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.
There's considerable confusion accumulated over the
expansion of regular builds with non-GUI targets:
- bare meant quite a lot including wireless;
- base meant GUI base actually,
so rescue, server, and especially jeos images had to
either cope with the bloat or start from scratch.
Let's fix "base" first...
Looks like QEMU 2.2.0 gets surprised with some kernel module
that udev tries to load during "Populating /dev:" boot stage;
dropping *both* +vmguest and +wireless dependencies does help
while switching to un-def kernel flavour doesn't (as is or with
any single one of these two deps removed).
Thanks user who reported the problem at opennet.ru
(alias "myhand").
altlinux-p7-kde4-20141212 images had a problem:
- the generic starterkit profile is tuned
for modest disk space consumption;
- KDE4 is quite extensive;
- kde4-regular subpackage tosses in most of KDE4 SC
...thus the livecd wouldn't autopartition enough
space for root filesystem when being installed.
volumes-profile-lite specifies 10Gb rootfs currently
which is just fine (-kdesktop one starts with 30Gb),
let's use it until vm-profile management is imported
into mkimage-profiles or things are redone completely.
There's no more livecd-gnome3-nosetup, it's been renamed
to setup-gnome3-done with no P:/O: to reduce repo spam.
Should have cherry-picked this into master back then :-/
I've been puzzled to find current live-webkiosk.iso trying to go
localboot with no timeout; this is basically a plug, VM images
might need more effort to go in.
There's GNOME 3.8 in p7/branch and it has no initial setup tool
thus no disabler is needed anyways; I've fixed this a while ago
but seem to have managed to lose the commit somehow :-/
So ALT Linux has got decent build tools but no one even knows
about those, right? And starting to use those is somewhat tedious
as it requires ALT installation with a specially crafted user account?
live-builder.iso has been useful enough to consider including its
relative into starterkits skipping regular builds for roughly the
same reasons as those for installers: we're better off helping users
to get onto the stable platform bandwagon than having them excited
with Sisyphus just to hit some wall down the road (even if we warn
of these walls well in advance most of the time), and while packages
tend to get into branches via sisyphus repo it's still better to stay
on the safe side with what we make easily available.
It's been ignored since the introduction of deflogin feature
and removal of build-vm specific ROOTPW variable handling;
might be adjusted to mandate this for vm/bare too, have to
think it over.
This has been asked for by lewellyn@freenode, why not.
NB: distro/.regular-sysv doesn't include use/net-eth/dhcp
(as it looks like asking for) since there's still hope
to get NM cooperating with sysvinit again.
Both locale and keyboard have been set up already,
no use to waste time on those (which results in 'us'
keyboard layout missing out totally, ironically).
Thanks aris@ for the tip.
This aims to work around URW fonts deficiencies combined
with the relatively high position of their standard ones
in default /etc/fonts/conf.avail/60-alt-post-user.conf;
see also #30293, #30294.
It's not EFI-bootable due to the extra size penalty
with current build/boot technology, but it's also
not ISO-hybrid as most images are made hybrid by
making them EFI-capable as well since the processes
are related.
Thanks dango at the forum for asking:
http://forum.altlinux.org/index.php/topic,33094.msg236808.html#msg236808
+net-eth covers both stage2 and base installation parts
while use/stage2/net-eth would result in an inobvious
catch when udev-rule-generator-net would hit install2
but not the installed system, and interface bindings
would be carried over from installer to the installation
but would *not* be updated in case of changed network
card(s) configuration.
It was spilled over an intermediate target and a mixin
for what looks like no good reason; let's factor these
in by means of +net-eth as *both* are really needed
for an installer anyways, and if/when we do installers
with those new and crippled ethernet interface names
this addition can be dropped to be used where required.
The bare intermediate target did use/stage2/net-eth but it
was found out by sem@ that an installed server would lack
udev-rule-generator-net package so its ethernet interface
naming would be inherited from install2 stage only; fix it
by adding use/net-eth and reduce the churn with +net-eth.
It looks *ugly* on-screen, at least within regular builds,
even if the screen is 166dpi.
Based on a quick experiment this morning I'd suggest using
fonts-otf-adobe-source-{code,sans}-pro instead -- and it's
available as use/fonts/otf/adobe now, incidentally.
The documentation is still built with it though as a2x/fop look
unhappy otherwise (as in replacing Cyrillic glyphs with "#"s).
Thanks mithraen@ for creating a universal frontend script
for udisksctl/pmount/hmount user mount tools; let's try it
within the minimalistic GUI image for the starters.
