Dan Jones

Getting stuff done with a smart speaker is supposed to be easy, and in most cases, the manufacturers have done a solid job making that a reality. And now Google Home has learned a new trick to make getting stuff done even easier.

As first reported by CNET, the smart speaker now supports handling two commands at the same time. So, instead of having to list off one command after another, you can just say something like, “Okay Google, turn on the TV and what’s the weather?” Your TV will come to life and then the smart speaker will tell you the weather.

The benefit here is that you don’t have to predetermine these commands in the Google Home app. So you can string commands together, like telling Home to turn on some lights and play some music, and the smart speaker should be able to handle that.

Don’t try three commands, though. Google Home will apparently just tell you it can’t handle that.

Google is also planning on rolling out predetermined commands, like telling Home a certain phrase to get a string of commands completed, but this should be one feature that goes a long way to making using Home more streamlined.

(This is a sponsored post.)

Storyblocks is exploding with over 400,000 stock photos, vectors, backgrounds and more! With its user friendly site, massive library to choose from, and fresh new content, there’s no stopping what you can do. All of the content is 100% free from any royalties. Anything you download is yours to keep and use forever! Right now you can get 7 days of free downloads. Get up to 20 photos, icons, and vectors everyday for 7 days. That's 140 downloads free over the course of the 7 days. Click on over and see where your imagination takes you! Start downloading now.

Direct Link to Article — Permalink

​7 Days of Free Stock Images is a post from CSS-Tricks

Recently I had my first experiences with the Windows Subsystem for Linux (WSL), evaluating its potential as an environment for getting work done. This subsystem, introduced to Windows 10 in August 2016, allows Windows to natively run x86 and x86-64 Linux binaries. It’s essentially the counterpart to Wine, which allows Linux to natively run Windows binaries.

WSL interfaces with Linux programs only at the kernel level, servicing system calls the same way the Linux kernel would. The subsystem’s main job is translating Linux system calls into NT requests. There’s a series of articles about its internals if you’re interested in learning more.

I was honestly impressed by how well this all works, especially since Microsoft has long had an affinity for producing flimsy imitations (Windows console, PowerShell, Arial, etc.). WSL’s design allows Microsoft to dump an Ubuntu system wholesale inside Windows — and, more recently, other Linux distributions — bypassing a bunch of annoying issues, particularly in regards to glibc.

WSL processes can exec(2) Windows binaries, which then run in under their appropriate subsystem, similar to binfmt on Linux. In theory this nice interop should allow for some automation Linux-style even for Windows’ services and programs. More on that later.

There are some notable issues, though.

Lack of device emulation

No soundcard devices are exposed to the subsystem, so Linux programs can’t play sound. There’s a hack to talk PulseAudio with a Windows’ process that can access, but that’s about it. Generally there’s not much reason to be playing media or games under WSL, but this can be an annoyance if you’re, say, writing software that synthesizes audio.

Really, there’s almost no device emulation at all and /proc is pretty empty. You won’t see hard drives or removable media under /dev, nor will you see USB devices like webcams and joysticks. A lot of the useful things you might do on a Linux system aren’t available under WSL.

No Filesystem in Userspace (FUSE)

Microsoft hasn’t implemented any of the system calls for FUSE, so don’t expect to use your favorite userspace filesystems. The biggest loss for me is sshfs, which I use frequently.

If FUSE was supported, it would be interesting to see how the rest of Windows interacts with these mounted filesystems, if at all.

Fragile services

Services running under WSL are flaky. The big issue is that when the initial WSL shell process exits, all WSL processes are killed and the entire subsystem is torn down. This includes any services that are running. That’s certainly surprising to anyone with experience running services on any kind of unix system. This is probably the worst part of WSL.

While systemd is the standard for Linux these days and may even be “installed” in the WSL virtual filesystem, it’s not actually running and you can’t use systemctl to interact with services. Services can only be controlled the old fashioned way, and, per above, that initial WSL console window has to remain open while services are running.

That’s a bit of a damper if you’re intending to spend a lot of time remotely SSHing into your Windows 10 system. So yes, it’s trivial to run an OpenSSH server under WSL, but it won’t feel like a proper system service.

Limited graphics support

WSL doesn’t come with an X server, so you have to supply one separately (Xming, etc.) that runs outside WSL, as a normal Windows process. WSL processes can connect to that server (DISPLAY) allowing you to run most Linux graphical software.

However, this means there’s no hardware acceleration. There will be no GLX extensions available. If your goal is to run the Emacs or Vim GUIs, that’s not a big deal, but it might matter if you were interested in running a browser under WSL. It also means it’s not a suitable environment for developing software using OpenGL.

Filesystem woes

The filesystem manages to be both one of the smallest issues as well as one of the biggest.

Filename translation

On the small issue side is filename translation. Under most Linux filesystems — and even more broadly for unix — a filename is just a bytestring. They’re not necessarily UTF-8 or any other particular encoding, and that’s partly why filenames are case-sensitive — the meaning of case depends on the encoding.

However, Windows uses a pseudo-UTF-16 scheme for filenames, incompatible with bytestrings. Since WSL lives within a Windows’ filesystem, there must be some bijection between bytestring filenames and pseudo-UTF-16 filenames. It will also have to reject filenames that can’t be mapped. WSL does both.

I couldn’t find any formal documentation about how filename translation works, but most of it can be reverse engineered through experimentation. In practice, Linux filenames are UTF-8 encoded strings, and WSL’s translation takes advantage of this. Filenames are decoded as UTF-8 and re-encoded as UTF-16 for Windows. Any byte that doesn’t decode as valid UTF-8 is silently converted to REPLACEMENT CHARACTER (U+FFFD), and decoding continues from the next byte.

I wonder if there are security consequences for different filenames silently mapping to the same underlying file.

Exercise for the reader: How is an unmatched surrogate half from Windows translated to WSL, where it doesn’t have a UTF-8 equivalent? I haven’t tried this yet.

Even for valid UTF-8, there are many bytes that most Linux filesystems allow in filenames that Windows does not. This ranges from simple things like ASCII backslash and colon — special components of Windows’ paths — to unusual characters like newlines, escape, and other ASCII control characters. There are two different ways these are handled:

1. The C drive is available under /mnt/c, and WSL processes can access regular Windows files under this “mountpoint.” Attempting to access filenames with invalid characters under this mountpoint always results in ENOENT: “No such file or directory.”

2. Outside of /mnt/c is WSL territory, and Windows processes aren’t supposed to touch these files. This allows for more freedom when translating filenames. REPLACEMENT CHARACTER is still used for invalid UTF-8 sequences, but the forbidden characters, including backslashes, are all permitted. They’re translated to #XXXX where X is hexadecimal for the normally invalid character. For example, a:b becomes a#003Ab.

While WSL doesn’t let you get away with all the crazy, ill-advised filenames that Linux allows, it’s still quite reasonable. Since Windows and Linux filenames aren’t entirely compatible, there’s going to be some trade-off no matter how this translation is done.

Filesystem performance

On the other hand, filesystem performance is abysmal, and I doubt the subsystem is to blame. This isn’t a surprise to anyone who’s used moderately-sized Git repositories on Windows, where the large numbers of loose files brings things to a crawl. This has been a Windows issue for years, and that’s even before you start plugging in the typically “security” services — virus scanners, whitelists, etc. — that are typically present on a Windows system and make this even worse.

To test out WSL, I went around my normal business compiling tools and making myself at home, just as I would on Linux. Doing nearly anything in WSL was noticably slower than doing the same on Linux on the exact same hardware. I didn’t run any benchmarks, but I’d expect to see around an order of magnitude difference on average for filesystem operations. Building LLVM and Clang took a couple hours rather than the typical 20 minutes.

I don’t expect this issue to get fixed anytime soon, and it’s probably always going to be a notable limitation of WSL.

So is WSL useful?

One of my hopes for WSL appears to be unfeasible. I thought it might be a way to avoid porting software from POSIX to Win32. I could just supply Windows users with the same Linux binary and they’d be fine. However, WSL requires switching Windows into a special “developer mode,” putting it well out of reach of the vast majority of users, especially considering the typical corporate computing environment that will lock this down. In practice, WSL is only useful to developers. I’m sure this is no accident. (Developer mode is no longer required as of October 2017.)

Mostly I see WSL as a Cygwin killer. Unix is my IDE and, on Windows, Cygwin has been my preferred go to for getting a solid unix environment for software development. Unlike WSL, Cygwin processes can make direct Win32 calls, which is occasionally useful. But, in exchange, WSL will overall be better equipped. It has native Linux tools, including a better suite of debugging tools — even better than you get in Windows itself — Valgrind, strace, and properly-working GDB (always been flaky in Cygwin). WSL is not nearly as good as actual Linux, but it’s better than Cygwin if you can get access to it.

Before the end of October, Google found itself embroiled in a huge controversy about hamburgers. More specifically, hamburger emoji.

After the release of Android 8.0 Oreo, it became known that Google had a very interesting idea for where the cheese should be located on a hamburger: on the bottom, below the meat patty. Above the meat was the lettuce and tomato, which many wouldn’t argue with. But that cheese location sent many people decrying Google via social media networks.

Seeing how it was so obviously important to so many people, Google’s CEO Sundar Pichai confirmed to the world that he would “drop everything” and fix the issue as quickly as he could. And sure enough, it looks like Pichai has kept his word.

As noted today by Emojipedia, Google has fixed the burger emoji in the recent Android 8.1 Oreo beta update. As you can see from the image above, the cheese has been moved to above the meat patty, sitting just under the tomato and below the piece of lettuce.

All is right with the world again.

Google has also “fixed” the beer emojis. In Android 8.0, the stein of beer was only half full and yet the top of the big glass was covered in froth, which didn’t make any sense at all. In Android 8.1 Oreo, those beer steins are completely full of liquid so the froth makes more sense.

The updated emoji will be available in Android 8.1, which Google just released the last developer preview for and will see a public launch in December.

Tactical Holiday Stocking – Have a holly MOLLE Christmas!

TriceraTACO Dinosaur Taco Holder

Set-top boxes are designed to give owners a one-stop shop for most of the things they would want to do on their TV, including apps to games. And now for Amazon Fire TV owners, that also includes browsing the web.

Today Amazon officially launched its Silk web browser for select Fire TV models as well as the Fire TV streaming stick and the Westinghouse/Element 4K television that has Amazon’s Fire TV software built in. The new app is available to download through the app stores on those devices.

It’s worth noting here that only the first- and second-generation Fire TV set-top boxes support the Silk browser, but that the newest model does not support the web browser yet. Amazon says an update coming in December will address that issue, so if you’ve got the newest Fire TV set-top box, you’ll have to wait a bit longer to get this functionality.

The Silk web browser will let you navigate the internet right from your TV, and according to AFTVnews, the in-browser video playback is pretty solid. Otherwise it sounds like a pretty straightforward web browser. You’ll even be able to add bookmarks for quick navigation.

Is web browsing on your TV something you want to do?

wtf human
www.patreon.com/Sephko

Walmart is an example of a commercial third place…a place people go to socialize that isn’t home or the workplace. But like Starbucks and McDonald’s, Walmart also functions as a replacement home for some people. Across America, Walmart parking lots fill up with the vans, RVs, and cars of nomads, vacationers, and the homeless. The NY Times sent a pair of photographers out to capture some of these parking lots at night.

There are standards of etiquette — do not, for instance, sit in the parking lot in lawn chairs — and also online rosters of no-go Walmarts. There is an expectation that you should buy something, but there is no parking fee. There is a measure of solitary privacy, even in a place that is deliberately accessible. Still that doesn’t prevent some people from leaving skid marks in the parking lot.

El Monte RV provides a short guide to Walmart camping and Allstays has a list of Walmarts that allow overnight parking.

SUPPORT LUNARBABOON!!!

Texts From Superheroes

Facebook | Twitter | Patreon