:: Re: [DNG] window manager fails to c…
Top Page
Delete this message
Reply to this message
Author: Ralph Ronnquist
To: dng
Subject: Re: [DNG] window manager fails to connect to XServer
On Wed, Mar 27, 2024 at 12:46:17PM -0400, Haines Brown via Dng wrote:
> Ralph, I've made no progress. X log halts with defintion of compose
> key and then nothing. I eliminated the command to define the compose
> key, but that didn't help.
> When you suggeted I put the pgrep lines in ~/.initrc it made good
> sense but the terminal from which I ran startx did not disply the
> test lines the resulting linex. My ~/.initrc looks like this:
> #!/bin/sh
> #
> # ~/.xinitrc
> #
> # Executed by startx (run your window manager from here)
> # line for test only
> pgrep -a Xorg >&2
>   if [ -d /etc/X11/xinit/xinitrc.d ]; then
>     for f in /etc/X11/xinit/xinitrc.d/*; do
>       [ -x "$f" ] && . "$f"
>     done
>     unset f
>   fi

> # two lines for test only
> sleep 2
> pgrep -a Xorg >&2
> exec startfluxbox
> The pgrep line run in a terminal works as expcted.
>     $ pgrep -a Xorg >&2
>     7871 /usr/lib/xorg/Xorg -nolisten tcp :0 vt1 -keeptty -auth 
>         /tmp/serverauth.8EwqRRpvho

Ok this confuses me greatly. Which terminal are you running that in?
Because it shows that Xorg is running, i.e. you have started Xorg, and
it is running.

> In order to see what the terminal reports when I startx, I log 
> in. At the $ prompt enter startx. When default window manager comes 
> up. I kill X to see what the terminal reported.      

Is this on a different VT while Xorg is running on vt1?
What's "the default window manager"?

> When you provided option (a) (nothing returned), I presume the normal
> lines resulting from startx are shown, but no additional lines due
> to pgrep.

Previously you showed an uoutput that included 1 pgrep line? Has this

I had thought the flow was the floowing:
1) boot the machine... and it comes up with a "login:" prompt on vt1
2) log in as non-root user
3) "$ startx"

But, it sounds to me you are doing something else.

Otoh, "startx" would happily start second (third, fourth ...) Xorg in
different VT yielding different DISPLAY settings (:1.0 :2.0 ...) but
it's not unheard of that <u>window managers</u> make the assumption
that there is only one Xorg with display 0:0 and then get terminally
unhappy when they can't interact with that.

If I read your paragraphs right. "startx" starts and runs with an Xorg
display taking over and covering... so it doesn't exit. It exits
when the window magager terminates. So the question whether it runs or
not is answered(?).

Further, given the ~/.xinitrc above, it also shows that fluxbox is
running. Because if it would terminate, then Xorg would terminate and
then VT would change back to text mode and the startx stderr output
would show up. As you say, that doesn't happen until you quit the
window manager.

Though, perhaps your paragraph about that paragraph relates to having
an edited ~/.xinitrc where the last line is "exec startfxce4" or

All in all, I think your results tell that either a) fluxbox comes up
and stays up until forcefully terminated, or b) fluxbox comes up but
terminates for some reason. In both cases Xorg comes up and stays up
until the window manager is terminated.

I thought we were in case b, and it shows that the reason of fluxbox
not being able to connect to the running Xorg belongs to that fluxbox
startup (and not that Xorg terminating for its own reasons).

Unfortunately I don't use fluxbox myself but I could possibly spin up
a qemu to explore it. As I'm in the midlle of an overseas move right
now (only across Bass Straight but evenso) it might take a moment, and
perhaps someone already running fluxbox could chip in with some
guidance. I do think it's pretty clear that the issue in your very
first error log does belong to fluxbox and not Xorg. (That's not a
"blame assignment" but rather for the purpose of focusing the bug


> _______________________________________________
> Dng mailing list
> Dng@???
> https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng