are for Qualcomm smartphone chipsets (SM845 and
SM8250). Don't think they're any use in network install images.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
grub2-2.06-3 changes where the unicode.pf2 font file is stored.
This changes the efi.tmpl to install it from the new location, which
means that it depends on grub2-2.06-3, but there is no way to express
this in the runtime-install.tmpl so if you see a failure like:
2021-07-08 16:10:05,586: OSError: nothing matching /var/tmp/lorax/lorax.t80f74er/installroot/boot/grub2/fonts/unicode.pf2 in /
it means the new version of grub2 wasn't in the repos you used when
running lorax.
Fixes#1165
These were apparently lost in the transition to livemedia-creator.
livecd-creator added them in it's code, and lmc based its live config
files on the the boot.iso configs which do not include it (on purpose).
Fedora has split xorg-x11-font-utils, with bdftopcf, mkfontscale and
fonttosfnt being split out into separate packages. Remove all of those too.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
In f34 and beyond, the old xorg-x11-server-utils package was split up
into seperate packages for each util. This was to allow them to rev at
their own pace instead of requiring all of them to rebuild at once.
See https://bugzilla.redhat.com/show_bug.cgi?id=1932754
and
https://fedoraproject.org/wiki/Changes/XorgUtilityDeaggregation
We need to adjust lorax (in f34+) to not try and remove the
xorg-x11-server-utils package (as it no longer exists) and also to
install the 2 utils that we need from it for installs.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
It never used vesa directly, it passes nomodeset, so make it less
confusing. Also simplify the menu text so that it fits on the screen
for distributions with long names.
anaconda in F34 and Rawhide recently stopped accepting params
without the inst. prefix, so 'rescue' does nothing except print
a warning now. We need to use `inst.rescue`. This has worked for
quite a long time so will be OK at least on all Fedoras and RHEL
8, not sure about RHEL 7.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
The difference between the anaconda-install-{env,img}-deps packages is how
they treat dependencies. The -env package leaves some dependencies as weak to
allow less featureful builds. The -img package hard-requires everything
Anaconda could potentially use and ensures everything works.
For boot.iso, the latter is preferable. Its usage moves some things from the
templates to that package.
A post-installation script in fedora-live-base.ks actually modifies
Lorax (not idempotently) with the change in this commit while it is
running. This modification belongs directly in Lorax instead.
Note a subtle distinction in behavior that has been preserved here.
Lorax will copy the livecd-iso-to-disk script from the installroot
(if present). Running livecd-creator will copy it from the existing
root filesystem instead.
Since Fedora 30, license files are missing from the ISO filesystem
of live or installer images (including official builds). The source
path to these files changed when they were moved into a subpackage
named fedora-release-common (or generic-release-common).
Also, copy the license files from the installroot, rather than the
existing root filesystem.
There's no reason for it to run, it can't notify anyone. But disabling
the service, or masking it, doesn't work so remove the service files
from the rootfs.
Resolves: rhbz#1888730
mk-s390-cdboot has stopped working because the kernel outgrew the
hard-coded offset it used when creating cdboot.img. IBM now has a script
in s390utils that can do the same thing so use the upstream script
instead.
This drops mk-s390-cdboot script, switches the s390 templates to use
mk-s390image from s390utils.
It adds @ROOT@ to cdboot.prm, and sets inst.stage2 so that the installer
image will be found when booting the iso.
Resolves: rhbz#1891778
Some of the options have been removed, others are now the default.
MOTD still needs to be printed, the boot environment doesn't include the
pam motd module.
Resolves: rhbz#1872892
Previously this symlinked them to /dev/null, which didn't really
accomplish anything since they get recreated. So just remove them so
python can decide whether or not to recreate them.
When we stopped caring about ppc and ppc64, we changed several
instances of three-item tuples:
("ppc", "ppc64", "ppc64le")
into...this:
("ppc64le")
which is not a single item tuple, but just the string "ppc64le"
in some extraneous braces. It so happens that the right thing
still happened in all relevant cases , we think, but it's wrong.
There's no need to be using an iterator at all for a single
item, so just change them all to == "ppc64le" or != "ppc64le" as
appropriate.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
AFAICS, the devices that need these firmwares - various boards
built by NXP, https://www.nxp.com - are all aarch64. So we don't
need to carry these firmware files in the installer env for other
arches.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
Mellanox Spectrum devices are switches intended for data centers.
It is I guess feasible that someone might want to install Fedora
on one, but from the product pages and data sheets, I believe
they all have management interfaces that do not require this
firmware to work, and that's what you'd use if you needed a
network connection during OS deployment. The firmware is only
needed for the actual switched interfaces, and we don't need to
make those work during installation.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
I based this on the output of a recent installer image build:
https://kojipkgs.fedoraproject.org/compose/branched/Fedora-33-20200904.n.0/logs/x86_64/buildinstall-Everything-logs/pylorax.log
I looked at every runtime-cleanup related error there and tried
to make appropriate changes. In many cases this means just
removing a line that isn't needed any more because the package
in question just went away or is no longer pulled into the
installer environment. In other cases packages changed name or
files moved around, and I tried to make appropriate updates. In
a few cases files moved to another package but I wasn't sure
enough it would still be safe to remove them so I just left them
in place. Most of the changes here I'm pretty sure should be
safe, though there *could* be unforeseen fallout from e.g. fixing
the removals from procps to be removals from procps-ng - it's
been years since that package was renamed, so something *could*
have started using those binaries in the meantime. I did at least
check that anaconda itself does not.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
These are for devices that just aren't going to be needed during
install, like video encode/decode accelerators, TV capture cards,
webcams, and some sound firmwares that should probably be in
alsa-firmware but aren't. This is a fairly conservative cut, I
will split some possibly more controversial cuts into separate
commits for ease of detachment. The linux-firmware WHENCE file is
an invaluable resource in figuring this out.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
