Is it time for a new Firefox thread yet?

ratchet_freak

@HardwareGeek said:
Doing something useful when you don't have enough memory to do it is the hard part. I'd still like to know how they think Rust sill solve this.

I suspect you guys are reading too much into some blogger's misunderstanding of what Rust is supposed to do. From the rust home page:

@rust-lang said:
Memory safety: no null or dangling pointers, no buffer overflows

Memory model: optional task-local GC, safe pointer types with region analysis

I don't see anything there about helping you deal with low memory stuff like everyone is fixated on. These features are meant to be a step up from dealing with memory in C++, and based on the descriptions, I believe it. Seems like most new languages have this sort of thing.

or they are just more focused on fixing memory leaks,

there are other languages that can do that

boomzilla

@ratchet freak said:

r they are just more focused on fixing memory leaks,
there are other languages that can do that

Yes, their thing for memory leaks strikes me as the main reason behind this. It does seem a bit crazy to use something brand new, but then what other language would be a reasonable replacement?

@HardwareGeek said:

@El_Heffe said:
... syntax like let x = 3 has been around for a long time. I seem to remember it being in some languages such as BASIC many years ago...
It was definitely in BASIC in the 70's. I thought it had died a well-deserved death almost that long ago.

Filed under: Apparently I was wrong.

Everything old is new again.

@too_many_usernames said:

This is why some people should not be allowed to create computer languages. In my industry, "let x = 3" would be interpreted as "3 is an allowable value for x" and wouldn't say anything about the current value of x.

If you want to set a value, the language should be something like set x = 3 which is astonishingly less ambiguous.

Put another way: anyone that thinks "let" is a synonym for "define" has serious linguistic issues.

Set doesn't work terribly well with immutable variables when you allow variable shadowing, because you're not changing the current value of anything; you're defining a new label with a different value to be used instead of the old value.
Consider:

let x = 1
while x < 10
  let x = x + 1
  print x
print x

Inside the loop, you're not setting the value of x, you're defining a new x to be used in the current scope. When you leave that scope (each iteration, or when you leave a loop), that definition is removed and the previous x becomes available again.
The above is equivalent to:

let x1 = 1
while x1 < 10
  let x2 = x1 + 1
  print x2
print x1

Or more simply:

while 1 < 10
  print (1 + 1)
print 1

Ben L.

@Salamander said:

@too_many_usernames said:
This is why some people should not be allowed to create computer languages. In my industry, "let x = 3" would be interpreted as "3 is an allowable value for x" and wouldn't say anything about the current value of x.

If you want to set a value, the language should be something like set x = 3 which is astonishingly less ambiguous.

Put another way: anyone that thinks "let" is a synonym for "define" has serious linguistic issues.

Set doesn't work terribly well with immutable variables when you allow variable shadowing, because you're not changing the current value of anything; you're defining a new label with a different value to be used instead of the old value.
Consider:
let x = 1
while x < 10
  let x = x + 1
  print x
print x
Inside the loop, you're not setting the value of x, you're defining a new x to be used in the current scope. When you leave that scope (each iteration, or when you leave a loop), that definition is removed and the previous x becomes available again.
The above is equivalent to:
let x1 = 1
while x1 < 10
  let x2 = x1 + 1
  print x2
print x1
Or more simply:
while 1 < 10
  print (1 + 1)
print 1

That problem also exists if you can declare multiple variables in one statement and you have shadowing. Like so:

x, err := DoSomething()
if err == nil && SomeCondition(x) {
    y, err := DoSomethingElse(x)
    if err == nil {
        x, err = DoAThirdThing(y)
    }
}
return x, err

The value of err is always what DoSomething returned because the line that declares y also declares a new variable named err which disappears at the end of the if statement.

@Ben L. said:

x, err := DoSomething()
if err == nil && SomeCondition(x) {
    y, err := DoSomethingElse(x)
    if err == nil {
        x, err = DoAThirdThing(y)
    }
}
return x, err
The value of err is always what DoSomething returned because the line that declares y also declares a new variable named err which disappears at the end of the if statement.

Congratulations on giving an example that is actually quite plausible.

dkf

@Ben L. said:

shadowing
[decent example]

And that's why shadowing is problematic. Yes, it might be a good idea in some circumstances, but more often indicates a problem. I prefer to be without it (since if I've got name overlap problems, my function is too long anyway and should be refactored).

Ben L.

@dkf said:

@Ben L. said:
shadowing
[decent example]
And that's why shadowing is problematic. Yes, it might be a good idea in some circumstances, but more often indicates a problem. I prefer to be without it (since if I've got name overlap problems, my function is too long anyway and should be refactored).

Generally, shadowing is useful when the variables are in different scopes, say a global variable and a local variable that just happen to be named the same thing.

@dkf said:

And that's why shadowing is problematic. Yes, it might be a good idea in some circumstances, but more often indicates a problem. I prefer to be without it (since if I've got name overlap problems, my function is too long anyway and should be refactored).

As far as I can tell, the main reason Rust has variable shadowing is so they can do stuff like:
let foo = get_some_object();
let foo = foo.get_some_attribute();
Which is the worst reason ever to allow it.

dkf

@Ben L. said:

Generally, shadowing is useful when the variables are in different scopes, say a global variable and a local variable that just happen to be named the same thing.

“Just happened to be named the same thing” is my point. If that's the problem, there's something major wrong at the language design level (and the code itself is probably a ticking time bomb). If you must have global variables in a language (and there are some good reasons for doing this) then at least they should require some explicit syntax to say “I want the global variable now, please” so that your code doesn't get broken by something you're not aware of.

And shadowing ought to generate some sort of warning at least, a “hey, you are aware that this is confusing as fuck?”

Wait . . . what? A global variable and a local variable with the same name? Even if it works and your program doesn't assplode, isn't that a really bad idea?

veggen

TRWTF are "developers" who find the concept of immutable variables new and confusing. Seriously people, a gazillion languages have everything immutable (all the functional ones, to begin with) and the benefit is inherent thread safety, and thus - easy concurrency. I'm not sure who here just enjoys playing dumb and shitting on everyone and everything and who is actually dumb, but both groups should pull heir shit together and stop the lame rant.

too_many_usernames

I'm trying to figure out what scope rules have to do with using 'set' or perhaps 'define' instead of 'let'.

Also - calling different storage locations at different scopes with the same name "shadowing" is a confusing use of the word "shadowing" . . . if you work in an industry with concepts like "shadow RAM".

immutable variables

I'll jump on the "Yet another instance of why computer programmers should pay attention in linguistics class" bandwagon for this one. If it's immutable it's a constant, not a variable. And just because functional languages implicitly create variables instead of explicitly, doesn't mean there aren't variables there.

flabdablet

@too_many_usernames said:

I'll jump on the "Yet another instance of why computer programmers should pay attention in linguistics class" bandwagon for this one.

What a lot of negative energy.

Lorne Kates

@too_many_usernames said:

"Yet another instance of why computer programmers should pay attention in linguistics class"

Referrrrrrrrrer.

Ben L.

@Lorne Kates said:

@too_many_usernames said:
"Yet another instance of why computer programmers should pay attention in linguistics class"

Referrrrrrrrrrrrrrrrrrrrer.

FTFY

Seriously, correct spelling is important.

@too_many_usernames said:

I'm trying to figure out what scope rules have to do with using 'set' or perhaps 'define' instead of 'let'.

It's not as important as what is happening; If you have prisoner245, and you set them free that is an action on prisoner245.
On the other hand, if you define prisoner245 to be a free person, you have two prisoner245's; one who is free, and one who is not.
Scope rules come into play as to what happens when you leave them; a definition cannot exist outside it's scope, but the act of setting prisoner245 free will still exist even if the scope in which it was changed does not.

@too_many_usernames said:

Also - calling different storage locations at different scopes with the same name "shadowing" is a confusing use of the word "shadowing" . . . if you work in an industry with concepts like "shadow RAM".

How? One is referring to a hardware concept while the other is a software concept. And they are both based on the definition of "shadowing"; to follow closely and secretly.

@too_many_usernames said:

immutable variables

I'll jump on the "Yet another instance of why computer programmers should pay attention in linguistics class" bandwagon for this one. If it's immutable it's a constant, not a variable.

final StringBuilder b = new StringBuilder(); // Is an immutable variable
final MyEnum e = MyEnum.Foo; // Is a constant
Object Xyz(final MyEnum e) { ... } // e is a immutable variable of function Xyz. It is constant within Xyz.

Enterprise_Architect

@boomzilla said:

Yes, their thing for memory leaks strikes me as the main reason behind this. It does seem a bit crazy to use something brand new, but then what other language would be a reasonable replacement?

