IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
This reverts commit 115a1901cd:
the change has not been tested properly unfortunately,
and it broke today's regular-rc builds fortunately;
there are no vulkan-{intel,radeon} packages in sisyphus
(only amdgpu), and these are present in lakostis@' repo:
http://www.unsafe.ru/lakostis/RPMS/ALTLinux/glvnd/repo/x86_64/RPMS.hasher/
Just drop the whole thing until it gets sorted out.
The whole RADEON_PACKAGES affair was introduced to deal
with fglrx/radeon incompatibility; it got basically
deprecated following fglrx removal from sisyphus,
and lakostis@ should have done "add" logic instead
of reusing the "set" one inappropriately.
Fixes: 85c52d71c6
See-also: https://lists.altlinux.org/pipermail/devel/2019-July/208126.html
Signed UEFI loader not required for aarch64.
NB: i586 images don't need UEFI SB either
and 32-bit shim is used for x86_64 images
along with proper 64-bit one.
This reverts commit 1b457a5d85.
It wasn't prudent to switch everyone to master FTP server;
Yandex mirror still has an order of magnitude more bandwidth.
The problem is that `chkconfig dm on' will enable
display manager service on *all* runlevels feasible
without paying any attention to its customary subset
of those; the solution seems just to avoid that.
Note that there's at least one more similar case
with networking services vs runlevels 2 and 3;
it's to be handled either in a similar manner,
or somewhat more generically.
Reported-by: Konstantin Savun and Speccyfighter
Suggested-by: Anton Midyukov <antohami@altlinux.org>
See-also: https://bugzilla.altlinux.org/36967
See-also: https://forum.altlinux.org/index.php?topic=36177.msg340553#msg340553
See-also: https://www.opennet.ru/openforum/vsluhforumID3/117762.html#81
It's actually sisyphus jeos image that needed this:
open-vm-tools got installed and then removed just
before the final reboot (unlike p8 jeos); not clear
what exactly caused this as subsequent installation
only pulled in libdnet and libicu56, neither of which
were excplicitly removed during cleanup.
Let's make the image a bit leaner until someone asks
for VMware support there.
We've decided that there's not much sense to continue building
server-openstack flavour of starterkits since it does take quite
an effort to even test properly, and there was rather no feedback
on those (maybe silent users will at least ping back this time).
Those interested are invited to follow server-v development
and contact Alexey Shabalin <shaba@altlinux.org> for cooperation.
PS: the actual pkglists are here, it's only the bunch of groups
tailored for standalone server-openstack with its group selection
menu layout that are being removed regarding the shared bits.
Add target use/bootloader/uboot. Das U-Boot -- the Universal Boot
Loader, used on single board PCs (also in qemu).
Use bootloader lists to provide better portability and to avoid
overcomplicating bootloader switch.
Added restriction for alterator-$$(BASE_BOOTLOADER)) only for
ISO images. Since the new uboot bootloader is not supported in
Alterator and is not used when building ISO images.
It could result in a broken config given broken chroot state;
it's of no practical difference as both cases just won't work
but we can at least provide meaningful diagnostics (and a bit
of perfectionism, heh).
Fixes: 2527fe94da2e810cf5d52cf548f84456fd5a6b4d
This feature ensures the availability and specification of the choice of
office package by analogy with the choice of browser. Additional options
may be set by explicit or indirect request.
gdm3 and lightdm have their own startup services,
and can no longer be started via prefdm.
Other display managers still run through prefdm.
See-also: https://bugzilla.altlinux.org/36769
This might evolve into some form of per-arch logic,
or otherwise x86 might get moved into a feature of
its own instead of being the presumed one.
The first step is done.
This reverts commit 226a2395e5.
Users are getting upset about NetworkManager being unable to
configure "System EthX" connections that are meant to set up
with "acc" (which isn't obvious at all).
See-also: https://forum.altlinux.org/index.php?topic=42519.0
elogind is the systemd project's "logind" extracted out to be
a standalone service. It integrates with PAM to know the set
of users that are logged in to a system and whether they are
logged in graphically, on the console, or remotely.
...through ensuring that the intentinal lack of dependencies
is actually *complete* and not still pulling the default elinks
in if nothing else was specified (this was still too much for
virtualization base images aiming to make use of mixin/p8).
chrony synchronizes time when connecting to a network using
NetworkManager. This feature is great for computers without
hardware clock, for example Raspberry Pi.
