[PATCH v3 3/4] dlls/{ntdll, kernel*}: provide a pseudo console environment for initial CUI processes not tied to an Unix tty

[Jacek Caban]

On 4/11/22 12:42, Eric Pouech wrote:
>> BTW, I'd also consider WARN() instead of once+FIXME() for 
>> console_ioctl().
>
> the rationale behind this one was to be able to track when some 
> program ends up using Console* interface while being attached to the 
> shell_no_screen console type
>
> this could be complicated to track or understand
>
> (actually an ERR would be perhaps more suited than a FIXME) 


ERR() is usually for cases that should not happen. Those ioctl()s may 
legitimately happen and in some cases they may even be desired bevavior, 
so I'm not sure that any debug channel that is turned on by default is 
appropriate. I'd hope that retval=0 from a console function call in 
relay logs will not be too hard to spot (and then +console could clarify 
that with a WARN()).


Thanks,

Jacek
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.winehq.org/pipermail/wine-devel/attachments/20220411/5af4d8e6/attachment.htm>