Maybe not reasonable, but how about C++? While it certainly lets you do stupid things with memory, it also gives you plenty of ways to not do that. Plus, the same tools can help you manage other, possibly more constrained resources in a predictable way.

But maybe I'm biased. My day job has me dealing with software that runs out of file descriptors long before it runs out of memory.

@Enterprise Architect said:

@boomzilla said:
Yes, their thing for memory leaks strikes me as the main reason behind this. It does seem a bit crazy to use something brand new, but then what other language would be a reasonable replacement?
Maybe not reasonable, but how about C++? While it certainly lets you do stupid things with memory, it also gives you plenty of ways to not do that. Plus, the same tools can help you manage other, possibly more constrained resources in a predictable way.

Firefox is written in C++

Enterprise_Architect

@El_Heffe said:

@Enterprise Architect said:
@boomzilla said:
Yes, their thing for memory leaks strikes me as the main reason behind this. It does seem a bit crazy to use something brand new, but then what other language would be a reasonable replacement?
Maybe not reasonable, but how about C++? While it certainly lets you do stupid things with memory, it also gives you plenty of ways to not do that. Plus, the same tools can help you manage other, possibly more constrained resources in a predictable way.
Firefox is written in C++

Yeah, I know. My point is it's not necessary to write an entire new language to avoid leaking memory.

Though in their defense, what better way is there to make sure a new programming language is actually useful than to write large real-world applications in it?

topspin

@El_Heffe said:

But that's OK. At least Firefox 28 now has volume controls for HTML5 video and audio.

Well, that is indeed something.

I think just a few months ago, Chrome added an indicator to tabs playing sound or using the camera. And I was thinking oh great, while I (and several others) submitted feature suggestions back in fucking 2007 for, what, Firefox 3.6??, that it should have volume controls / indicators / mute button / something like that for every tab. Because hunting down sound bluring from a random tab is annoying.

topspin

@Lorne Kates said:

AND THEY CALLED THEIR LANGUAGE RUST! Why not "obsolete"? Or "decayed corpse pile"? AH! AH! AHHHHHHHHHH *poof*

At least the features it has sound interesting, especially when you consider it compiles to run on bare metal.
I mean, they didn't build it on top of the worst language ever and then call it HACK.

@Ben L. said:

But that's not possible. Because Mozilla refuses to tell me what version their browser is. @fire2k said:
@El_Heffe said:
@dkf said:
their policy is that you're not really supposed to care what the product version of the browser is,
Because they live in a bizarro fantasy world where new versions don't break anything or cause any incompatibilities. Which is especially hilarious in the case of Firefox, who now goes out of their way to make sure every new version breaks everything.
The current beta contains their new UI which really goes out of their way and fully breaks everything forever

QFT

The problem with Australisaurus is not the shitty UI. The default UI for Firefox has been shitty since version 4.0, and barely tolerable before that. But, it was easily changed. A couple clicks here. A couple clicks there. Install a theme. Drag the buttons around, and in approximately 2 minutes and 13 seconds you've got a nice browser.

Australapithicus, however, breaks things in ways that are difficult or impossible to fix. The 'Classic Theme Restorer' extension tries to undo some of the damage but has many limitations. For example, you can only use it with the default theme. Since the default theme is shit, this greatly reduces your ability to restore Firefox to a usable state. More examples, here.

The excuse "Most people don't use those features" is meaningless bullshit. Removing features, especially features that have existed for several years, benefits no one. If most people just use Firefox with the default shitty UI and never do any customization, that's fine. Maybe they like it that way. Maybe they are too lazy or stupid to even think of changing it. Whatever.

But the ability to make all those changes and customizations should remain in place. Having all those options available has absolutely no effect on the lazy, stupid people because they never see them or go looking for them, but they still exist for the people who want them. Everyone wins, everyone is happy.

About a month ago I switched to a Firefox fork called 'Palemoon' that someone on here mentioned. It's maintained by some guy in Sweden and he's patching in the latest bug/security fixes while still keeping most of the features that have been removed from the UI over the last couple of years.

blakeyrat

@El_Heffe said:

Removing features, especially features that have existed for several years, benefits no one.

On the contrary, it benefits everybody: fewer features leads to faster, more complete QA and fewer shipped bugs.

Of course the catch here is that Mozilla doesn't bother QAing the product either, so.

