Sunday, September 17, 2017

OpenIndiana Hipster: Console workaround, revised audioctl script, and an extended middle finger to Oracle


This year has turned out to be one in which I've been happily knee deep in OpenBSD (6.0, 6.1 and -current), only venturing into illumos territory on the rarest occasions, either as a quick sensory fix via the OpenIndiana 2016.10 snapshot live DVD or a two- or three-day fling with Tribblix. Otherwise, I've been using OpenBSD on this machine (a Dell Inspiron 1525) on a steady basis for productivity and creativity with no sense of missing apps or features. (So there, Linux!)

This all changed recently with initial unconfirmed reports of Oracle having gutted its Solaris team down to a skeleton crew for maintenance purposes, which had been interpreted in some quarters to mean that this awesome operating system - and the de facto "official" UNIX for those on x86 hardware - is being left to wither on life support just long enough for current support contracts to run out, rather than being actively developed and improved. After almost two weeks of keeping mum on the issue, Oracle has finally confirmed the layoffs via a Worker Adjustment and Retraining Notification (WARN) filing in California. As The Register also reports, Oracle has revised its product roadmap to show that Solaris.next (which will mean further Solaris upgrades occurring in a rolling fashion along the 11.x stream) will not be unleashed until 2018, rather than 2017 as originally planned. El Reg further speculates that this future vision of Solaris may entail it being relegated to an emulated cloud offering, or some such.

Um, okay.

I'm somewhat allergic to all that corporate "cloud" blither-blather, as well as the notion of surrendering as many computer services as possible to a remote host, so Oracle's latest cloud rah-rah session doesn't make me feel confident about Solaris' future as a freestanding operating system.

And so, as my way of extending a middle finger to what I initially felt was (and may still prove to be) Oracle's further mishandling of Solaris, I decided to spend some quality time with illumos, the community-driven successor fork of Sun's OpenSolaris, whose throat was slit by the meanies at Oracle, post-acquisition. Over the past week or two, therefore, I've been vacillating between OI Hipster and Tribblix, both as a way of commemorating the venerable Solaris while also showing contempt for its neglectful foster parent. I don't have anything substantively new to say about Tribblix other than it's a total gas to work with, and so for now I'll limit my comments to OI Hipster.

Workaround for system crash when exiting X11 to console

I have no complaints whatsoever about the OI Hipster implementation of the MATE desktop, though if I had my druthers it would default to the Nimbus theme out-of-the-box in honor of its OpenSolaris lineage, but that's a microscopic complaint. (Being a grown-up, and given that changing the theme is beyond effortless, I'm okay with my current state of drutherlessness.)

On the other hand, I start to twitch and then break out in hives if I can't exit to the console every now and then to experience the spartan beauty of UNIX sans all that GUI goop. Unfortunately, when you enter the appropriate command in OI Hipster (pfexec svcadm disable lightdm) on my particular machine, the system exits MATE and simply hangs before getting to the console command line. This also happens when exiting twm, so the issue is with how my particular graphics driver and adaptor handle an X11 shutdown, rather than being specifically a MATE issue. (See my bug report for more details.)

In the meantime, a reasonable workaround is to enter the above command for exiting MATE and let the system crash, and then do a hard shutdown (in my case by holding down the power button until the machine stops writhing and goes limp), and then once you've rebooted add mate-session to your .xinitrc file, which means you start MATE at the command line with startx rather than the graphical login screen. This also means exiting MATE (or any X11 window manager or desktop environment) simply by becoming root and then issuing the reboot command. (From MATE you can either issue the poweroff command or select Shutdown from the System dropdown menu in order to turn off your computer. In a barebones X11 window manager like twm, you're stuck with the former option.

Revised crank script

Some time ago I wrote about the lack of (or otherwise lukewarm) mouse-driven audio control in OI's MATE implementation. My solution then was a workaround via two shell scripts, which I christened crank (for increasing the volume to 100%) and ucrank (for returning the volume to 50%). This time around, I decided to streamline the workaround by condensing it to a single shell script (still called crank) with the following lines:

read a
audioctl set-control -d /dev/sound/audiohd:0mixer volume $a
audioctl set-control -d /dev/sound/audiohd:0mixer headphones $a:$a
audioctl set-control -d /dev/sound/audiohd:0mixer speaker $a:$a


So rather than the bludgeon-like all-or-nothing of crank and ucrank, the user can fine tune the volume with a single command and enter the precise volume level. (I've got an unhealthy lust for tinnitus, so I never settle for less than 100.)

Parting thoughts

I have used OI enough over the past three years (151_a8 and Hipster) that I have developed a deep affection for it, and thus even though OpenBSD remains my natural habitat I can't resist brief or extended sojourns with the operating system (along with its illumos brethren) that carries on the innovative spirit of Sun Microsystems far more than Oracle ever could (or should I say would ever want to).

While projects like OI may suffer from having far fewer developers than the various operating systems based around the Linux kernel (because, after all, there is no "Linux operating system", contrary to what the lazy and careless penguinistas in the media would have you believe), I believe the future is bright for illumos if those overseeing it continue to improve and innovate on the OpenSolaris code and continue to show what's possible with business-led open source projects.

Another factor that may pay dividends for illumos in general is Joyent's acquisition last year by Samsung. Given that one of Joyent's high-profile offerings is SmartOS, a specialized illumos distro, it's hard not to see how the support of a behemoth like Samsung wouldn't trickle down to the overall illumos project.

Wednesday, July 19, 2017

Wednesday, July 12, 2017

Tribblix m20: Getting started and performing a minor upgrade

I've been taking Tribblix (version m.20) for a spin for the first time since writing about it in this space last year, when it was version m.16. I was very impressed by the speed and simplicity of installation and use, particularly for a SunOS 5.11-derived system. My only real complaint at the time was that some of the packages were a little unfinished, particularly that ugly duckling of a desktop environment known as CDE, with whom I am inexplicably smitten. (It's kind of like that odd-looking girl you fell for in grade 9 but were too nervous to approach. All these years later you search out her yearbook picture from that year, mall bangs and all, and wonder to yourself, "What was I thinking?", only to find yourself tossing and turning in bed at 3:46 a.m., obsessing over what could have been. That approximates how I feel about the Common Desktop Environment.)

Four upgrades later, some of the packages are a little more finished (or else are imported directly upstream from illumos or sideways from OpenIndiana). In my previous Tribblix fling I didn't use it (on the technical level) as extensively as I could have, opting merely to poke and prod around the perimeter and simply get a general feel for it. This time around, however, I decided to wade a little deeper in.

Installation

Nothing substantively different to report here from what I've written about previously, other than to say that rather than entering ./live_install.sh -B c1t0d0 kitchen-sink to commence installation, I instead did ./live_install.sh -B c2t0d0 x11 x11-extras retrodesktop. (The difference is due to the "kitchen-sink" option now automatically installing a graphical login that takes you straight to Xfce4, which is fine if that's what you're into. I'd rather have just a plain ol' text login. Also, c2t0d0 is the label Tribblix gives to my Dell Inspiron 1525's hard drive, as opposed to the ThinkPad T60p I was using the previous time.)

Housekeeping

Akin to Solaris 10, the Tribblix install process doesn't give you the opportunity to create a new user until after you boot into your new system. In this case, your initial post-installation login credentials are the same as for the live DVD, which means logging in as "jack" with an eponymous password, and switching to root with "tribblix" as the password. Therefore, my first order of business was fourfold:
  1. add myself as a user
  2. assign root privileges to myself,
  3. create a password for my own account, and 
  4. change the root password.
This was accomplished the old-fashioned way at the command prompt, using my own credentials for the sake of demonstration:
$ su
Password: tribblix
# useradd -d /export/home/jed -m -s /usr/bin/ksh -c "James Deagle" jed
# usermod -P "Primary Administrator" jed
# passwd jed
New Password: __
Re-enter new Password: __
# passwd passwd
New Password: __
Re-enter new Password: __
(That last step, "passwd passwd", is the command for changing the root password.)

The only other housekeeping matter that I would normally take care of so soon after first booting into a new system is to create my own /bin directory for storing any scripts or programs written by Yours Truly: 
$ pwd 
/export/home/jed 
$ mkdir bin 
$ cd bin
$ pwd 
/export/home/jed/bin

Packages

I'm not going to rewrite my instructions on using this system's handy-dandy package manager (called zap), because then I'd have no reason to force you to read my previous post on the topic. I will say, however, that I highly recommend you install the pkgsrc overlay (courtesy of Joyent), as it will give you an additional boatload of packages (albeit some duplicates of native binaries from Tribblix/illumos). Just how many? That's a great question...let's ask UNIX:
$ pkgin avail | wc -l
16066
Onwards and up(grade)wards...