This one can be used to override the default content
of .disk/info file (used by propagator but can find
some other uses within installer as well); the reason
being that ISO9660's Volume ID is up to 32 characters
and a file lacks that particular limitation.
desktop+lxqt+extra contains some qtdesktop bits that do not
relate directly to lxqt. Therefore, it makes it difficult
to use use/x11/lxqt for the more basic builds based on lxqt.
- added netsurf - lightweight html4 browser
- added otter-browser - opera12-like browser
- drop qupzilla, konqueror: removed from Sisyphus
- palemoon not work on non-x86 systems
The original commit broke system tar2fs use by accidentally
moving TOPDIR definition into a separate shell execution;
thanks iv@ for spotting and fixing it promptly.
Co-authored-by: Ivan A. Melnikov" <iv@altlinux.org>
Brief version: otherwise DNS resolving might fail to perform
until dnsmasq service restart or system reboot (ouch!).
The problem:
1) NetworkManager requires dnsmasq;
2) dnsmasq can win a race against dhcpcd on ethernet.
The result:
1) /etc/resolv.conf looks fine;
2) dnsmasq is running;
3) resolving beyond /etc/hosts fails.
Suggested-by: Mikhail Efremov <sem@altlinux.org>
The overcomplicated scheme tacked the difference between
p7/t7 and sisyphus-armh of those days that's not there
anymore; all of those repos carry kde4-plasma-nm and not
plasma-applet-networkmanager.
It's been replaced by e18 and eventually the rolling
enlightenment package long ago, in fact missing by p8;
moksha might be a worthy replacement but now just drop
the good ol' stable bits in the profile either.
It's missing in sisyphus (and the last maintainer
was desperate with efforts required to keep it going),
just drop the cruft; if it ever returns, we'll revert this.
These have been broken by p8 already, and missing in sisyphus;
kde4 is broken in sisyphus but that's still not irreversible
(unlike our old and unmaintained TDE build, sigh).
Yandex mirror can lag behind at times:
http://mirmon.altlinux.org/sisyphus.html#ru
...and we've got 1GbE there now, not that bad
compared to 100Mbps back then...
NB: yandex traffic might be billed as "local"
in many regions, this should be revisited.
tar2fs comes from m-p, not from mkimage. Also, we should
use $TOPDIR from shell, not $(TOPDIR) from make, when
calling it.
Note: this is a security fix for environments relying
on packaged mkimage-profiles with sudo enabled for the
builder user.
Fixes: f293239d5b
None of Elbrus systems I had a chance to use as root
had any problems regarding TRIM feature of SSDs these
used to carry/support; let's start with e2k and maybe
carry over to x86 later.
This has manifested on e2k for me -- still using older
make-initrd lacking R: file -- so let's get this kludge
back for a while, or until any reasonable make-initrd
around has *everything* it needs to create images
listed in its runtime requirements.
Fixes: 2b3455c29a
filesystem package has been severely broken in sisyphus
since late August unfortunately, and I was too busy to
even notice that for too long; let's plug this kludge
until it's fixed now that I've finally seen that crap :-/
See-also: https://bugzilla.altlinux.org/35350
sin@ was kind enough to just stick mount.cifs into initrd
regardless of its presence in the chroot in question;
let's look first and only add what's found.
This started as a stopgap fix after make-initrd 2.2.0
which happened to collide with cifs-related m-p commits
in a somewhat unfortunate manner...
ldv@ has dropped sysklogd from sisyphus due to
its long-standing FTBFS and "other problems":
http://git.altlinux.org/tasks/archive/done/_207/211999/logs/events.6.1.log
The current recommended replacement is rsyslog
which is fine by itself but really overcomplicated
for the barebones logging system we usually need...
There's an ISO9660 COPY tag for license info file;
make use of it, factoring use/docs/license out
while at that.
One of the goals was to make it hold the reference
to reference to GPL in regular builds and starterkits
;-)
Every .iso was assumed to be bootable since the very beginning[*],
and isoboot images were deemed to be x86 isolinux ones; this didn't
change with basic ppc/armh support as I never ran into hardware
that would _boot_ those ISOs, not only run the code, and it was
only e2k isodata project that finally forced this refactoring.
It's still not perfect: pack and syslinux features still end up
somewhat interwoven, and too much places care for architecture
the image is being built for (instead of archdep features tossing
their appropriate bits and pieces in).
Should help:
- any-arch regarding isodata images;
- {x86,aarch64}/efi by decoupling isoboot and isolinux;
- ppc{,64} as introducing yaboot support will be easier now;
- mipsel{,64} too, hopefully.