Ben L.

@blakeyrat said:

@El_Heffe said:
Removing features, especially features that have existed for several years, benefits no one.

On the contrary, it benefits everybody: fewer features leads to faster, more complete QA and fewer shipped bugs.

Of course the catch here is that Mozilla doesn't bother QAing the product either, so.

removed rarely-used render-text-in-a-legible-font feature to improve page load speeds

@blakeyrat said:

@El_Heffe said:
Removing features, especially features that have existed for several years, benefits no one.

On the contrary, it benefits everybody: fewer features leads to faster, more complete QA and fewer shipped bugs.

I stand corrected. Removing features benefits no one, except the laziest, shittiest developers who don't give two fucks about the people who use their software. Congratulations. Have a cookie. You should work at Mozilla, it sounds like you would fit right in there.

LoremIpsum

As these threads get longer and more frequent, I'm beginning to worry about the long-term future of Mozilla too, and not just Australis. I've just been checking my add-ons list and checking if they're available on Chrome, and it looks like all the mission-critical ones are. I still can't help but think that moving to Google is an even worse idea, though. Oh well...

anachostic

@blakeyrat said:

@El_Heffe said:
Removing features, especially features that have existed for several years, benefits no one.

On the contrary, it benefits everybody: fewer features leads to faster, more complete QA and fewer shipped bugs.

Please explain that benefit to Opera users who got fucked by a Chromium dildo. I'm sure they'll understand.

Onyx

@anachostic said:

@blakeyrat said:
@El_Heffe said:
Removing features, especially features that have existed for several years, benefits no one.

On the contrary, it benefits everybody: fewer features leads to faster, more complete QA and fewer shipped bugs.

Please explain that benefit to Opera users who got fucked by a Chromium dildo. I'm sure they'll understand.

Opera: Ok guys, so, we're getting rid of Presto* and switching over to Webkit
Well, I guess, if you can't maintain it and keep it current with standards I guess that's fine...

Opera: Also, we're not gonna fuck up speed dial which every other browser ripped off and called it live tiles or whatever and did it poorly
Oh, ok, much appreciated, well, I guess you're gonna keep it as close as possible to old ver...

Opera: Oh btw, Dragonfly? Yeah, gone, go download Firebug or something.
But... well, granted, it probably depended on Presto so...

Opera: We're also mail client, news client and chat...
Ummm... Not that I used them myself but could you...

Opera:... and mouse gestures, and bookmarks...
Now wait just a damn minute...

Opera: Oh, and also, Qt has to go, I mean, not like it works on any platform imaginable and has great performance, noooo, GTK is a much better choice!
...

Opera: And in line with that, we're only interested in the Windows version. You dirty alternative OS people might get something one day... if we can be bothered.
$#%#^%$

No, I'm not bitter, why do you ask?

* Opera's rendering engine up to version 12

blakeyrat

This situation is obviously fictional, because there is no such thing as an "Opera user". But let's assume for the moment that they exist...

@Onyx said:

Opera: Oh, and also, Qt has to go, I mean, not like it works on any platform imaginable and has great performance, noooo, GTK is a much better choice!

So they switched to GTK...

@Onyx said:

Opera: And in line with that, we're only interested in the Windows version.

BECAUSE THEY THINK GTK WORKS BETTER IN WINDOWS!?!???!? Holy shit. No wonder there's no such thing as an Opera user.

topspin

You seem to be implying that GTK sucks donkey balls on Windows only. It is way inferior to QT everywhere.

dkf

@topspin said:

You seem to be implying that GTK sucks donkey balls on Windows only. It is way inferior to QT everywhere.

I've never used Qt, but I did use GTK for a short contract (because the customer wanted it that way) back in 2002. It sucked then, but maybe it is better now?

LoremIpsum

@dkf said:

@topspin said:
You seem to be implying that GTK sucks donkey balls on Windows only. It is way inferior to QT everywhere.
I've never used Qt, but I did use GTK for a short contract (because the customer wanted it that way) back in 2002. It sucked then, but maybe it is better now?

I'm interested in knowing more about what in GTK sucks so much. I have been using Pidgin on Windows for 8 years and it seems to work just fine. It has a few peculiarities but I don't know whether those are down to GTK sucking or just Pidgin not working very hard on their supplemental dialog boxes.

blakeyrat

@LoremIpsumDolorSitAmet said:

I have been using Pidgin on Windows for 8 years and it seems to work just fine.

