public inbox for isar-users@googlegroups.com
 help / color / mirror / Atom feed
From: "Roberto A. Foglietta" <roberto.foglietta@gmail.com>
To: Jan Kiszka <jan.kiszka@siemens.com>
Cc: Michael Adler <michael.adler@siemens.com>, isar-users@googlegroups.com
Subject: Re: [PATCH] feat: add zstd support to image conversions
Date: Tue, 24 Jan 2023 04:17:54 +0100	[thread overview]
Message-ID: <CAJGKYO5qxEdBEWFsRLq-1KuE8+2VRbBNYfCfdKLdBa=PL-YDow@mail.gmail.com> (raw)
In-Reply-To: <6d7ab6e6-6260-6a85-587e-9609e9314645@siemens.com>

[-- Attachment #1: Type: text/plain, Size: 1461 bytes --]

Il Lun 23 Gen 2023, 20:22 Jan Kiszka <jan.kiszka@siemens.com> ha scritto:

> On 23.01.23 15:00, Michael Adler wrote:
> > This patch enables ISAR to build zstd compressed images.
> >
>
> I was briefly thinking you had solved the dpkg vs. zstd issue - just a
> dream :)
>

Hi Jan.

I think you are speaking about zstd compression for the debian package that
is 1% bigger but 13x times faster in expanding. [1]

Ubuntu and Arch adopted zstd but not debian which is still using xz and as
far as I have seen neither in parallel using option -T.

I think you should ask them to produce all package with

export XZ_DEFAULTS='-T 8'

and enabling dpkg to set internally

XZ_DEFAULTS='-T 0'

unless it is differently set. What Do you think about changing ISAR to use
zstd or xz -T without repackaging everything at fetch time?

It would be a solution: fetch, repackage with zstd and the downloads remain
for a long time like a permanent cache and also the benefits.

The counter effect is that those debian packages and those produced by ISAR
will be not anymore compatible with a standard debian and this could be a
nasty surprise in some cases for some users.

The alternative is to repackage everything with xz parallel. Usually a
modern laptop has 8 pipes but I have the sensation that with -T 0 it counts
just the cores because on my machine spawn 4 not 8.

Best regards, R-

[1]
https://archlinux.org/news/now-using-zstandard-instead-of-xz-for-package-compression/

[-- Attachment #2: Type: text/html, Size: 2303 bytes --]

  reply	other threads:[~2023-01-24  3:18 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-01-23 14:00 Michael Adler
2023-01-23 19:22 ` Jan Kiszka
2023-01-24  3:17   ` Roberto A. Foglietta [this message]
2023-01-24 12:05     ` Roberto A. Foglietta
2023-02-09 11:35 ` Uladzimir Bely

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='CAJGKYO5qxEdBEWFsRLq-1KuE8+2VRbBNYfCfdKLdBa=PL-YDow@mail.gmail.com' \
    --to=roberto.foglietta@gmail.com \
    --cc=isar-users@googlegroups.com \
    --cc=jan.kiszka@siemens.com \
    --cc=michael.adler@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