* I knew of school addon images baked with mkimage-profiles-desktop
but postponed and then neglected the whole problem for years...
Linux Driver Management (aka LDM) allows easy configuration of different
hardware. Currently this hook does GPU configuration on systems with
multiple GPUs (aka Optimus/PRIME).
The goal was simple: every use/e2k/{x11,install2}/$MACHINE
must be self-sufficient regarding platform support, that is
it must depend on the corresponding CPU-specific target.
A bit less make-tech but still better overall.
NB: llvm cleanup is not needed anymore after upgrade to
Mesa-17/llvm-3.7, should be dropped like this year.
Forgot to do that before merging, sorry.
(xorg-drv-ati cares for dependencies
but we still prefer modesetting driver
in the installers to make them versatile)
The places of the terms *do* affect the sum in apt's case;
start with lowlevel items like SYSTEM_PACKAGES and end with
high-level ones like THE_LISTS to reduce the chance of getting
hit by premature virtual dependency expansion/fixation.
Adapt live and rescue features accordingly.
See-also: https://bugzilla.altlinux.org/show_bug.cgi?id=30806#c5
Suggested-by: Leonid Krivoshein <klark@altlinux.org>
The infamous systemd infiltration through secondary
"targets" using virtual dependencies started with
syslogd-daemon, continued with network-config-subsystem
and later with ntp-server; all of its implementations
pulling in the main package which is explicitly unwanted
in sysv-based images.
Let's try employing pkgpriorities.
The common problem was network-config-subsystem getting
resolved into something completely wrong (like net-scripts
or systemd-networkd) *before* it got specified precisely;
let's just avoid the common cause, that is, a metapackage.
See-also: https://bugzilla.altlinux.org/show_bug.cgi?id=30806
It's not much use for it to stay without the actual
pointer to the place where NM GUIs are referenced,
I've almost started out implementing the "missing"
bit myself right now :-/
It's legit here as I know no Elbrus users among
non-Russian-speakers, at least so far; should move
to generic l10n feature (which will definitely see
its glory by then).
This is to avoid extra actions when one actually needs
KOI8-R (which is still hardwired in some lcc messages).
Should all be rolled into l10n feature some day...
The early scheme consisted of boot.conf template
with a separate hook filling it in; this was nice
for serial console setup script which could just
amend the kernel command line as needed _but_
this got changed towards the more generic scheme
(breaking the hook in question, unfortunately).
This one supports multiple kernels (and tweaks boot
labels aiming to simplify target kernel selection
for the CPU at hand) instead of an earlier attempt
to fill in a template.
No need for the template so just drop it.
This one might probably change into core functionality but right now
it should help when you get not the branding, init system, or whatever
alternative you do _not_ need.
See-also: https://bugzilla.altlinux.org/show_bug.cgi?id=30806
We definitely don't want to see rekonq anywhere,
it's grossly incompetent as a web browser;
and Qupzilla is now being developed under KDE
umbrella and the new name Falkon.
This looks like missing in server starterkit for me;
builder one can be installed by hand if one knows how
to do that but let's better be explicit about that.
Putting any configuration into /etc/net/ifaces/lo/resolv.conf
makes etcnet *overwrite* /etc/resolv.conf, while putting that
into /etc/resolv.conf itself makes e.g. vzctl --nameserver
*append* to what's been specified.
Reported-by: Gleb Fotengauer-Malinovskiy <glebfm@altlinux.org>
...explicitly; this has been triggered by the change to
livecd-install as of 0.9.10-alt2 (doesn't R: alterator-grub
anymore to allow non-grub platforms as well).
This should be a no-op _together_ with the mentioned change
by now as grub is selected explicitly; will take more attention
for e.g. e2k as the problem just shifted here from livecd-install.
Reported-by: Leonid Krivoshein <klark@altlinux.org>
Those images who are fine with ALSA might still benefit
from addons like udev-alsa to have mixer levels restored
on boot; and PulseAudio-bearing images can be based upon
the same intermediate targets now without receiving the
addons they don't need.
This has been long overdue: pretending that "http server"
is an Apache2 with PHP5 was wrong for ages, and providing
apache2/nginx, php5/php7 choice along with more databases
was waiting for its user proposal.
Follow these changes in server feature as well.
Suggested-by: Altexander
See-also: https://forum.altlinux.org/index.php?topic=40290
This one enables rw session support for Rescue images;
the nitpick is that syslinux' gfxboot *will* set this up
when booting the same ISO from USB Flash media instead of
optical one (CD/DVD-ROM/RW), and we don't put gfxboot into
a standalone rescue image.
