Uhhh... that was the OP's point:

No matter how good it is, no torrent client will allow you to choose the files inside a zip file inside the torrent.

Wish I had mod points again so I could pull you above 0.

The decisions and direction of Ubuntu are disturbing when considering longer-term usage. LTS releases last for 5 years, but will you really still want to be using Ubuntu 5 years from now?

I just installed 12.04 Server and they have added an advertisement to the MOTD for their paid Landscape service, an ad which is displayed at every logon. Not only that, but the new "better" dynamic MOTD system is very opaque and not easy to see how you can customize or get rid of that advertisement. (For anyone curious, the best solution I found was to simply remove executable rights from the scripts I didn't want to run in /etc/update-motd.d. I only left 00-header and 99-footer).

Ubuntu seems all about trying to "upgrade" the standard GNU/Linux experience, from Upstart replacing Init and UUID-encoded mount devices to a horrible mess of network configuration scripts replacing /etc/interfaces and the early use of GRUB 3. I realize that there are probably improvements to be found in some of these changes, but there are also a LOT of new bugs and idiosyncrasies that everyone has to deal with now because they jumped the gun and changed for things were ready and stable.

For new servers, Debian stable has replaced Ubunut LTS as my go-to distro. It's been a welcome return to sanity.

Bad form to reply to oneself, but I forgot my favorite:

More in the stop dicking around with the UI category, we have an idiot named Carlo Alberto Ferraris to thank for destroying standalone image viewing (when you open an image file directly, or via right-click and View Image). He was so very offended by that page having a white background that he felt it necessary to ruin a feature that's been standard in browsers for over a decade.

This isn't just an issue of changing something for the sake of change, it's a plain stupid idea in the first place. First, a dark background when most websites and images are very light is jarring. Second, centering the image makes it harder to click on for actions like saving or copying. And third, it destroys the usability of a very common entire class of images.

Open this image of waveforms in Firefox 13+ to see the problem. Transparent GIFs have the same issue. The solution? Yet another addon to fix stupid Firefox developer mistakes.

This hasn't been true since version 12. In that release they added an auto-updating service that runs in the background and handles installing updates without the user's permission.

I had Firefox 13 set to "Check for updates, but let me choose when to install" when version 14 was released. It bugged me a couple of times to install 14 and I said no each time. Then, one morning I open Firefox to see 14 has been installed, completely without permission. I checked the update settings and it was still set to "ask me". Looking at the update log showed that 14.0.1 had been installed as a "security update".

Few things piss me off more than software doing things I've explicitly told it not to. Firefox auto-updating wouldn't be so bad if the moronic development leads would:

• Stop dicking around with the user interface. At this point they're only changing it just to change it because they can (new toolbar buttons are just white on Aero glass? What kind of idiots are running this show again?).
• Stop dicking around with the basics. In 14 they changed the mouse-wheel scroll timing because some dev retard though it should be "smoother". The "scroll time" was doubled, making mouse scrolling like walking through freezing molasses. Thank frak there's an about:config setting for it (general.smoothScroll.mouseWheel.durationMaxMS should be ~200), or I'd have ditched Firefox same-day.
• Stop changing performance settings to satisfy memory "leak" morons. Just because a web browser is using 1GB of memory (on your 8GB system) doesn't mean it has a memory leak. It means that web pages are filled with images, and decoded images are big. Throwing away that memory every time you switch tabs means that all those images have to be re-decoded when you switch back. But what the hell -- now they can claim "OMG, Firefox 13 uses less memories than Chrome!!11!". Stupid.

Just a few of the things I hate about the new Firefox system. The ONLY reason to stick with Firefox is the addons. Mozilla is betting the entire farm against the Firefox addon ecosystem -- if (or rather, as soon as) Chrome catches up, people (including a lot of "power" users) will start leaving in droves.

Firefox has been taken over by the same kinds of people that have poisoned GNOME for years. They think dictating to users what they do and do not like and what they will and will not do is the correct way to design software. They are dead wrong, something the failure of GNOME 3 should have taught them, but just hasn't managed to sink in yet (if it ever does).

