[blfs-support] Problem with gcc-6 and qt--5.6

Ken Moffat zarniwhoop at ntlworld.com
Sat May 21 17:47:27 PDT 2016

On Sat, May 21, 2016 at 06:56:44PM -0500, Bruce Dubbs wrote:
> It does not 'appear' to be volatile, it *is* volatile.
> $ grep run /etc/fstab
> tmpfs          /run         tmpfs    defaults            0     0
> See LFS Section 8.2.
Well yes, but the last time I looked on a desktop there seemed to be
some other directories in /run which did get created.  At the moment
I'm on my netbook with an old system, but I see directories in /run
for ConsoleKit, cups, dbus, laptop-mode-tools, lock, mount,
pm-utils, udev, wpa_supplicant so a lot of things manage to write

> I have in my ~/.bash_profile
> export XDG_RUNTIME_DIR=/run/plasma-runtime-bdubbs
> I suppose a generic value would be
> export XDG_RUNTIME_DIR=/run/plasma-runtime-$USER
> That type of thing is not mentioned in the book.  Where do you think it
> should be?  Possibly configuration for xdg-utils, but plasma may be
> appropriate.

It was in the context of "there was a message from plasma in the
logs when I tried to work out why kde did not work for me" and
"messages when looking at that gstreamer package which broke parole
on AMD for me" (vaapi, I think).

I put that into .xinitrc on the machine where I was doing that, but
of course export does not run mkdir,

> Actually I do not think that XDG_RUNTIME_DIR is strictly needed.  It just
> prints a warning and creates the file in /tmp.
>   -- Bruce
If it is needed (for the moment, I'm clutching at straws trying to
find out why plasma is broken for me), ~/.bash_profile seems
unlikely - will it get read, and early enough, when using sddm ?

If it is only a warning, it doesn't really matter.  I know plasma is
slow and bloated, but when it has worked for me it was an
interesting alternative approach.

This email was written using 100% recycled letters.

More information about the blfs-support mailing list