I don't know if this counts as a proper segue, but the zap package management utility is also used to carry out binary upgrades of the system itself. As explained on the Tribblix home page and paraphrased here, the steps are as follows:
# zap refresh,
# zap update TRIBzap-upgrade
# zap upgrade list (to see if an upgrade is available)
# zap upgrade m20.1

And to boot into your newly-upgraded system:
# beadm activate m20.1
# init 6

Final Thoughts

Especially considering that this illumos distro is the work of just one man, Tribblix is a beauty to behold, and also serves as Solaris in a Hurry for someone who isn't into putting their entire life on hold in order to get an OS installed. (The older I get, the more I fall into that category.) The fact that some of the packages still have some rough edges is entirely a moot point - Tribblix provides a framework and proof-of-concept for the idea of an OpenSolaris derivative that is small and fast, and makes an excellent starting point for those wishing to spin their own distros, whether along the same lines as Tribblix or off in some new and unforeseen direction.



Formal complaint to Toronto Sun: Khadr editorial amounts to unlawful accusation

The following is a formal complaint submitted to the Toronto Sun on July 12, 2017.

_____


I am writing to submit a formal complaint regarding your July 9, 2017 editorial, Khadr payoff a slap in the face to all who serve, and will escalate the matter to the National NewsMedia Council should the results of the current complaint submission process prove unsatisfying.

In the editorial in question, you refer to Canadian citizen Omar Khadr as "(an) expert bombmaker and the killer of U.S. Army Sgt. Christopher Speer." Nowhere does this piece mention that there were no eyewitnesses to the alleged crime, that the process by which Mr. Khadr "confessed" to the killing of Sgt. Speer involved intense physical and psychological torture (including threats of anal rape), nor that the tribunal in front of which he was tried without legal representation has since been discredited. Furthermore, the editorial neglects to mention that Mr. Khadr, who was 15 at the time of Sgt. Speer’s killing, was indeed a child soldier according to internationally-recognized guidelines, for which the threshold is 18. Also, the editorial omits the fact that Mr. Khadr was raised by a family with direct jihadist involvement, and thus grew up subject to what can only be called brainwashing, which itself is psychological abuse of a different sort, and the key ingredient to the making of any child soldier. (Although he was 15 at the time of his alleged crime, one can only assume that his inculcation began much earlier, and that even if he wanted to leave such circumstances he likely didn’t feel like he had a lot of say in the matter.)

For the reasons stated above, and given that the way in which Mr. Khadr’s detention and tribunal were managed would have had the case thrown out in any other circumstances, I believe Postmedia is engaging in unlawful allegation by labelling Mr. Khadr a “killer” without any attempt at providing a reasonable context, and thus also misleading readers to unwittingly assume that Mr. Khadr was convicted in a process that would satisfy the threshold in our own domestic legal system.

I consider this conduct unlawful on Postmedia’s part, not to mention an abuse of its power as a nationwide media conglomerate, and as such I would like to see your organization issue a retraction of the above quoted passage, as well as a formal apology to Mr. Khadr and to readers who deserve more responsible behavior from their news media.

Friday, July 7, 2017

OpenBSD 6.1: XFCE 4 posing as MATE, the joys of syspatch, doas mount and no-fi connectivity

About seven months have elapsed since the last time I shared any episodes from my ongoing UNIX adventures. This prolonged silence, methinks, was probably due to the fact that I was on a single installation (OpenBSD) almost the whole time, which is quite a feat considering my distrohopping tendencies. I was so busy using UNIX for my own day-to-day tasks, rather than endlessly reinstalling and tweaking entire systems, that I simply felt like I had nothing to say, UNIX-wise. (The important thing to note here, for those who are quick to write OpenBSD off as some "security only" distro with no desktop potential, is that I have been getting along just fine using it as an all-purpose desktop system for tasks that were often multimedia in nature. So there!)