See-also: http://altlinux.org/propagator
See-also: http://altlinux.org/remount_rw
Radeon HDMI as the default soundcard isn't really optimal,
and trying to push 44100 kHz samples towards 48 kHz interface
isn't going to sound great; provide sane ALSA dmix defaults
for Elbrus-401 PC workstation.
BASE_PACKAGES_REGEXP and THE_PACKAGES_REGEXP,
to be exact; the lack of handling these appears
to have been the culprit of firefox missing in
vm images which use/browser/firefox.
This reverts commit 41a3f09132:
at least build-vm doesn't do *_REGEXP (which is worth fixing
but all relevant branches got firefox-classic_theme_restorer
so this tweak is now irrelevant and a bit wasteful).
It's strange but ve/builder and distro/live-builder
along with distro/regular-builder are all mostly
"independent", that is duplicating functionality
without any reasonable gain; spotted finally.
The early builds used to rely upon a non-committed
rootfs/files/etc/X11/xorg.conf within this feature
which was a bit annoying and would have screwed an
Elbrus system based on any other GPU.
So let's provide some flexibility by packaging it.
These are creeping throughout m-p and aren't really needed
by default as these should be enabled upon configuration;
let's provide a single switch off point.
No need to deduce kernel version again,
just save it in a temporary file.
The main reason to change what worked is
that e2k kernel-image package has Linux bits
named as image-$kver and not vmlinuz-$kver;
the guessing logic taking all of this into
account resulted in non-aesthetic patch.
NB: there's a duplicating script within
kernel feature; it wasn't easy to avoid
this and it might differ when handling
multiple kernels, I didn't think much
about this now as vm images tend to ship
with the sole one.
This is unfortunate but Sisyphus' Xfce can't really strive with its
default /etc/xdg/xfce4/xfconf/xfce-perchannel-xml/xsettings.xml,
namely IconThemeName; while p8's one still does pretty good.
Thanks antohami@ for his analysis and workaround suggestion.
See-also: https://bugzilla.altlinux.org/32924
See-also: https://lists.altlinux.org/pipermail/devel/2017-February/202348.html
Suggested-by: Anton Midyukov <midyukov-anton@ya.ru>
Drop duplicated code pulled into armh feature in haste
and another half-copy that made it into build-ve feature;
make ve.mk more explicit about what it actually does
regarding default nameservers (and switch it to level3
while at that, just for diversity they preach).
Refer to net-dns feature where appropriate
(it actually started out as an extension of
net feature but the reasons to separate it
quickly became apparent to me).
Was broken in too many places including several libraries
x11vnc currently links against getting removed, several
utilities missing, and no way to guess what to do...
One could stumble upon http://altlinux.org/vncinst
or find the implementation in installer.git but it's
not exactly user-friendly.
An initial (and quite naive) implementation was done
for conf.d/ files but features do need it as well --
which became obvious during a casual inspection of
distro/alt-workstation's target graph (which occured
broken).
The code is a messy hackery unfortunately, regexes used
are fragile regarding e.g. [A-Z_] in target names.
Read with care.
NB: lib/ might need this too but it's rather stable
and contains some special cases that would obstruct
the regexes even more...
In this case it's rather worth it to examine build.log
than read documentation again (as vm.txt should have been
read or at least skimmed through to get sudo setup ready,
and the problem might be either an environment one or a bug).
rpm-4.13 stopped bringing alt-gpgkeys in, and specifying it
explicitly for STAGE1_PACKAGES didn't help somehow; don't presume
/usr/lib/alt-gpgkeys/ existence anyways.
elinks doesn't require libX11 (unlike links2),
and webclient can become e.g. rekonq pulling
in kde4libs -- that's exactly what I ran into
while preparing 201612xx server starterkits.
For some reason DOT_BASE only got BASE_PACKAGES_REGEXP so e.g.
browser feature use well could result in the lack of a browser
if no metapackage pulls one in!
X-Brown-Paper-Bag :-(
It's at least removing the very obvious user->root
attack through (maliciously) modifying bin/tar2fs
and waiting for it to be run; if mkimage-profiles
is installed system-wide as a package, the script
from /usr/share/mkimage-profiles will be tried so
those willing to allow vm/* build to themselves
can provide for a passwordless sudo (as described
in doc/vm.txt) to run a root-only writable script,
not user-writable.
Still not perfect but a step away from the abyss.