Then you are forever disqualified from ever again judging the quality of software.

boomzilla

@blakeyrat said:

@LoremIpsumDolorSitAmet said:
I have been using Pidgin on Windows for 8 years and it seems to work just fine.

Then you are forever disqualified from ever again judging the quality of software.

I've never used Pidgin. What's wrong with it on Windows? Is this just another "non native widgets" thing? Or is there actually something wrong?

Onyx

@boomzilla said:

I've never used Pidgin. What's wrong with it on Windows? Is this just another "non native widgets" thing? Or is there actually something wrong?

Non-native widgets, mostly. Also, it pulls the entire Gtk toolkit with it instead of just including whatever DLLs it needs to run. I have no idea how and where it installs it, and if other Gtk applications can use it, but given the amount of Gtk applications available for Windows that attempt to use a shared Gtk install (from my personal experience only Pidgin and Gimp do that), and amount of Windows applications that use Gtk at all, it seems like a bit of an overkill.

My point was that it's not that it works badly, it's that they switched from Qt, which will bend over backwards so you don't have to worry about the platform to get things looking mostly right (OS conventions not withstanding), and hell, includes QtWebKit in it's essential modules, to Gtk which is pretty much "Linux first, port it to other platforms later", and then focused mostly on a Windows release. Kinda boggles the mind.

Note: I don't know how up to date QtWebKit is. I'm gonna guess it's lagging behind the main release, but it should be possible to recompile it with code from WebKit trunk anyway.

boomzilla

@Onyx said:

My point was that it's not that it works badly, it's that they switched from Qt,

Yeah, I get that. I was just wondering if blakey's problem was his native widget OCD thing, or something that actually sucks. I totally agree that Qt is better than GTK, but that doesn't mean that decent programs (as judged by sane humans) can't be developed using GTK.

@Onyx said:

Note: I don't know how up to date QtWebKit is. I'm gonna guess it's lagging behind the main release, but it should be possible to recompile it with code from WebKit trunk anyway.

I don't know either, but I'd be surprised if you still could.

Lorne Kates

@boomzilla said:

I was just wondering if blakey's problem was his native widget OCD thing

So let me get this straight. Just because there's a consistent, tested and expected set of widgets already available, you think that's a REASON to "just keep using them"? Luddite.

boomzilla

@Lorne Kates said:

@boomzilla said:
I was just wondering if blakey's problem was his native widget OCD thing

So let me get this straight. Just because there's a consistent, tested and expected set of widgets already available, you think that's a REASON to "just keep using them"? Luddite.

Dramatization of blakeyrat encountering a GTK widget in Windows. Of course, it's all good fun when MS confuses us all with Win8.

dkf

@LoremIpsumDolorSitAmet said:

I'm interested in knowing more about what in GTK sucks so much.

It just took so much work to make anything work, and even more work to make it work well. It felt like hammering nails made of potato chips into granite using a giant multi-colored squeaky hammer.

boomzilla

@dkf said:

@LoremIpsumDolorSitAmet said:
I'm interested in knowing more about what in GTK sucks so much.

It just took so much work to make anything work, and even more work to make it work well. It felt like hammering nails made of potato chips into granite using a giant multi-colored squeaky hammer.

I've only ever built anything with it via wxWidgets, so I've never had to hammer the chips myself.

blakeyrat

@Onyx said:

Non-native widgets, mostly.

It's not that they're non-native, it's that they're non-native and they suck shit. Firefox uses non-native widgets, but it (for the most part) gets them correct and so there no complaint. But GTK widgets ignore the user's colors/theme, ignore the OS colors/theme, don't work with speech recognition, don't work with touchscreen/tablet devices, etc.

The GTK dialogs are also hilariously awful. You can't just dismiss an atrocity like the GIMP open dialog with, "oh it works a little different than Windows," it's more like, "oh it works a little different than ANY PREVIOUS COMPUTER EXPERIENCE WEAK HUMAN FLESH-CREATURES HAVE HAD IN THE PAST MY ROBOT BROTHER!" Oh and another the 56,000 other things it doesn't support in Windows, it's also oblivious to the fact that NTFS is not case-sensitive and thus even something as simple as alphabetizing the list of files is completely broken.

boomzilla

@blakeyrat said:

The GTK dialogs are also hilariously awful. You can't just dismiss an atrocity like the GIMP open dialog with, "oh it works a little different than Windows," it's more like, "oh it works a little different than ANY PREVIOUS COMPUTER EXPERIENCE WEAK HUMAN FLESH-CREATURES HAVE HAD IN THE PAST MY ROBOT BROTHER!" Oh and another the 56,000 other things it doesn't support in Windows, it's also oblivious to the fact that NTFS is not case-sensitive and thus even something as simple as alphabetizing the list of files is completely broken.

