From: Henning Schild <henning.schild@siemens.com>
To: Jan Kiszka <jan.kiszka@siemens.com>
Cc: isar-users <isar-users@googlegroups.com>
Subject: Re: [PATCH 1/7] dpkg-raw: Respect file permissions defined by recipe
Date: Mon, 7 Jan 2019 18:51:53 +0100 [thread overview]
Message-ID: <20190107185153.48ff9b83@md1za8fc.ad001.siemens.net> (raw)
In-Reply-To: <73e2f06f-9ece-1c7c-739f-b572a109179c@siemens.com>
Am Mon, 7 Jan 2019 17:54:26 +0100
schrieb Jan Kiszka <jan.kiszka@siemens.com>:
> On 07.01.19 17:28, Henning Schild wrote:
> > Am Mon, 7 Jan 2019 15:26:16 +0100
> > schrieb Jan Kiszka <jan.kiszka@siemens.com>:
> >
> >> On 07.01.19 15:20, Jan Kiszka wrote:
> >>> On 07.01.19 15:19, Henning Schild wrote:
> >>>> Am Mon, 7 Jan 2019 14:28:47 +0100
> >>>> schrieb Jan Kiszka <jan.kiszka@siemens.com>:
> >>>>
> >>>>> On 07.01.19 14:20, Henning Schild wrote:
> >>>>>> Am Wed, 2 Jan 2019 12:34:11 +0100
> >>>>>> schrieb Jan Kiszka <jan.kiszka@siemens.com>:
> >>>>>>> From: Jan Kiszka <jan.kiszka@siemens.com>
> >>>>>>>
> >>>>>>> dh_fixperms overwrites the permissions do_install defined
> >>>>>>> carefully. Skip this step to avoid that.
> >>>>>>>
> >>>>>>> Fixes: f301ccb2b5b1 ("meta/dpkg-raw: build raw packages like
> >>>>>>> all others") CC: Henning Schild <henning.schild@siemens.com>
> >>>>>>> Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
> >>>>>>> ---
> >>>>>>> meta/classes/dpkg-raw.bbclass | 4 +++-
> >>>>>>> 1 file changed, 3 insertions(+), 1 deletion(-)
> >>>>>>>
> >>>>>>> diff --git a/meta/classes/dpkg-raw.bbclass
> >>>>>>> b/meta/classes/dpkg-raw.bbclass index 8d11433..10fb1b9 100644
> >>>>>>> --- a/meta/classes/dpkg-raw.bbclass
> >>>>>>> +++ b/meta/classes/dpkg-raw.bbclass
> >>>>>>> @@ -56,9 +56,11 @@ EOF
> >>>>>>> deb_create_rules() {
> >>>>>>> cat << EOF > ${S}/debian/rules
> >>>>>>> #!/usr/bin/make -f
> >>>>>>> +
> >>>>>>> +override_dh_fixperms:
> >>>>>>> +
> >>>>>>> %:
> >>>>>>> dh \$@
> >>>>>>> -
> >>>>>>
> >>>>>> I think it is not a good idea to do that in general. While you
> >>>>>> might have found an example where dh_fixperms caused problems,
> >>>>>> there are probably many where it helps. Say people use "cp" to
> >>>>>> fill ${D} or "echo" to fill ${D}/bin/
> >>>>>
> >>>>> I'm open for better suggestions.
> >>>>
> >>>> The suggestion is to do that in the one recipe that you need it
> >>>> for, and not touch the general case.
> >>>
> >>> ...except for causing that regression: Keep in mind that we used
> >>> to respect permissions defined by the user before the switch to
> >>> packaging via Debian!
> >
> > True, but there is a changelog section that even tells users how to
> > disable certain dhs for their recipes.
> >
> >> To make my issue more concrete: Consider you want to package
> >> secrets this way. Then it would be rather ugly to even temporary
> >> have them group or even work readable during packaging and
> >> installation - in case you suggestion should be to adjust the
> >> permissions in a postinst.
> >
> > Having secrets in your repo and build process would be ugly as well,
> > many spots where they could leak. So i do not think that is a good
> > example.
> > And i am not talking about a postinst, but a rules file that does
> > exactly what yours does. See what example-raw does for dh_usrlocal,
> > if you bring your rules you do not get the defaults.
> > Looking at the man-page i see a lot of "removes permission", where
> > documentation seems to be the only exception. Again secret does not
> > seem to be a good example. (except you place it in
> > usr/share/doc ;) )
> >
> > What exactly is your motivation for the change?
>
> Allow to ship files that are not world-readable by defaults. That's a
> pretty common pattern, e.g. to add pre-generated keys, certificates,
> wifi passwords etc.
I think i got that but i am not sure how they would become world
readable, not from the man-page nor from the code. Except you got the
location "wrong". So a full path example with the permissions before
and after is what i was asking for.
Henning
> So I don't think it is a good idea that dpkg-raw now breaks this use
> case, sometimes silently(!), and forces users to overload their rules
> files. I'm not even sure that it makes sense for Debian to add these
> permissions to during the fixperms phase, but I didn't dig into that
> details.
>
> Jan
>
next prev parent reply other threads:[~2019-01-07 17:51 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-01-02 11:34 [PATCH 0/7] Vacation hacks Jan Kiszka
2019-01-02 11:34 ` [PATCH 1/7] dpkg-raw: Respect file permissions defined by recipe Jan Kiszka
2019-01-07 13:20 ` Henning Schild
2019-01-07 13:28 ` Jan Kiszka
2019-01-07 14:19 ` Henning Schild
2019-01-07 14:20 ` Jan Kiszka
2019-01-07 14:26 ` Jan Kiszka
2019-01-07 16:28 ` Henning Schild
2019-01-07 16:54 ` Jan Kiszka
2019-01-07 17:51 ` Henning Schild [this message]
2019-01-07 18:56 ` Jan Kiszka
2019-01-09 15:37 ` Baurzhan Ismagulov
2019-01-10 1:19 ` Jan Kiszka
2019-01-15 8:19 ` Jan Kiszka
2019-01-15 10:10 ` Baurzhan Ismagulov
2019-01-15 10:13 ` Jan Kiszka
2019-01-02 11:34 ` [PATCH 2/7] bootimg-efi-isar: Run grub-mkimage without own config Jan Kiszka
2019-01-07 13:22 ` Henning Schild
2019-01-07 14:02 ` Jan Kiszka
2019-01-02 11:34 ` [PATCH 3/7] bootimg-efi-isar: Retrieve boot disk during runtime Jan Kiszka
2019-01-02 11:34 ` [PATCH 4/7] isar-cfg-localepurge: Also set /etc/default/locale Jan Kiszka
2019-01-02 11:34 ` [PATCH 5/7] Fix LIC_FILES_CHKSUM paths Jan Kiszka
2019-01-02 11:34 ` [PATCH 6/7] Move repo variables to proper conf file Jan Kiszka
2019-01-02 11:34 ` [PATCH 7/7] Move debian distro conf files to meta Jan Kiszka
2019-01-04 13:49 ` [PATCH 8/7] bootimg-efi-isar: Fix serial console setup for grub Jan Kiszka
2019-01-04 13:50 ` [PATCH 9/7] bootimg-efi-isar: Reformat generated grub.cfg Jan Kiszka
2019-01-07 6:06 ` [PATCH 10/7] bitbake.conf: Clean up and enhance OVERRIDES Jan Kiszka
2019-01-07 8:17 ` Claudius Heine
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=20190107185153.48ff9b83@md1za8fc.ad001.siemens.net \
--to=henning.schild@siemens.com \
--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