Re: [PATCH 12/20] host/rootfs: Do not read from /dev/tty1

[Alyssa Ross <hi@alyssa.is>]

[-- Attachment #1: Type: text/plain, Size: 1426 bytes --]

Demi Marie Obenour <demiobenour@gmail.com> writes:

> On 9/19/25 14:22, Alyssa Ross wrote:
>> Demi Marie Obenour <demiobenour@gmail.com> writes:
>> 
>>> On 9/8/25 05:19, Alyssa Ross wrote:
>>>> Demi Marie Obenour <demiobenour@gmail.com> writes:
>>>>
>>>>> This breaks debugging because data written goes both to Weston (or its
>>>>> subprocesses) and to getty.
>>>>>
>>>>> Signed-off-by: Demi Marie Obenour <demiobenour@gmail.com>
>>>>
>>>> Can you give some more detail?
>>>
>>> I was having problems logging in until I made this change.
>>> login was complaining that the username I passed was invalid,
>>> which turned out to be because it included only a subset of
>>> the characters I typed.
>> 
>> Okay, interesting.  I haven't seen that.  I think I did this because the
>> example systemd unit for Weston[1] sets StandardInput=tty-fail.  What do
>> you think of that?
>> 
>> [1]: https://wayland.pages.freedesktop.org/weston/toc/running-weston.html#running-weston-from-a-systemd-service
>
> In that case, I think it is best to make sure any child
> processes Weston spawns redirect stdin to /dev/null.

That's not an upstream behavior, though, so I wonder why nobody has
encountered this problem before.  Weston seems to still work fine after
your change, so I'm fine with applying your patch as is.  I'm just
really confused why the upstream recommendation is like that in that
case.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 227 bytes --]