Re: [OE-core] [PATCH v2] pseudo: Update to 1.9.3 release

[Mark Hatle <mark.hatle@kernel.crashing.org>]

I do think we should take it, but it does require the dpkg change as RP mentions 
below.

This should fix the known openat2 issue with RHEL9 and similar systems, and then 
maybe deal with issues elsewhere as well.  (realpath, statx and other differences)

--Mark

On 1/20/26 4:31 AM, Richard Purdie via lists.openembedded.org wrote:
> On Tue, 2026-01-20 at 09:59 +0100, Yoann Congal wrote:
>> Le mar. 20 janv. 2026 à 09:52, Richard Purdie via
>> lists.openembedded.org
>> <richard.purdie=linuxfoundation.org@lists.openembedded.org> a écrit :
>>> Pulls in the following changes:
>>>
>>>    Makefile.in: Bump version to 1.9.3
>>>    configure: Minor code quality changes
>>>    pseudo: code quality scan - resolved various potential issues
>>>    makewrappers: improve error handling and robustness
>>>    Update COPYRIGHT files
>>>    ports/linux/pseudo_wrappers.c: Call the wrappers where possible
>>>    ports/linux/pseudo_wrappers.c: Workaround compile error on Debian
>>> 11
>>>    ports/linux/pseudo_wrappers.c: Reorder the syscall operations
>>>    ports/unix/guts/realpath.c: Fix indents
>>>    pseudo_util.c: Skip realpath like expansion for /proc on Linux
>>>    test/test-proc-pipe.sh: Add test case for proc pipes
>>>    ports/unix/guts/realpath.c: realpath fails if the resolved path
>>> doesn't exist
>>>
>>> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
>>
>> Do you think I should take this for stable inclusion? (leaning toward
>> yes)
> 
> Good question. Copying Mark.
> 
> I suspect it will be easier all around if we did take it. There are a
> lot of fixes in here and it is hard to know which ones any given user
> might need or run into. There isn't any real feature work, just various
> fixes for new glibc calls or existing code. The failure modes for
> pseudo are problematic for end users who wouldn't understand many of
> them. The realpath/proc fixes are the most invasive piece.
> 
> I would suggest we wait a little bit before taking it, just so we can
> test things with master and ensure there is are no regressions.
> 
> You would need to take the dpkg ADMINDIR change as well as the pseudo
> changes won't work without that.
> 
> I'm open to other thoughts either way...
> 
> Cheers,
> 
> Richard
> 
> 
> 
> -=-=-=-=-=-=-=-=-=-=-=-
> Links: You receive all messages sent to this group.
> View/Reply Online (#229656): https://lists.openembedded.org/g/openembedded-core/message/229656
> Mute This Topic: https://lists.openembedded.org/mt/117360386/3616948
> Group Owner: openembedded-core+owner@lists.openembedded.org
> Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [mark.hatle@kernel.crashing.org]
> -=-=-=-=-=-=-=-=-=-=-=-
>