Yeah, it's pretty lame the way they handled this. It's very clearly a move to push people at Windows 8 by removing "value-add" stuff from Win7 and Vista.

If you're looking for gadgets, most of the old ones (and a bunch of new ones) have been hosted here:

http://gallery-live.com/

It's nice to see someone else who doesn't see showing file extensions as some kind of panacea for computer issues. It really wouldn't help at all.

I like this idea, but it seems like an unsolveable problem: How do you decide what is "executable" and what isn't? Sure, there are well-defined executable binaries for each operating system, but there are also a myriad of other ways you can run code. Whether it's a Bash script, VB script, embedded macro of some kind, etc. In Windows, the shell (Explorer) runs ShellExecute on the file to invoke the filetype's registered handler -- how would it know ahead of time what that handler might do? I could display an image or it could load embedded bytecode and run it through an interpreter.

And, on Windows at least, this "glow" already exists to some extent. Files downloaded through most (all?) browsers these days are given the "mark of the Internet", and alternate data stream tag that says "this came from the web". When a user runs a program (or certain other file types), a dialog box is displayed that says "Warning! This is an untrusted program that may break your computer!". Users simply ignore this -- will a glow be much different?

I think I have to disagree with you on both counts here. First, because even if systems had been designed that way at the beginning, they would have evolved to what we have today. The reason is simplicity: We simply want the computer to "open" whatever it is we're interested. I don't want to be bothered every time, trying to decide if the file I'm looking at falls under "document" or "program". Say, for example, an HTML file. It's a "document", sure, but it also may contain various forms of scripting which falls (or should) under "program". Computers are supposed to take care of easy repeated tasks, surely knowing how to display a given type of file falls under that.

Second, and partially because of the first, I think users would very quickly train themselves to start right-clicking anything that didn't work when they left-clicked it. Download a good program and left-click it by habit -- oops, didn't work, hmm, try right clicking -- it works. Now they download an evil cat picture and left-click it -- oops, that didn't work, hmm, try right-clicking -- it works (and they're now infected).

Humans are very good at making deductive steps towards solving a problem -- "I tried left-click and that didn't work, but I know that right-clicking also does things to files I download, I will try right-clicking. It worked, therefore right-clicking is the answer when left-clicking doesn't work. Stupid computers, why are they so hard to use?" -- and then the whole system is void. If you observe users trying to figure something out, you quickly see they come up with all sorts of odd and unusual ways of going about it, just because "that's what works".

Fundamentally it comes down to understanding the separation of the two kinds of files and why it's important to treat them differently. This requires technically informed users -- the very same flaw as simply displaying file extensions.

That's a nice ideal, but the reality is that many up-to-date "stable" distribution releases are still using kernels which are susceptible the leap second problem (and haven't had the patch back-ported to them). Ubuntu 8.04 LTS server is supposed to be supported until April 2013, and on my (updated!) system,

# uname -r
2.6.24-28-server

I like the idea of stable releases, but this is a glaring problem with the entire idea. Everyone extolls the wondrous virtues of package managers for Linux-based systems, but the dirty secret is that unless you stay bleeding-edge (which is usually the opposite of "server"), you'd better be happy with the 4-year old version of Apache, PHP, MySQL, and the Linux kernel you're running. Sure, it's possible to manually download and install packages from a newer release (assuming you can get past the dependency hell usually associated with it). Sure, it's possible to try and splice in (or "pin" packages using Debian parlance) from a newer repository. Sure, it's possible to install from source, compiling and installing everything by hand. But once you do any of these you've given up 90% of what makes the package manager useful and are just asking for dependency problems in the future.

And, all that aside, do you even know if the patch released to fix this problem is included in your distribution-released kernel? If you're not rolling your own kernel it can be nigh to impossible to know what's included and what's not -- in that case it doesn't even matter if it's up-to-date.

I suppose that's fitting, given that Do Not Track is absolutely nothing more than a strongly^W weakly worded letter anyway.

