A clarification about (Linux) Mesa / Nouveau Drivers

Two of the subjects which I like to blog about, are direct-rendering and Linux graphics drivers.

Well in This Earlier Posting, I had essentially written, that on the Debian 9 , Debian /Stretch computer I name ‘Plato’, I have the ‘Mesa’ Drivers installed, and that therefore, that computer cannot benefit from OpenCL, massively-parallel GPU-computing.

What may confuse some readers about this is the fact that elsewhere on the Internet, there is speak about ‘Nouveau’ Drivers, but less so about Mesa Drivers.

‘Mesa’, which I referred to, is a Debian set of meta-packages, that is all open-source. It installs several drivers, and selects the drivers based on which graphics hardware we may have. But, because ‘Plato’ does in fact have an nVidia graphics card, the Mesa package automatically selects the Nouveau drivers, which is one of the drivers it contains. Hence, when I wrote about using the Mesa Drivers, I was in fact writing about the Nouveau Drivers.

One of the reasons I have to keep using these Nouveau Drivers, is the fact that presently, ‘Plato’ is extremely stable. There would be some performance-improvements if I was to switch to the proprietary drivers, but making the transition can be a nightmare. It involves black-lists, etc..

Another reason for me to keep using the Nouveau Drivers, is the fact that unlike how it was years ago, today, those drivers support real OpenGL 3, hardware-rendering. Therefore, I’m already getting partial benefit from the hardware-rendering which the graphics card has, while using the open-source driver.

The only two things which I do not get, is OpenCL or CUDA computing capabilities, as Nouveau does not support that. Therefore, anything which I write about that subject, will have to remain theoretical for now.

I suppose that on my laptop ‘Klystron’, because I have the AMD chip-set more-correctly installed, I could be using OpenCL…

Also, ‘Plato’ is not fully a ‘Kanotix’ system. When I installed ‘Plato’, I borrowed a core system from Kanotix, before Kanotix was ready for Debian / Stretch. This means that certain features which Kanotix would normally have, which make it easier to switch between graphics drivers, are not installed on ‘Plato’. And that really makes the idea daunting, to try to switch…

Dirk

Installing a “Wacom” graphics / digitizer tablet under Linux.

I’ve just received my “Wacom Intuos PT S 2″ digitizer tablet – aka graphics tablet – which I had specifically bought, because there is some support for the Wacom series of tablets under Linux. I was able to get it working 100%, and also did get the ‘wireless kit’ to work, that I had ordered with it.

Under Linux, we need to have at least these two packages installed, in order to get this hardware to work:

• ‘xserver-xorg-input-wacom’
• ‘libwacom2′

If available, the following should also be installed:

• ‘libwacom-bin’

Additionally I should mention that I did this on the Debian / Stretch computer I name ‘Plato’, on which I have Plasma 5 as my desktop manager. Although Debian and Linux do support Wacom, apparently, Plasma 5 as such does not. This effectively means that individual applications may have tablet-support, but that We’re not given a graphical module for the System Settings panel, from which to customize our tablet. For that reason, I needed to write a shell-script which does this for me:

#!/bin/bash

xsetwacom --set "Wacom Intuos PT S 2 Finger touch" Touch off
xsetwacom --set "Wacom Intuos PT S 2 Pen stylus" Button 2 key "shift"
xsetwacom --set "Wacom Intuos PT S 2 Pen stylus" Button 3 key "ctrl"

xsetwacom --set "Wacom Intuos PT S 2 Pad pad" Button 1 key "button +3"
xsetwacom --set "Wacom Intuos PT S 2 Pad pad" Button 3 key "del"

xsetwacom --set "Wacom Intuos PT S 2 Pad pad" Button 8 key "ctrl" "c"
xsetwacom --set "Wacom Intuos PT S 2 Pad pad" Button 9 key "ctrl" "v"

xsetwacom --set "Wacom Intuos PT S 2 (WL) Finger touch" Touch off
xsetwacom --set "Wacom Intuos PT S 2 (WL) Pen stylus" Button 2 key "shift"
xsetwacom --set "Wacom Intuos PT S 2 (WL) Pen stylus" Button 3 key "ctrl"

xsetwacom --set "Wacom Intuos PT S 2 (WL) Pad pad" Button 1 key "button +3"
xsetwacom --set "Wacom Intuos PT S 2 (WL) Pad pad" Button 3 key "del"

xsetwacom --set "Wacom Intuos PT S 2 (WL) Pad pad" Button 8 key "ctrl" "c"
xsetwacom --set "Wacom Intuos PT S 2 (WL) Pad pad" Button 9 key "ctrl" "v"



Presumably, any other users with Plasma 5 desktop-managers will need to write similar scripts. Simply typing in the command ‘xsetwacom’ without any arguments, will display its basic usage.

It should also be noted, that after an X-server module has been installed, that will act as the input driver, at the very least, the X-server also needs to be restarted, before that driver will be loaded.

(Updated 2/18/2019 : )

Revisiting HTML, this time, With CSS.

When I first taught myself HTML, it was in the 1990s, and not only has the technology advanced, but the philosophy behind Web-design has also changed. The original philosophy was, that the Web-page should only contain the information, and that each Web-browser should define in what style that information should be displayed. But of course, when Cascading Style-Sheets were invented – which in today’s laconic vocabulary are just referred to as “Styles” – they represented a full reversal of that philosophy, since by nature, they control the very appearance of the page, from the server.

My own knowledge of HTML has been somewhat limited. I’ve bought cuspy books about ‘CSS’ as well as about ‘JQuery’, but have never made the effort to read each book from beginning to end. I mainly focused on what some key concepts are, in HTML5 and CSS.

Well recently I’ve become interested in HTML5 and CSS again, and have found, that to buy the Basic license of a WYSIWYG-editor named “BlueGriffon“, proved informative. I do have access to some open-source HTML editors, but find that even if they come as a WYSIWIG-editor, they mainly tend to produce static pages, very similar to what Web-masters were already creating in the 1990s. In the open-source domain, maybe a better example would be “SeaMonkey“. Beyond that, ‘KompoZer‘ can no longer be made to run on up-to-date 64-bit systems, and while “BlueFish”, a pronouncedly KDE-centric solution available from the package-manager, does offer advanced capabilities, it only does so in the form of an IDE.

(Updated 03/09/2018, 17h10 : )

KdeWallet, and Using Smb4K, under Plasma 5

One fact which I’ve written about before, is that I have an up-to-date Linux computer, that uses the ‘Plasma 5′ desktop manager, which is actually the successor to ‘KDE 4′. When using this desktop manager, we can still install numerous packages that ‘belong’ to the old, KDE 4, and most of them will continue to work. One of those is ‘smb4k’, which is a point-and-click utility, to mount a network SMB share – aka, a Windows-file-share, such that it will be visible in our home folder, as though that share was a local sub-folder.

There exist command-line methods to do the same thing, which would mount that network share, and declare it’s of the ‘cifs’ file-system-type, but the use of a simple GUI to do so may be easier.

But then one problem which ensues, is that Smb4K will use the KDE 4 Wallet, to store our password, for that share. It will function in this way, by depending on the package ‘kde-runtime’. In truth, this latter package probably pulls in numerous (old) KDE 4 libraries, and not just the old KWallet, but the existence of this KDE 4 Wallet, on our Plasma 5 machine, is most obvious…