Re: CTWM and XDM

Tue, 19 Aug 2025 23:58:13 -0700

 
 
----------------
To: [email protected] ([email protected]);
Subject: CTWM and XDM;
20.08.2025, 07:51, "BC HABS" <[email protected]>:
+ ctwm@
- all
 
 
 
----------------
To: The CTWM Mailing List ([email protected]);
Subject: CTWM and XDM;
19.08.2025, 22:04, "Greg A. Woods" <[email protected]>:

At Tue, 19 Aug 2025 11:26:58 +0100, BC HABS <[email protected]> wrote:
Subject: Re: CTWM and XDM



 If it is possible to build *without* workspaces, then I haven found
 it, or to not have that section in the .ctwmrc file and not have
 issues, but having just one workspace defined is no hardship and it
 seems to have fixed the issues being experienced


It is impossible to compile ctwm without workspace support.

However not defining any workspaces should just work.

On a quick walk-through of the code the only thing I noticed was that
the flag variable field used to indicate whether any workspaces are
defined or not is not explicitly initialized in the InitScreenInfo().
However since it's a plain normal boolean flag and the struct is
allocated with calloc(), the default should be "false" anyway.

I just tried running my latest build of ctwm with "-f /dev/null", which
will of course avoid defining any WorkSpaces, and it works just fine. I
can also send it a SIGHUP and it restarts just fine.

BTW one can test ctwm in an Xnest display with relative ease provided
one allows for controlling which window manager is run from one's
~/.xinitrc through, e.g., an environment variable. For example I do
this with the following commands:

    (
        cd;
        WM="/work/woods/f-ctwm/build-NetBSD-x86_64/work/woods/f-ctwm/ctwm -v -f /dev/null" \
        XNESTDISPLAY=$display \
        xinit -- $X11PATH/bin/Xnest;
        cd -
    )

It's helpful to have your ~/.xinitrc detect when it is run from a
terminal and redirect all output to a file for analysis too, though
without doing so you'll see any errors wherever you run xinit from.
 

--
                                        Greg A. Woods <[email protected]>

Kelowna, BC +1 250 762-7675 RoboHack <[email protected]>
Planix, Inc. <[email protected]> Avoncote Farms <[email protected]>
 
Good day and thank you for the comments.
 
...
>> I just tried running my latest build of ctwm with "-f /dev/null", which
>> will of course avoid defining any WorkSpaces, and it works just
>> fine.
...
 
--dumpcfg  will show the built in internal cfg ... this is the cfg used if no file is specified or can be found on the usual paths.
 
On my system, that internal cfg has a 'workspaces' defined.
 
Whether specifying  -f /dev/null causes the internal cfg to be used I am unclear (for what it is worth, an AI earch enquiry seemed to inficate it does use the internal built-in cfg if -f /dev/null is used).
 
Does this mean my reported experience is backed up ... I dont know, I cant say.
 
Regards
Habs
 
 
Further to my earlier post above ... (and I should have tried before posting),  despite the AI results I mentioned,   it does not look like from testing that the  -f /dev/null uses the internal cfg.    It does indeed seem to just start ctwm   with no workspaces (or menus or basic setup etc that the internal cfg indicates).   My error ... sorry.
 
I am not sure then, when I keep getting the 'crashes' as originally reported when not using a workspaces entry ... perhaps other settings in the cfg file  rely on   a workspaces entry somewhere in the code.
 
I will experiment further based on the - f /dev/null  advice etc.
 
THank you
Habs

Reply via email to