There's no sense to duplucate sysklogd requirement
in plethora of fallback places when a specific feature
responsible for comprehensive init system choice has been
implemented since; just use/init/sysv as needed.
systemd-214 has major problems with starting these services,
and KDC should definitely not even try to start up before
setup (which is not feasible given that these are LiveCDs).
domain-client pkglist inhabitants and net-usershares feature
are nasty enough to bring a lot of extra garbage in unfortunately
(alterator-auth, alterator-kdc, alterator-net-shares involved).
Commit 78f2158 left those images which required NM but not
its applet explicitly broken (as in "no applet at all");
this should probably be redone alike to browser feature
but let's provide a 20140612 band-aid at least.
I thought about this again and came to conclusion that bringing
alteratord and ahttpd up or down should rather be done at the same
time as backends and frontend are useless without each other
(at least for your average sysadmin).
Matt Lewandowsky suggested that alteratord be enabled
but ahttpd be disabled so as to avoid extra port being
listened to out-of-box but to be able to turn web GUI on
when needed.
Suggested by frbrgeorge@ and sounds quite reasonable
given that live-rescue.iso is very immature yet and
there are only a few -- but compelling -- reasons
to provide a graphical rescue image, gparted being
one of them for sure.
The logic is pretty much the same as with live.mk,
even somewhat extended as this has actually been
the driver of this change: some images like icewm
or lxqt-based ones might show off other browsers
explicitly (in addition to zerg@'s request).
Firefox was the very reasonable default for initial livecd
implementation but now that at least initial browser chooser
infrastructure is in place it's time to un-hardwire its use.
It's _the_ default but switchable now so that images providing
a comprehensive browser can avoid feature duplication.
This should better lurk here unless someone (including myself)
either forgets or doesn't realize the inobvious chain of the
assumptions made for the read-only warranty to actually work.
This is actually downplaying: the net effect is that
a few images continue to carry krb5-ticket-watcher
and have avahi service enabled by default while most
of the images have one "weird" item less on their menus
and a few hundred kilobytes less in total RSS.
These are not really needed by default in lightweight
distros sporting reduced application and services set,
and the heavier ones will take relatively less hit
by default while being more ready as ALT Domain clients
(which was the whole motivation behind adding the feature
and the corresponding pkglist).
A virtual machine lacking DHCP client seems much less useful,
and being able to shut one down via emulated ACPI button press
seems like no luxury either.
led-ws kernel flavour has gained kernel-modules-vmware
recently, let's add this to the appropriate targets.
It's used in regular-jeos already but THE_ part was missing.
This should provide the fix for #30024 (thanks ildar@):
vmware relies on mptspi.ko by default and that's been
split out into scsi and not guest subpackage for led-ws.
BTW use/install2/vmguest instead of a bunch.
plasma-applet-networkmanager has been superseded by a bunch
of kde4-plasma-nm* packages; only the main one has been included
in regular-kde4 flavour since the switch resulting in the lack of
VPN/mobile connectivity options.
My opinion still is that plasma-applet-networkmanager should be
returned as a metapackage for p7/branch timespan so that images
could be built no matter whether it's sisyphus or p7 at hand.
Oh well.
Package has been prepared by shaba@ and sem@,
and it looks like ALT Linux with un-def kernel
is one of the few (or just the one) distribution
running on Hyper-V Gen.2 rather flawlessly
thanks to efforts by boyarsh@ and vitty@.
Just a convenient knob for a few things done previously
to help spare the interesting bits from being overrun.
NB: live_rw isn't added although it might be useful
since that would yield too many boot targets overall;
this is likely to change some day, hopefully when
media type detection/handling is implemented.
Let's bump syslinux timeout to 20 seconds either
so that iKVM users have a chance to select anything
and not just see the default booting after a few
screen area size changes.
cvltonemap is no more available in sisyphus/p7;
xsane and usbutils were sorely missing (thanks dd@).
NB: fim is currently i586-only, need to fix or drop it.
server-zabbix.iso is ready for deployment,
and live-zabbix.iso is zabbix agent + firefox
(needs at least Server to be specified properly
within zabbix_agentd.conf when booted).
It used to be added in server-ovz but it really belongs to
the underlying server-mini already as more images built
upon that one should perform correct shutdown given ATX
compatible case/mobo or a VM that can do ACPI.
This reverts commit fd8f375573.
xdg-su is broken (some would say beyond repair) regarding DE
detection and handling of various GUI helper utilities,
especially in graceful fallback department; only a few images
can get imagewriter until this is fixed, let kde4 be the one.
