Re: Update of wip/sddm to version 0.21

●To: Pkgsrc Users <pkgsrc-users%netbsd.org@localhost> ●Subject: Re: Update of wip/sddm to version 0.21 ●From: Stepan Ipatov <st.ipatov%gmail.com@localhost> ●Date: Fri, 2 Jan 2026 02:51:30 +0100

On 12/30/2025 7:07 PM, David H. Gutteridge wrote:

> Have you considered sending the patches upstream? (I ask this being a
> guilty party who sometimes doesn't do this; the scale can get
> unmanageable quickly.)


Thank you for the kind words.

Yes, I am planning to send the patches upstream to the sddm developers in the near future.


In the meantime, could you please update the package from wip once again?

I've found a rather annoying issue: when sddm is started during system boot, the keyboard does not work. It happens due to the way sddm selects the virtual TTY for starting the X server. It calls ioctl(fd, VT_OPENQRY, &vt) and gets the first free TTY. The problem is that at the stage of system boot when the rc scripts are executed, the ioctl() call returns /dev/ttyE1. Apparently, sddm and getty do not coexist well on the same TTY. The only workaround I've found so far is to delay starting sddm until init has already spawned getty. Once that happens, ioctl() returns the expected /dev/ttyE4.


I've committed an updated version to wip/sddm. Only two files were changed:

 * The sddm.sh rc script

* patch-src_common_VirtualTerminal.cpp — fixed the TTY device naming scheme, since wscons does not support more than 8 devices



--
Thanks in advance,
  Stepan Ipatov

●References: ●Re: Update of wip/sddm to version 0.21 ●From: David H. Gutteridge

●Prev by Date: diff from 2025-12-23 22:39 to 2025-12-29 01:06 ●Next by Date: Re: python313 (2025Q4) fails to build (link) on hppa -current ●Previous by Thread: Re: Update of wip/sddm to version 0.21 ●Next by Thread: Re: Update of wip/sddm to version 0.21 ●Indexes: ● reverse Date ● reverse Thread ● Old Index Home | Main Index | Thread Index | Old Index