Within the past few days, however, I had succumbed to some old bad habits, and took OpenIndiana Hipster for a spin after an extended absence. I installed from the 2016.10 iso and then was able to patch it up to whatever was current as of July 2. And while some bugs and hiccups (not to mention a lack of some of my favorite multimedia apps) sent me scurrying back to OpenBSD before long, I nevertheless enjoy the overall feel and aesthetics of that system's iteration of the MATE desktop environment (which itself is a successor fork of Gnome 2).



OpenIndiana Hipster's default MATE desktop, as shown 
from a 2016.10 live DVD session

Once I had OpenBSD (version 6.1) back up and running, I installed the XFCE 4 desktop environment with the intention of going for the default configuration (as usual), but after an errant mouse click found myself faced with the prospect of configuring it from scratch, one panel at a time. Still longing for OpenIndiana's aesthetic, I proceeded to see just how much I could tweak XFCE 4 until it bore a passing resemblance to OI's default MATE desktop configuration. (One essential part of this task included downloading and installing the "Nimbus" window theme. It also helps that the default typefaces for xfce4-terminal, Monospace Regular 12-point, is the same as that for mate-terminal.) After a few hours that felt more like 20 minutes, here is the fruit of my labor:



OpenBSD 6.1 running XFCE 4, tweaked to somewhat resemble MATE.


Akin to an Elvis impersonator, the above looks like like Solaris/OpenIndiana MATE if you squint your eyes and don't look too closely at the applications button. (I should note here that I haven't tried installing MATE on OpenBSD yet, as I'm not 100% certain yet that a polished version is available on that system. Nevertheless, it appears to be in the "work-in-progress" stage.)

And for anyone who has fiddled with Oracle Solaris 11.x, all it takes is a simple change of background wallpaper to make the above look even more familiar:



And the "Desktop Least Likely To..." award goes to OpenBSD 6.1
pretending to be Oracle Solaris 11.3.

The above pic is the exact configuration of my XFCE desktop as of this writing. The visual aesthetics of Oracle Solaris crossed with the speed of OpenBSD is sending conflicting information to my brain, though not in an entirely bad way. (File under "Recreational High".)

Odds and ends

syspatch - As of OpenBSD 6.1, you can use the syspatch command to scan for system patches (syspatch -c) or to simply let 'er rip and apply all available patches (syspatch with no options).

doas mount/umount - While running OpenBSD 6.0, I had been using hotplug-diskmount for mounting and unmounting of usb drives/devices. This time around (after reinstalling from scratch with 6.1 DVD iso), I somehow couldn't get that utility to work, but came across a simple method in an excellent online article, Installing OpenBSD 6.1 on your laptop is really hard (not). The steps, as shown by that author and followed by me, are to first append the following lines to /etc/doas.conf (or create a new file with those lines if it doesn't yet exist)

permit nopass jed as root cmd mount
permit nopass jed as root cmd umount

Next, create a directory from which to access your USB content (in my case /home/jed/usb), and then mount or unmount as follows:

$ doas mount /dev/sd1i /home/jed/usb
$ doas umount /dev/sd1i

Brother, can you spare a compatible wifi card? - I love my current laptop, but the one blind spot is that it uses a Broadcom BCM4215 wifi card which, as far as I can tell after much Googling, only works with Microsoft Windows. (That may not be entirely true - it appears Ubuntu has a driver, firmware-b43legacy-installer, but my particular card may not be covered.)


Thursday, June 15, 2017

Don't taint Orthodox Christian ecumenicism with lies of omission

A June 1 article in the National Catholic Reporter demonstrates how to mislead (and therefore manipulate) an unwitting reader by either downplaying important facts, or in this case omitting them altogether.

In Orthodox not interested in reunion with Rome, Jesuit Fr. Thomas Reese tries to make the case that Orthodox bishops in Eastern Europe, as well as rank-and-file Orthodox Christians, are not interested in reuniting with the Church of Rome, and are therefore less ecumenical than their Roman Catholic counterparts.

One of the fundamental problems with Fr. Thomas' article is that it does not establish a frame of reference for what reunification would look like, aside from the Vatican being "open to a less intrusive role for the pope in the Eastern churches than in the West." (Does "less intrusive" mean that the Eastern churches would be exempt from papal primacy, or does it mean the pope would simply take a softer approach to how he rules over the East while allowing Orthodox bishops to refer to him as a mere patriarch?)

Fr. Thomas also neglects to mention if Orthodox Christians would be expected to budge from their own theological positions (where they differ from those of Rome), or if East and West would simply agree to disagree on such matters.

On that topic, Fr. Thomas writes that "the touchy issue has always been the role of the papacy", thus disregarding altogether one of the primary theological divisions, namely the filoque in the Nicene Creed. While to some it may seem like a matter of splitting hairs, whether one believes the Holy Spirit proceeds from the Father only (as the Orthodox believe) or from the Father and the Son (as Roman Catholics believe) nevertheless points to a stark difference in how one views the Holy Trinity. Neither side should be expected to "reunite" with the other at the expense of such a core belief.

It is very telling that the author employs a Reality Distortion Field worthy of Steve Jobs by writing that the two sides "accept the same Nicene Creed", and leaving it at that. Intentional or not, this amounts to a lie of omission in the service of over-simplifying (and thus distorting) the true reasons behind the appearance of Orthodox resistance to reunion with Rome. Such a lie (that being the impression Fr. Thomas creates that there is no dispute over the Nicene Creed) also makes it easier to paint Orthodox Christianity as being less ecumenical than Roman Catholicism, with the unwritten implication that the Orthodox are somehow more close-minded because of it.

One dimension left unexplored by Fr. Thomas is the possibility that perhaps Orthodox Christians, along with their bishops, feel some sort of trepidation at being a smaller fish potentially being eaten by a much larger one. Looking at it through the other end of the telescope, I suspect that to Roman Catholics, reunion with Orthodox Christianity may feel like an opportunity for expansion of their own interests. (I have a hard time believing that Roman Catholics would view their post-reunion pope as "just one of the guys" among the other patriarchs.) Likewise, for Orthodox Christians that same reunion may feel like the beginning of a typically Western incursion, albeit in priestly garb, and under the pretense of ecumenicism.

Whether this last point is a factor in the results of the Pew Research Center study quoted by Fr. Thomas is merely speculation by Yours Truly. Nevertheless, it points to an issue that is far too complicated to allow for accusations that Orthodox Christians are less ecumenical than Roman Catholics, and whatever such a generalization would seem to imply about them.

As for me, ecumenicism and Christian unity between East and West shouldn't mean the same thing as "sameness", nor should it require outright reunion between Orthodoxy and Roman Catholicism. There are enough ideals and issues around which members of both sides can passionately lock arms in solidarity without the need to merge their respective organizational structures, and thus risk losing that which makes each unique and beautiful.


Thursday, June 1, 2017

Don't burden Wonder Woman with American jingoism

The following is a letter to the editor of the New York Post.

_____

 
Dear Editor,

Maureen Callahan's appraisal of the upcoming Wonder Woman movie (Why 'Wonder Woman' is less American than ever, June 1, 2017) seems to center on the extent to which it fails to be a propaganda exercise for the cause of American Exceptionalism. In the process, she betrays the national insecurity and defeatism at the heart of mindless flag-waving for its own sake.

America is already great in many ways, even if its sons and daughters have been repeatedly deceived into harm's way in the name of (or within the highly-charged milieu of) patriotism. Think of the Gulf of Tonkin "incident" that was fabricated by elites to lure unwitting Americans into the Vietnam war (which didn't actually have anything to do with "stopping communism"), or the "pre-emptive" oil war in Iraq that was waged by George W. Bush under the false pretense of the "Weapons of Mass Destruction" that were already known not to exist.

When will more Americans realize that the sense of "patriotism" they've been trained to feel is actually a form of mind control that keeps them supporting people and causes that are diametrically opposed to their own interests? When will they stop offering themselves up for manipulation by politicians such as President Trump, whose numbskulled appeals to patriotism of the most jingoistic variety were designed to turn the crowds at his campaign rallies into arenas full of angry barking dogs?

Patriotism in and of itself is a complicated and contradictory affair, and so I do not share Ms. Callahan's dismay that today's Wonder Woman doesn't honor her roots as wartime propaganda with an hourglass figure.

Finally, Ms. Callahan concludes her piece by asking "Why can't Wonder Woman make America great again?", which inadvertently amounts to the author bashing her own country. Looking at the bigger picture, a country that is truly great doesn't need to constantly preen and flex its muscles for the benefit of the rest of the world - it simply needs to let its actions speak for themselves.

James Deagle
Ottawa, Canada