Everything surrounding the feature is a complete joke, from Mozilla introducing it in the first place to people who think it will really make even a small difference. Do Not Track is absolutely the same as walking around a bad neighborhood with a Do Not Mug sign hanging around your neck. It will not work.

For everyone babbling about "if everyone does it then advertisers won't respect it!", stop and think about what you're saying. If it can't work for everyone then it isn't really working for anyone. As far as I'm concerned, kudos goes to Microsoft for this move. Either they realize the stupidity of the header and are making a point or they are just blundering about and raising the issue by accident.

We already have an opt-in do-not-track feature that actually works -- it's called AdBlock, NoScript, etc, etc. Do Not Track is really nothing more than a PR stunt by those pushing it — not even worth the 6 bytes it takes to send it on each HTTP request.

I used to see the 10 second UI hang as well, probably because that's the interval at which Firefox saves session data (to restore in case of a crash). I'm not sure how bad it still is on newer versions, but I resolved it on 3.6 by installing Session Manager (to replace the built-in session state handling) and changing the session save interval from 10 seconds to 20 (in SM's options). As a bonus, Session Manager is really nice for saving and loading tab and window sessions for later use.

You might still see the UI freeze up during session saving if writing to your hard disk is being slow for some reason, though. It's one of those things I didn't understand why it wasn't done on a background thread.

I'm glad to have caught your eye with my comment! I've seen your name come up (well, I'm assuming it's you based on you Slashdot user) on several of the Bugzilla threads I found looking for answers to these issues. Since you seem to be familiar with it, I have a quick question regard image handling, if you're willing.

You mention image.mem.min_discard_timeout_ms. I've already set that one pretty high (1 hour (which really means, 30-90 minutes, right?)), but was wondering if it applies to closed tabs as well as background tabs. If it does, then that might be something to consider changing. I'd suggest that closed tabs can have their images discarded pretty quickly, but background tabs should keep them longer (honestly, I'd prefer to have a setting that prevents background tab images from ever being discarded).

Can you describe just briefly what image.mem.max_decoded_image_kb and image.mem.max_bytes_for_sync_decode control? I haven't had much luck finding documentation for these options, but have seen several people (yourself included) suggest modifying their values. Deciding how to adjust them is much easier if I had an idea of what they are doing.

One last thing -- I've seen mention that Firefox has a (possibly soft) cap set on the total number of bytes that decoded images can consume. Is that true, and/or related to image.mem.max_decoded_image_kb?

I really appreciate you letting me bend your ear over this issue. While I don't love many of the recent changes to Firefox, the way images are decoded and discarded has been one of the biggest annoyances by far and was the first problem I noticed when I finally went from 3.6 to 10. It's nice to see someone is working to improve the situation :)

If by fixed you mean browser usability was sacrificed in order to make the apparent memory usage drop, then yes. My biggest complaint with these memory "improvements" is in regard to image handling:

- Images are now decode-on-draw meaning they display slower and background tab images are not decoded. Browsing an image gallery or some other image-heavy site is now obscenely painful in Firefox.

- Decoded images on background tabs only live for 10-20 seconds and then are discarded at which point they must be re-decoded when the tab is activated. Long-lived tabs like Gmail now flicker every time you switch back to them as images are re-decoded.

These are just the two that come to mind right away. Luckily they can be fixed by tweaking some about:config settings (image.mem.decodeondraw and image.mem.min_discard_timeout_ms). Unfortunately many cannot be fixed so easily.

I'm really tired of the Firefox devs choosing (usually wrong) user complaints over good design and usability practices.

C:\Windows\System32\drivers\etc\hosts

Unfortunately that is also the arcane spell to summon APK -- get away while you still can!

Here's another one!

FCKGW-RHQQ2-YXRKT-8TG6W-2B7Q8

While on this topic, does anyone have any suggestions for good seedbox providers? I've come across a couple that look attractive, but they often blacklist public trackers (which largely voids the whole point of a seedbox).

If you've had a good or bad experience with a provider I'd love to hear it.