Buildroot Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Peter Korsgaard <peter@korsgaard.com>
To: "Yann E. MORIN via buildroot" <buildroot@buildroot.org>
Cc: "José Luis Salvador Rufo" <salvador.joseluis@gmail.com>,
	"Yann E. MORIN" <yann.morin.1998@free.fr>,
	"Edgar Bonet" <bonet@grenoble.cnrs.fr>
Subject: Re: [Buildroot] [PATCH 1/1] support/scripts/check-merged: shift args processed by getopts
Date: Sun, 30 Nov 2025 18:06:26 +0100	[thread overview]
Message-ID: <87tsybh2lp.fsf@dell.be.48ers.dk> (raw)
In-Reply-To: <aSv3lJlD1bNlcRGs@landeda> (Yann E. MORIN via buildroot's message of "Sun, 30 Nov 2025 08:51:48 +0100")

>>>>> "Yann" == Yann E MORIN via buildroot <buildroot@buildroot.org> writes:

 > José, All,
 > On 2025-11-29 21:34 +0100, José Luis Salvador Rufo spake thusly:
 >> Shift the arguments processed by getopts so that only the root directories
 >> to check remain as arguments for the subsequent `for root; do` loop.

 > The actual issue this triggers is that shecks in custom skeleton cause
 > build failure:

 >     support/scripts/check-merged -t skeleton -u -b /usr/src/simplek8s/rootfs-skeleton
 >     The skeleton in -t is not properly setup:
 >      - /usr/bin should exist, be a directory, and not be a symlink
 >      - /usr/lib should exist, be a directory, and not be a symlink
 >     The skeleton in skeleton is not properly setup:
 >      - /usr/bin should exist, be a directory, and not be a symlink
 >      - /usr/lib should exist, be a directory, and not be a symlink
 >     The skeleton in -u is not properly setup:
 >      - /usr/bin should exist, be a directory, and not be a symlink
 >      - /usr/lib should exist, be a directory, and not be a symlink
 >     The skeleton in -b is not properly setup:
 >      - /usr/bin should exist, be a directory, and not be a symlink
 >      - /usr/lib should exist, be a directory, and not be a symlink
 >     [...]

 > The reason this was not caught earlier, is that the script was only
 > tested against overlays, not custom skeletons. The difference between
 > skeletons and overlays, is that the tests are strict for skeletons, and
 > lax for overlays. Thus, with a custom skeleton, where the merged
 > symlinks are missing, the tests go wrong and report am incorrect issue,
 > while with overlays, the checks fails, but they are not critical for
 > obverlays, so  the script is silent and suceeds.

 > I think a better commit log would be:

 >     support/scripts/check-merged: fix checking custom skeletons

 >     When using a custom skeleton where the merged symlinks ar emising,
 >     the build fails with errors like:

 >         support/scripts/check-merged -t skeleton -u -b /usr/src/simplek8s/rootfs-skeleton
 >         The skeleton in -t is not properly setup:
 >          - /usr/bin should exist, be a directory, and not be a symlink
 >          - /usr/lib should exist, be a directory, and not be a symlink
 >         The skeleton in skeleton is not properly setup:
 >          - /usr/bin should exist, be a directory, and not be a symlink
 >          - /usr/lib should exist, be a directory, and not be a symlink
 >         [...]

 >     Commit 793ebd5d2809 (support/scripts/check-merged: use getopts
 >     instead of getopt) intoduced a flawed use of getopts: unlike getopt,
 >     getopts does not conume the positional arguments. This causes the
 >     check for directrory validity to all check each option as if they
 >     were directories.

 >     For overlays, this is transparently ignored, because the checks are
 >     only lax for overlays (missing symlinks are OK).

 >     However, for skeletons, the checks are strict. Because of that, a
 >     missing symlink is considered an error, when it should be considered
 >     as being OK.

 >     The fix is to actually consume the positional args to only keep the
 >     list of directpries to validate, like is done for example in
 >     support/download/dl-wrapper.

 >     Fixes: 793ebd5d28095c8df45a0d183d273d8a84b3f0a4
 >     Reviewed-by: Yann E. MORIN <yann.morin.1998@free.fr>
 >     Acked-by: Yann E. MORIN <yann.morin.1998@free.fr>
 >     Cc: Edgar Bonet <bonet@grenoble.cnrs.fr>
 >     Signed-off-by: José Luis Salvador Rufo <salvador.joseluis@gmail.com>

 > Thank you José for reporting and fixing the issue! Thank you Edgar for
 > giving a helping hand!

Committed with Yanns commit message (minus a few typos), thanks.

-- 
Bye, Peter Korsgaard
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

      reply	other threads:[~2025-11-30 17:06 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-11-29 20:34 [Buildroot] [PATCH 1/1] support/scripts/check-merged: shift args processed by getopts José Luis Salvador Rufo
2025-11-30  7:51 ` Yann E. MORIN via buildroot
2025-11-30 17:06   ` Peter Korsgaard [this message]

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=87tsybh2lp.fsf@dell.be.48ers.dk \
    --to=peter@korsgaard.com \
    --cc=bonet@grenoble.cnrs.fr \
    --cc=buildroot@buildroot.org \
    --cc=salvador.joseluis@gmail.com \
    --cc=yann.morin.1998@free.fr \
    /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