The whole story with this installer has been due to a query
at #altlinux whether there's a distribution image similar
to altlinux-p7-rescue.iso which appears to be booting under
Hyper-V Gen. 2 without a hitch; changing just the kernel
towards the newer one made the user rather happy since
everything worked out-of-box for him, even unimportant bits.
Of course it's mostly due to boyarsh@'s preceding work :)
NB: there's no use to build i586 version as that hypervisor
lacks CSM (BIOS implementation) and would only boot UEFI
compatible operating systems with pretty strict requirements.
It has dawned on me that gdm2.20 is more widely useful
within sysvinit based builds: lightdm with gtk greeter
would fail to poweroff/reboot while this one would not.
aen@ asked to ship this one as well; no problem given mixins,
still being able to *switch* the init instead of regrafting
would be very beneficial.
NB:
- wdm can't do autologin;
- wdm can poweroff/halt;
- wdm+autologin won't work under systemd (via prefdm.service);
- nodm will work under systemd;
- nodm will ruin consolehelper -> livecd-install by root's PATH;
- gdm2.20 is lightweight, feature complete and sovereign enough.
The reason for an explicit cleanup is that VNC installation support
is now left in by default (see #29901); thus this commit is only
keeping the status quo for this image.
It's still a GUI installer but pretty much barebone one by now...
in particular, it needs no xorg-drv-$hardware being mostly targeted
at VMs and demo appliances where fbdev is rather enough.
Cleanup extra kernel drivers too.
The mixin concept and name has been borrowed from Ruby
language -- it's a kind of thing that can be added to
more or less whatever suitable; the problem it tries
to solve is that incrementally building up the image
configuration breaks when one would like to change
something that's been configured in early enough so that
grafting early will warrant a lot of duplication later on
but inheriting too much things that need to be changed
gets too much hackery in.
It started while trying to build an installer image
using configuration bits and pieces collected while
bringing an installable LiveCD together: there are
just too many livecdish things in a LiveCD to try
and rebase the actual desktop configuration things
onto an installer, so putting these into a mixin
to be reused within both livecd and installer
seems the way to go.
This time it autostarts using livecd-fgfs and primus
if possible; firefox and GUI mixer are the notable loss
but the clarity of "boot into FlightGear" should sort of
compensate for that.
Ah, and Tu-154 by default.
Current Sisyphus' xorg-drv-intel works somewhat better
with recent kernel drivers on my HD4000 GPU, and icewm
is not compositing at all; providing another test/backup
image fitted with newer kernel should do no harm.
This package has been built and recommended by cas@;
it requires Qt5 which hasn't been needed for anything else
included in regular builds so far so let's extend kde4 one
to begin with.
lightdm isn't going to turn off the system properly
with no systemd-logind around ("for no good reason",
that is); good ol' wdm for installed system and the
similarly ol' autologin just work though.
nodm is not gonna cut it since user PATH is weird
within the session breaking livecd-install by putting
/usr/sbin before /usr/bin while it shouldn't be there
in the first place.
As it happens regular-rc testing has shown that cinnamon,
gnome3 and kde4 flavours included NM via their pkglists
and dependencies (which used to result in live feature
enabling NetworkManager service wholesale when found);
now when handling default services has become more strict
it became apparent that these images have got their LiveCD
mode running without network by default (installation does
set that up though).
It looks like an easy way to just stick +nm into .regular-desktop
dependencies but then razorqt, sugar, xmonad would get NM which
is not what they're gonna handle; e17/e18 too.
Well actually it shouldn't -- except for rEFInd the boot manager:
branding graphics within the build environment are used to add
a single background image to EFI/refind/icons/ thus the change.
Wonder how this got lost though as this screenshot:
http://en.altlinux.org/File:Altlinux-rescue-uefi-memtest86.jpg
clearly illustrates it was working back in December indeed!
TDE images are pretty modest regarding resource consumption
thus suitable for older hardware; a slower flash drive can
stall indefinitely showing slideshow and not going any further
with actual package installation so let's put a cap on that.
This change is done to reduce ambiguity in some cases;
the previous intention has been to ease navigation when
staying in a particular directory, now it's been changed
in favour of convenient toplevel `git grep' in fact.
Both variants have their pros and cons, I just find myself
leaning to this one by now hence the commit. Feel free to
provide constructive criticism :)
Some path-related bitrot has also been fixed while at that.
Well, some of the maintainers clearly prefer t7/branch
to publish their works; at least GNUstep and TDE packages
are updated there and might migrate to p7/branch later.
No need to fight that, really.
KVM and VirtualBox support packages are pretty tiny
but essential when these images get deployed within
virtual environments for any reason, let's add 'em.