It turns out that -regular pulls in too much
(including xfce-polkit that doesn't work with
sysvinit); let's provide means to have it both
ways, need xfce4 4.12-alt5 or later metapackage
for this change to be effective.
The issue at hand it _nmconnect group that's supported
in ALT NM for polkit-less cases; it has to be added
*before* 50-users hook from deflogin feature fires off.
The existing systemd-related hook has been renamed
to reflect its job better.
This one is related to forensics mode handling in mkimage:
there's a regex that's expecting the old string ending with
hash value, and it stops working when this is changed here
(at least as of mkimage 0.2.18).
So partially undo the change for this particular file to avoid
refind's example menu for regular-rescue.iso (or any other image
involving forensics mode), and hope to revert this commit some day.
This keyword was misused due to the lack of understanding
of its meaning; thanks zerg@ for the hint, let's hide the
kernel boot parameters not expected to be modified by user
under the hood (and vice versa in some cases).
Note that this is just a keyword shift in all cases
but mediacheck where "showopts" is now added.
NB: this keyword is only relevant for gfxboot.
Now that we don't ignore X-Alterator-PackageList
in pkggroup files these few places referencing
group names as list names have broke; bring 'em
in line through this new and shiny function.
This has been spotted by rider@ and reproduced by me as well:
some touchpads would work in livecd/installed system but not
within the installer itself.
Commit 514652f has broke GLOBAL_CLEANUP_PACKAGES by accidentally
excluding it from export (in favour of GLOBAL_CLEANUP_BASE_PACKAGES
that's been added then); fix that.
This script was completely careless regarding the chance
to meet an empty variable resulting in plain "rpm -qa"
and subsequent attempt to, well, remove *all* packages.
Thanks zerg@ for being persistent this time, even if
he could probably find the culprit and send in this patch.
:)
The problem at hand was that use/x11/xorg has been final,
and zerg@ just couldn't switch from nouveau to nvidia
when kdesktop needs that one.
Initial approach included a "big" FREE/PROP switch that
chose the particular KMODULES/PACKAGES to get added to
THE_* but that fails to achieve e.g. nvidia+radeon combo;
looks like these need individual switches.
The use case at hand was: "we'd better backup this system
to a flashdrive before installing" (given quad-core CPU
and half-terabyte HDD); pxz is pretty tiny, no worries.
There were two problems:
- the latest pgsql related groups made installation
impossible (yes, that last minute change);
- hardware testing shows that use/stage2/kms is now
requisite as xorg-drv-fbdev might just refuse to work
with what looks like a perfectly good framebuffer...
Do away with them *quick*.
"Failsafe install" disabling APIC/LAPIC looks somewhat obsolete
by now; the only reasonable part seems to be the attempt to force
VESA videodriver for the installer (should be done within installer
itself though).
"Forensic mode" submenu has fallen apart after the original commit
as the tricky logic in mkimage::tools/mki-copy-efiboot failed to
pick up the new variant; this should all be redone (solo@ has
started doing something but it needs a time-consuming review).
Fixes: 79d0208841
use/docs/license will copy the texts contained in branding
package ("notes" one) over to the image's rootdir so these
can be read with ease; otherwise one has to look up the
right package at best (or unpack squashfs, no user can be
really expected to do that just to *read* a *license*).
This was originally profiles/scripts.d/01-copy-license
script from m-p-d; got cut down heavily.
The problem at hand was that an installer component
of a "DVD class" image does use/cleanup/installer
while installable LiveCD component gets broken by that
(livecd-install -> installer-scripts-remount-stage2
which gets removed as installer-*).
Split those.
Package profiles -- the ones allowing for a multi-purpose
installer -- have been basically overlooked during previous
mkimage-profiles development, unfortunately.
This is the very basic part: put them into pkg-groups.tar.
THE_* variables serve user needs while shim belongs
to either SYSTEM or COMMON level packages, not needed
explicitly for stage1 though (mkimage will put it there
when needed) so it's just COMMON.
It's not reasonable for use/firmware/laptop to depend on
use/firmware/wireless as some laptops come without WiFi
cards and wireless userspace to use those is specified
elsewhere anyways.
This partially reverts commit 30d3838: trying to use/rescue
with e.g. distro/simply results in conflict between SysVinit
and systemd-sysvinit; INIT_TYPE had to relation to RESCUE_LISTS
in the first place. Ugh.
This has long been a TODO item but an elegant solution
just didn't come until the night before starterkits...
some services (mostly those operating on real hardware)
do not fit virtual environments at all, won't even start.
