Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

current xfce, xfce4-terminal

17 views
Skip to first unread message

Ewald Pfau

unread,
May 22, 2013, 7:04:34 PM5/22/13
to

Thanx there is a fine xfce terminal, giving me decent access for most of
what I do with computing stuff (starting with mutt & tin, most editing via
joe in emacs-mode, etc). From time to time it is as well nice to have
Thunar (after having convinced it, that there is really nothing to do for it
in automatted ways) and occasionally Orage (but I haven't reminders stored
in it).

But I do not run xfce. Ever since, it's nice to have GNUstep running
instead.

****

Now I just entered a bulk-upgrade for everything, which arrived the other
day for slackware-current.

After upgrade, the xfce4-terminal was gone.

Not exactly: I still had the window , some bit less in height, but no
letters in it, only empty rectangles instead. I have the colors configured
black on white, this remained the same.

I thought, maybe it needed some kind of initialisation by xfce? So I fired
that one up, but all of xfce had that same syndrome, no letters, only empty
rectangles.

(Btw. xfce complained that 'Desktop' is not writable. That's fine with me.
This is an empty file with rights of zero, owned by root.)

So I stepped back, rewinded upgrades of xfce. Terminal still would not work.

So I rewinded all upgrades with the latest back-ups, as kept via rsync (this
is from better not just to delete when synchronising, but moving old
versions of what has been upgraded to a secure place).

[[ Just for the record, when not keeping track of kde, it looks like this:
# rsync -auvPb --delete # --backup-dir=$wt/../.g-r-del # --partial-dir=$wt/../.g-r-part # --exclude slackware/kde/* --exclude slackware/kde/ # --exclude slackware/kdei/* --exclude slackware/kdei/ # --exclude source/kde/* --exclude source/kde/ # --exclude source/kdei/* --exclude source/kdei/ # rsync://$wh/slackware/$ws/ $wt ;
with reasonable values of $wt, $ws and $wh ]]

Now everything was running again.

Now I upgraded xfce-path, by its own, this made xfce work with a readable
font. But now, xfce4-terminal would not start at all any more, complaining
about some initialisation deficits.

Now I just inserted the previous version of xfce4-terminal (with the new
xfce), and voilà, it would run again.

****

There was no reboot in between. I once started xfce via xdm in root console,
where it usually is via startx (to GNUstep) from user console (being able to
switch to xfce from inside there).

While doing some stuff, now it seemed the machine overly quiet, so I looked
at the temperatures (via gkrellm) - now what! this mounted! Normal
operation on a lame day makes for some 40 to 60 Celsius degrees, now it was
80 Celsius degrees (for the Fahrenheit fans: with 100 degrees, water is
boiling), with CPU load being less than 20 percent.

So I ressorted to the classical approach to cure a broken computer and
rebooted it.

It seems like xfce had mixed up somehow the acpi-or-whatever-behaviour and
had quietly left behind an insane state!

One should warn about this!

Terminal and thunar and orage are nice to have, anyway.
But I'll better not touch xfce for now.

****

Besides I'd like to insert here as well, that for my purpose, the kernels
since 3.9.2 seem to be broken. One may find hints to the broken behaviour,
experienced by others, when asking some search engine starting with
something like 'kworker'.

In effect, the worker stuff somehwere deep inside the kernel is running in
kind of wild circles, permanently, making for 15 percent of real CPU load,
all the time (in my case - others reported for up to nearly 100 percent,
rendering the machine useless then for any purpose).

$(uname -a) now says: 3.8.13 #1 SMP PREEMPT Sun May 19 22:16:20 CEST 2013
i686 Intel(R) Core(TM)2 CPU T7400 @ 2.16GHz GenuineIntel GNU/Linux

If anyone asks, I'll append the .config's here or by mail.
It's made to fit to an hp nc4400 portable.
Similarly, there is stripped .config's for Pentium-M for an hp nc4200.

Until some few weeks before, I had 3.9.1 running nicely on the latter
machine with no problems. Since then I migrated everyday stuff from one
machine to the other (nc4200 has PATA drives, for which it is getting a bit
hard to find replacement now - but ok, that machine had served more than 4
years, day by day, without any problems, hardly ever switched off.)

One should not shoot oneself into one's foot - at the end of migrating,
somehow I've managed to throw away all versions of the adopted and nicely
stripped .config for the 'new' machine, for which I had a stable kernel
running, then, but could not reproduce that one anymore, only after some
extra sessions, now, somehow. But the lost one has always been the better
one, in such a case.

Stephen Morgan

unread,
May 23, 2013, 10:20:52 AM5/23/13
to
The rectangle thing in XFCE is caused by pango, you presumably didn't
run slackpkg install-new before upgrade-all. The remedy is to run

slackpkg reinstall pango harfbuzz

Ewald Pfau wrote

Ewald Pfau

unread,
May 23, 2013, 7:08:43 PM5/23/13
to
Stephen Morgan <str...@darkstar.example.org>:

> slackpkg reinstall pango harfbuzz

Thank you, reinstall did the trick!

Respectively
< cd slackware/l; upgradepkg --reinstall pango-1.34.1-i486-1.txz >
as I had installed harfbuzz before already manually.

Think of me happily sitting in front of an upgraded machine, as everything
works again after having redone all upgrades in slackware-current (remaining
careful to not approach xfce desktop for the moment, though).

Ewald Pfau

unread,
May 24, 2013, 8:16:50 PM5/24/13
to
>[me yesterday:]

> [..] everything works again after having redone all upgrades in
> slackware-current (remaining careful to not approach xfce desktop for the
> moment, though).

As I saw, in slackware-current one pkg for xfce has changed again, so I gave
it a try - yes, there is no surprises anymore, as I had it yesterday with
rising temperature.

Xfce looks fine now on this laptop as well.

Additionally, I took a bit more thoroughly a look with kernel 3.9.4, which
just arrived, maybe to find how to avoid that nasty erroneous CPU load.

In short: Kernel 3.9.4 is working now as well.

Found a configuration detail for to cure the problem. In ALSA sound, PCI
devices, Intel device, I just accepted the proposed ticks, for being
included, as presented by default, when entering this item newly. Before, I
had switched off in this Intel section all details in support for other
sound devices.

Bingo!

****

( Sound is working for me as well via internal card, as externally via USB.

To have cards numbered in ALSA properly, this may be configured by the names
of the cards as found by <cat /proc/asound/cards>, and the names of the
sound modules, in </etc/moprobe.d/modprobe.conf>, so e.g.:

options snd_hda_intel id=Intel index=1
options snd_usb_audio id=UA25 index=0

will introduce "UA25" as default to the machine, instead of the built-in
soundcard "Intel". This is default then for the browser and other programs,
which look for "card 0" instead of being configured via ALSA name.)

0 new messages