public inbox for isar-users@googlegroups.com
 help / color / mirror / Atom feed
From: Alexander Smirnov <asmirnov@ilbers.de>
To: Jan Kiszka <jan.kiszka@siemens.com>,
	isar-users@googlegroups.com, Baurzhan Ismagulov <ibr@ilbers.de>
Cc: Claudius Heine <ch@denx.de>
Subject: Re: [PATCH 2/2] centralize multistrap configuration generation
Date: Thu, 15 Feb 2018 13:29:11 +0300	[thread overview]
Message-ID: <63ff2a0a-4303-1449-e2e4-41e2049aba7c@ilbers.de> (raw)
In-Reply-To: <a26c7f8f-ffd5-d395-fa45-8be699ba8de4@siemens.com>

On 02/15/2018 01:15 PM, Jan Kiszka wrote:
> On 2018-02-15 11:05, Claudius Heine wrote:
>>>> While this moves us in the right direction, it still leaves some
>>>> duplications behind (setup.sh) - and the big question if we should
>>>> invest any further in this horrible multistrap mess.
>>>>
>>>> I've just enabled arm64, and it took a while to figure out that I was
>>>> missing some lines in setup.sh. In contrast, qemu-debootstrap just
>>>> worked (I used it to debug the multistrap output).
>>>>
>>>> So, before we move on here: What all does multistrap currently give us,
>>>> compared to standard debootstrap? I only see "multiple repositories" so
>>>> far. But as Claudius is going to massage that part heavily anyway, I
>>>> really wonder if it's worth building on the multistrap indirection any
>>>> further.
>>>>
>>>> How about something like this?
>>>>
>>>> - [qemu-]debootstrap --variant=minbase ...
>>>> - inject all additional apt sources, including user-provided ones
>>>> - set pinnings and priority, including user-provided ones
>>>> - apt-get dist-upgrade
>>>>
>>>
>>> I'd say that multistrap is a tool which maps user-friendly config file
>>> to specific apt repositories configuration. If you look in the
>>> do_build() log for buildchroot, you could see what it exactly does:
>>>
>>> 1. Creating list of sources in rootfs/etc
>>>
>>> 2. Get packages.gz for all the sources:
>>>
>>> apt-get  -o Apt::Architecture=armhf -o
>>> Dir::Etc::TrustedParts=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/etc/apt/trusted.gpg.d
>>> -o
>>> Dir::Etc::Trusted=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/etc/apt/trusted.gpg.d/trusted.gpg
>>> -o Apt::Get::AllowUnauthenticated=true -o Apt::Get::Download-Only=true
>>> -o Apt::Install-Recommends=false -o
>>> Dir=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/
>>> -o
>>> Dir::Etc=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/etc/apt/
>>> -o
>>> Dir::Etc::Parts=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/etc/apt/apt.conf.d/
>>> -o
>>> Dir::Etc::PreferencesParts=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/etc/apt/preferences.d/
>>> -o APT::Default-Release=* -o
>>> Dir::State=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/var/lib/apt/
>>> -o
>>> Dir::State::Status=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/var/lib/dpkg/status
>>> -o
>>> Dir::Cache=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/var/cache/apt/
>>> update
>>>
>>> 3. Install packages:
>>>
>>> apt-get -y  -o Apt::Architecture=armhf -o
>>> Dir::Etc::TrustedParts=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/etc/apt/trusted.gpg.d
>>> -o
>>> Dir::Etc::Trusted=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/etc/apt/trusted.gpg.d/trusted.gpg
>>> -o Apt::Get::AllowUnauthenticated=true -o Apt::Get::Download-Only=true
>>> -o Apt::Install-Recommends=false -o
>>> Dir=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/
>>> -o
>>> Dir::Etc=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/etc/apt/
>>> -o
>>> Dir::Etc::Parts=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/etc/apt/apt.conf.d/
>>> -o
>>> Dir::Etc::PreferencesParts=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/etc/apt/preferences.d/
>>> -o APT::Default-Release=* -o
>>> Dir::State=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/var/lib/apt/
>>> -o
>>> Dir::State::Status=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/var/lib/dpkg/status
>>> -o
>>> Dir::Cache=/home/builder/isar/build/tmp/work/raspbian-jessie-armhf/buildchroot/rootfs/var/cache/apt/
>>> install ${LIST_OF_PACKAGES}
>>>
>>>   > That would kick out all the multistrap complications.
>>>
>>> For me the plan above intersects with these multistrap steps, you have
>>> to re-implement them from the scratch. In this case I'd vote for the
>>> community-proved solution.
>>
>> For me that is a reason to use debootstrap instead of multistrap, since
>> multistrap is not community-supported. If we implement some of those
>> feature, we might as well do it in debootstrap inself. Just take a look
>> at their TODO list:
>>
>>    Features:
>>      ++ second stage via chroot debootstrap/debootstrap
>>      ++ debootstrap/deb file to record deb destinations/information
>>
>>      -- configuration file
>>          -- versus command line
>>          -- support for sources (vs mirrors)
>>          -- faux-pinning for packages
>>
>> https://anonscm.debian.org/cgit/d-i/debootstrap.git/tree/TODO?id=5d42162bc0fe266b280d81445271e3f56b9560a8
>>
>>
>> That shows to me that upstream might be interested in some of these
>> improvements we would need to implement.
> 
> Exactly. We should really use the community-proven, continuously tested
> official approach of Debian, in that in a way that it foresees for it.
> 
> Let's bury multistrap. Its a dead horse we rode to long now.
> 

Nobody is against official tools. :-) This question is open for a quite 
long time, and the conclusion for now: there is *no* tool that fits Isar 
requirements for 100%. So in general I only want upstream-first way of 
tools modification for Isar.

Alex

Alex


  reply	other threads:[~2018-02-15 10:29 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-02-14 13:15 [PATCH 0/2] Consolidate " claudius.heine.ext
2018-02-14 13:15 ` [PATCH 1/2] meta/classes/base: extend sys.path with lib directory claudius.heine.ext
2018-02-15  7:28   ` Jan Kiszka
2018-02-14 13:15 ` [PATCH 2/2] centralize multistrap configuration generation claudius.heine.ext
2018-02-15  7:28   ` Jan Kiszka
2018-02-15  8:16     ` Claudius Heine
2018-02-15  8:50       ` Jan Kiszka
2018-02-15  9:34     ` Alexander Smirnov
2018-02-15 10:05       ` Claudius Heine
2018-02-15 10:15         ` Jan Kiszka
2018-02-15 10:29           ` Alexander Smirnov [this message]
2018-02-15 11:27           ` Baurzhan Ismagulov
2018-02-15 13:17             ` Claudius Heine
2018-02-15 13:18             ` Jan Kiszka
2018-02-15 11:42   ` Baurzhan Ismagulov
2018-02-15 12:08     ` Claudius Heine
2018-02-15 12:37       ` Alexander Smirnov
2018-02-15 13:20         ` Claudius Heine
2018-02-15 13:39           ` Alexander Smirnov
2018-02-16  9:31             ` Claudius Heine
2018-02-16 10:35               ` Alexander Smirnov
2018-02-16 11:35                 ` Jan Kiszka
2018-02-15 13:53           ` Jan Kiszka

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=63ff2a0a-4303-1449-e2e4-41e2049aba7c@ilbers.de \
    --to=asmirnov@ilbers.de \
    --cc=ch@denx.de \
    --cc=ibr@ilbers.de \
    --cc=isar-users@googlegroups.com \
    --cc=jan.kiszka@siemens.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox