Ubuntu 14.04 LTS Trusty Tahr - The thing that thousands thought to try

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > Ubuntu 14.04 LTS Trusty Tahr - The thing that thousands thought to try

evol262: Nov 30, 2010; #!/usr/bin/perl

Stanley Pain posted:

Units sold/shipped is a bloody dubious metric to use at best. Subscriber + Market Share numbers paint a different picture. In the US it's about 50/50 Android/iOS and in the rest of the world it's 35/65 give or take ~5%.

Citation desperately needed if you're asserting that the iOS market share is anywhere near 65% (and not actually 10-20%) given that China and India each have a market three times larger than the US and both markets are >80% Android. iOS has a lot of adherents, certainly, but it is not anywhere near even market share, much less dominant.

# ¿ Aug 12, 2014 18:48

Adbot: ADBOT LOVES YOU

# ¿ May 21, 2024 09:44

evol262: Nov 30, 2010; #!/usr/bin/perl

computer parts posted:

Mobile traffic is different from actual device marketshare.

The discussion is about whether icons with a visual similarity to iOS icons are intuitive or not, not anything about mobile traffic, and mobile traffic has nothing to do with the number of people who actually own iOS devices and see those icons.

# ¿ Aug 12, 2014 19:22

evol262: Nov 30, 2010; #!/usr/bin/perl

computer parts posted:

I think mobile traffic is pretty important when discussing Safari.

It's actually completely unimportant when discussing the Safari icon, because that's what we're discussing, not Safari.

Whether or not iOS users account for a significant proportion of mobile traffic in the US isn't really relevant, since it indicates nothing more than heavy mobile use. Think of it this way, if 10% of people own iOS devices and account for 99% of traffic, 90% of Ubuntu users (on average) will still have zero familiarity with the icons used on iOS.

# ¿ Aug 12, 2014 19:56

evol262: Nov 30, 2010; #!/usr/bin/perl

Powered Descent posted:

In my experience this is the fault of Windows, not Linux. I once set up a Windows 8 machine to dual-boot to Ubuntu, and it worked great... until the next time I booted into Windows, which would helpfully (and silently) "fix" the EFI settings for me, removing the Ubuntu entry, so I'd have to dick around in the bios menus to be able to boot to Ubuntu again. Annoying as hell.

(Note: this was maybe six months ago; no idea if Microsoft has fixed that since then.)

As long as you use grub-efi or grub2, you don't really have to worry about this. BCD is harder to manage (a lot of system updates will "fix" it), but Ubuntu's installer should put grub2 at bootx64.efi, but an EFI shell can select it even if you overwrite it from Windows somehow.

# ¿ Aug 12, 2014 21:32

evol262: Nov 30, 2010; #!/usr/bin/perl

Stanley Pain posted:

Re-Read what I wrote

I did, and it reads the same way -- like unverified statements that need a citation if you're going to throw out absurd percentages. Maybe you didn't say what you intended to.

Subscribers mean nothing in countries where you use SIMs (Asia and large parts of Europe). Market share is units sold in most of the world.

# ¿ Aug 13, 2014 02:32

evol262: Nov 30, 2010; #!/usr/bin/perl

Stanley Pain posted:

Units sold does not mean what you think it means. Unless I'm not remembering the stats properly, units shipped = Units the OEM has sold, and not units bought at retail.

Yes and no. That's only a valid metric for the US, as are subscriber numbers. But units sold and units shipped are different numbers.

Stanley Pain posted:

The greatest metric for this is mobile web usage which has been at 50/50 Android/iOS for a while now. To put this to rest there are MANY more people than you think that will recognize a Safari like web browser icon. Certainly not 10%. Is it stupid they basically cloned a bunch of apple icons? Sure, I hate apple and don't want anything to even remind me of them

This is also only true for some countries. On average, iOS is between 10% and 30% of marketshare by mobile traffic, depending on the country, with English-speaking nations and some outliers (Brunei, Japan) very high on the iOS scale and others (Austria, Bangladesh, Indonesia, India) very high on the Android scale. You may notice a correlation between nations with very large populations and large internal infrastructure (where mobile traffic wouldn't show up on US servers) using Android.

Certainly in the English-speaking nations and wealthier countries, iOS has more market share (both in proportion and sometimes in absolutes), and if that's the market they're going for, it may be ok. But both on mobile traffic and device sales, iOS is pretty outstripped if they're looking to make "Phones for People" the same way they made "Linux for People". I don't think an Ubuntu Phone would succeed in any real way, especially with Sailfish already out there with real hardware on a real OS which also happens to be mostly Android compatible, but aping iOS is still pretty stupid, particularly given that "globe-ish thing" is a pretty universally identifable browser logo (IE, Firefox, Android Browser, debatably Chrome).

# ¿ Aug 13, 2014 18:29

evol262: Nov 30, 2010; #!/usr/bin/perl

ShadowHawk posted:

I've been getting occasional nagging-sounding email asking me to keep putting out new Wine betas for 12.04.

I'd rather not since it's old and extra work, but I'd really like to know what sort of use case desktop users might have for beta Wine releases but not the latest LTS.

As someone who still has to backport fixes to RHEL5 sometimes, welcome to long-term support.

Latest is good for us, but it's almost always users who have some application which won't work with a new version of python or oracle or their hardware is so old they can't get drivers for it on kernels past 3.2 or 2.6.28 or something. I'd almost guarantee the users requesting it are doing so because the WINE devs just fixed a bug in some game they think their 6800GT (or whatever) will run fine if only nvidia hadn't stopped supporting it in the binary driver years ago.

# ¿ Aug 18, 2014 00:19

evol262: Nov 30, 2010; #!/usr/bin/perl

ShadowHawk posted:

I'm pretty sure those users can use the old nvidia drivers on 14.04 though. We still ship nvidia-304 packages. (Default is -331)

I don't actually track nvidia's versions anymore, but my comment was a little hyperbolic rather than specific versions.

I'd be really surprised if there weren't cards that worked in 12.04 but not 14.04, though

evol262 fucked around with this message at 03:02 on Aug 18, 2014

# ¿ Aug 18, 2014 03:00

evol262: Nov 30, 2010; #!/usr/bin/perl

keyvin posted:

OK, I have googled and I can't figure this out. I have a chromebook with xubuntu 14.04 on it. It has a tiny SSD, and I would like to keep some gog.com games on external storage. I was getting ready to write a script that listened to dbus for the message that removable media has been attached and mounted in /media. It would then scan for a folder and create a symlink to it in the .wine/drive_c folder.

This seems overly complicated. Given ubuntu's reputation for user friendliness there should be an easy way to do this. Maybe without the symlink but directly mounting the storage as .wine/drive_c in fstab, using my user as the user and the UUID as the device? I know jack poo poo about udev/thunar automounting.

Does it always get mounted at the same place (probably)? If so, just make a symlink. If it's not mounted, your symlink will be broken, but that's not the end of the world, and it'll work fine again once the drive gets mounted again, unless I'm missing something.

# ¿ Sep 15, 2014 18:52

evol262: Nov 30, 2010; #!/usr/bin/perl

keyvin posted:

Yeah, that will work great for now. You are an awesome resource evol262. Do you have an amazon wish list so I can show a little material appreciation when I get paid again?

I appreciate the offer, but I don't really buy things, and helping people is its own reward anyway~

# ¿ Sep 16, 2014 14:15

evol262: Nov 30, 2010; #!/usr/bin/perl

Xenomorph posted:

With so many darned Linux, FreeBSD, and Windows systems to keep working, if I can Google what I'm trying to do, quickly find a solution (such as "heartbeat") and then get it working in just a few minutes with very minimal config (3 short files), I try to stick with it.

I will keep "keepalived" in mind. I also read about someone switching to "corosync" because of the heartbleed/pacemaker segfault issue.

All my VM tests with heartbeat have been running well. I'm going to push the setup to a bigger VM server and then have some others test the setup.

Heartbeat is deprecated and essentially dead. Use corosync for the future

# ¿ Oct 2, 2014 21:21

evol262: Nov 30, 2010; #!/usr/bin/perl

fourwood posted:

I would really doubt it's as good as OS X. I say this without having tried it, but... I just don't see it happening given Canonical's focus on not-the-desktop as of late.

It depends on desktop environment. KDE is best, Unity is OK, other DEs are frankly crap (though you can work around that by changing font sizes in tiling WMs). Handling high DPI screens and touchscreens properly is a major focus of GNOME and the successors to X.org.

It's pretty easy to get the OSX experience by just halving your resolution (which is essentially what OSX does), but high DPI is a better experience on Windows.

# ¿ Nov 5, 2014 16:35

evol262: Nov 30, 2010; #!/usr/bin/perl

Yvershek posted:

Any idea what it's using when I'm able to connect while trying from a disk? It's strange to me that installing onto a drive makes it stop working.

"lsusb -k" from the disk.

Then "dpkg -s /path/to/driver.ko"

Install that package, or list it here. The installer kernel should be the same as the installed kernel, but I'm not 100% sure of that.

# ¿ Dec 2, 2014 07:11

evol262: Nov 30, 2010; #!/usr/bin/perl

Xenomorph posted:

Well, my network has pretty newish Cat6 cables, newish Gigabit switch, and newish Gigabit router. The distances between the cable coming into the house, the TV tuner, and the computer is under 10 feet or so. I don't think the network is the issue.

Besides, the TV going out is an issue with the cable service. Like, from the coax straight to my "real" TV, so networking isn't causing the dropouts there. If my house had just a TV (and no computers or network), I'd still get a signal drop every now and then (thanks, Charter!).

Windows running VLC is using the network, and the TV dropouts cause it no problem. Just like my "real" TV, it just gets the blocky image and momentary pause in sound, then resumes like normal.

Linux running VLC is where the issue is. It drops audio and never "recovers". I have to re-start the stream to it (which usually involves me having to physically hit a kid on the keyboard), which has become such a pain that I've simply stopped using Linux at home. I've only been using it recently to watch TV, and that doesn't work right anymore.

I'd like to figure out what or where the problem is. Is VLC on Linux that much different than it is on Windows? Is the AC3 codec it uses that much different than on Windows? Shouldn't these codecs and streaming programs have some sort of error recovery? If its getting millions of packets streamed to it, why gently caress up completely and shut everything down when 1 packet has an error?

Is there another DLNA compatible media player out there that I could try? I've only used VLC on Windows, Linux, Android, and iOS... Which is another thing, VLC on iOS doesn't decode AC3, which is a huge pain. Other streaming apps have in-app purchases for AC3, but not VLC.

You can try using mplayer or something else. Are you transcoding to h.264? Use that instead of mpeg2.

I don't think vlc uses liba52, so you may need to open a bug against vlc about it, but I'd try mpeg4 with ac3 or aac first

# ¿ Feb 24, 2015 22:50

evol262: Nov 30, 2010; #!/usr/bin/perl

Xenomorph posted:

I use an HDHomeRun Prime (CableCard). It is MPEG2 + AC3, only, as that is what the cable company & over the air stations broadcast with.

Does mplayer let me manually open an http stream? Does it handle plain old MPEG2 & AC3?

UPnP has been hilariously broken in VLC on Windows in every build after 2.0.8 (even the 3.0.0 builds are still broken), so "turning on TV" consists of me just double-clicking a Playlist file that I've updated with a text editor to point to the stream.

If mplayer works I could easily just switch to that.

Is it proper DLNA with CDS and CMS and all that, or just a raw http stream?

# ¿ Feb 25, 2015 04:36

evol262: Nov 30, 2010; #!/usr/bin/perl

It's a WD device, so it's probably exporting both.

He said NFS, but it's likely that CIFS will work at the same path (and //server/export is CIFS anyway, not server:/export)

# ¿ Feb 28, 2016 05:33

Adbot: ADBOT LOVES YOU

# ¿ May 21, 2024 09:44

evol262: Nov 30, 2010; #!/usr/bin/perl

MrMoo posted:

I think RHEL minor versions have the same limitations if a lot of updates are required, they only support the last two I think?

Security updates go to EUS/z-stream as well as bug fixes to critical packages (there's a list), which is supported for two years. If it's really important, you can do AUS (6 years).

ABI/API is the same, though, so your kernel's never gonna break.

# ¿ Mar 8, 2016 20:10

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > Ubuntu 14.04 LTS Trusty Tahr - The thing that thousands thought to try