It's been gfxboot-free but no user visible facility to select locale
has emerged through these years; it's been decided to put gfxboot
until some text chooser is available (thanks aen@ for discussion).
This is a minimalistic ALT-based system installer tailored
for those who know how to bring up networking and apt-get
the packages they actually need; thanks frbrgeorge@ for
proposing the specification as well as sem@ and glebfm@
for discussion.
No mc, no glibc-locales, even no man and interactivesystem!
Packages included: apt basesystem openssh vim-console
PS: Sisyphus-based regular build is not the main goal though
thus the p7/branch {bri,klu}dge.
This image is largely a rebase of server-ovz.iso onto regular-server;
it's not feasible to provide a single image that would install either
"mini" server or openvz/kvm one based on user choice during boot alas
(even if both ovz-el and std-def kernels are provided within "ovz" ISO
and vzctl&co could be stuffed into a package list/group).
Maybe this is fixed some day...
OpenVZ related part is now a reusable use/server/ovz target,
and service related groups which have been largely taken from
rider@'s server-light project are now use/server/groups/base.
TerraSync might come handy (just as online manuals) but one's
going to need internet access for that so let's put at least
DHCP-over-Ethernet configuration preset in.
There's a beautiful airliner model out there thanks to the guys
at flightgear.ru, and it was asking to be included but its unclear
licensing status; now that 3.0 is GPLed I'm glad to add this package.
This might belong to test.mk actually but it's been instrumental
in getting bumblebee support operational within these LiveCDs;
icewm and sysvinit are a commonplace among those currently
but aren't set in stone for that matter.
I'm fed up with graphical software occasionally making it
into regular-rescue.iso and bloating it for no good reason
given that window managers or xinitrc aren't included.
Dank Bagryantsev asked if it could be added to available packages
at least; well it is there now but not in default install
as aptitude is currently unsupported.
This kernel can help save almost 50 megabytes of image size
and shave off several megs of RAM consumption as well which
is important after the installation has been through.
Adding rescue image was requested by Speccyfighter (in Russian):
http://lists.altlinux.org/pipermail/community/2013-December/681045.html
...and it seems hard but doable if one doesn't mind barebone rescue;
still efi-shell shouldn't spoil x86_64 build as that one won't fit CD-R
and doesn't have to anyways.
un-def got unsuitable due to initial ramdisk migrating from
initramfs to tmpfs by default in newer kernels, and propagator
was using pretty kludgy way to determine that /dev has been
mounted already; led-ws (and supposedly lks-wks) have stumbled
upon this earlier.
20130822 version has been fixed regarding that.
...instead of installer-distro-desktop which pulls in
alterator-auth which breaks things big time for sysv-tde
installer image due to avahi-daemon lazy to run.
Of course it's the last step of installation that has to fail.
And I've been considering this for several months already anyways.
It's proprietary now but still very useful with no free software
alternatives for UEFI platform so far; let's include efi-memtest86
into the rescue image at least.
distro/.regular-install depends on use/luks now too;
this isn't a hard decision but so far looks good
given the overall functionality range and balance
within regular builds.
I've been considering a way to avoid confusion between:
- a tde based livecd with systemd;
- a tde based livecd with sysvinit;
- a tde based installer with sysvinit
and finally came to conclusion that regular-sysv prefix
will be common for installers with sysvinit within regular.mk
and p7.mk; this might be not perfect but should be good enough.
Note that while regular-sysv-tde.iso is buildable and installs
just fine at the moment I don't plan to publish Sisyphus based
installer builds as a rule since these require extra knowledge
regarding daily use (starting with http://altlinux.org/changes
and/or https://lists.altlinux.org/mailman/listinfo/sisyphus ML
subscription).
Sisyphus-based rescue image is fine as well as LiveCDs are;
installable LiveCDs (most of regular-*.iso) are actually
risky in case user actually installs *and* updates those
having ignored the red "unstable" status in the branding,
and that's the line I'm not going to cross that often.
This image family doesn't inherit from distro/.installer
and thus could miss the bootloader setup with no problems
(at least until an installed system would attempt booting
without a bootloader).
The whole thing requires some more thinking over and probably
moving more bits to mixins but in the mean time let's make sure
the bootloader *is* configured.
distro/.regular-install is now factored out to be reused in
tde based installer (and potentially more images later on).
This implies sysvinit at the moment which might change too
but looks just fine right now.
What was added for networked desktop installers is now needed
for regular desktop installers as well; moved to a mixin.
These will probably get reshaked at some point.