That's...weird. I've never seen that behavior before. I just tried some GTK apps (on Linux, not OSX or Windows) and it does case insensitive alphabetizing. A little googling tells me this seems to be some sort of localization issue on Macs when the locale is anything other than C. Mine is en_US, BTW. Does the same problem show up on Windows?

flabdablet

@El_Heffe said:

About a month ago I switched to a Firefox fork called 'Palemoon' that someone on here mentioned. It's maintained by some guy in Sweden and he's patching in the latest bug/security fixes while still keeping most of the features that have been removed from the UI over the last couple of years.

I'm considering cutting the school over to Palemoon to avoid the howls of outrage from my user population once I'm forced to move off Firefox 26ESR, but I'd want to be assured it had some kind of future first. If I understand Mozilla's plans correctly, they're removing a whole bunch of stuff from the underlying UI toolkit that's of no futher use to Asstralis. Seems to me that at some point Palemoon would have to diverge from the Firefox codebase enough that keeping the fixes up to date would overwhelm some guy in Sweden.

ender

@blakeyrat said:

Oh and another the 56,000 other things it doesn't support in Windows, it's also oblivious to the fact that NTFS is not case-sensitive and thus even something as simple as alphabetizing the list of files is completely broken.

Err, what? That's not a GTK+ on Windows screenshot - GTK+ on Windows does not have this problem (and it also follows the Windows theme better than many non-GTK+ Windows programs):
GTK+ file open dialog demonstrating that Blakey has no idea what he's talking about again

GTK+ file open dialog demonstrating that Blakey has no idea what he's talking about again

blakeyrat

@ender said:

Err, what? That's not a GTK+ on Windows screenshot

Lies. It's from Inkscape, which is a GTK program. I don't know what your + indicates, but as it happens I also do not give a shit.

That said, it's not a *current* GTK screenshot, it's a couple of years old. But, guess what? It was a pretty unacceptable bug back then as well.

@ender said:

(and it also follows the Windows theme better than many non-GTK+ Windows programs):

Then why are all your window and button captions in some weird-ass martian language instead of English? Answer me that, smartboy!

ender

@blakeyrat said:

That said, it's not a current GTK screenshot, it's a couple of years old.

Inkscape is a GTK+ program, but that screenshot is pretty obviously showing it running in X11 on OS X, not Windows (unless maybe you meant X-Windows? Nah.).

blakeyrat

@ender said:

Inkscape is a GTK+ program, but that screenshot is pretty obviously showing it running in X11 on OS X, not Windows (unless maybe you meant X-Windows? Nah.).

Oh crap, you're right. That was on my old OS X box, wow that is pretty old.

Either way, WTF still applies. HFS+ isn't case-sensitive either.

boomzilla

@blakeyrat said:

@ender said:
Inkscape is a GTK+ program, but that screenshot is pretty obviously showing it running in X11 on OS X, not Windows (unless maybe you meant X-Windows? Nah.).

Oh crap, you're right. That was on my old OS X box, wow that is pretty old.

Either way, WTF still applies. HFS+ isn't case-sensitive either.

It's not a disk thing. It's a macports thing and locales.

Onyx

@blakeyrat said:

Oh and another the 56,000 other things it doesn't support in Windows, it's also oblivious to the fact that NTFS is not case-sensitive and thus even something as simple as alphabetizing the list of files is completely broken.

The only program I saw that does that is Filezilla, but it does that on Linux as well. At first I though it's something to do with FTP itself but it does it for the local file list too. Never been bothered enough to care since I usually just use scp or rsync if I have shell access.

Last time I used Pidgin on Windows it was a far cry from that Gimp screenshot, it just loaded the Tango theme completely ignoring everything else but window decorations. Then again, it was ages ago so I have no idea if it got any better. I never much minded myself, though I did find it funny it saved it's settings in C:\user\whatever-windows-decided-is-settings-dir-this-version\.purple. Nice "hidden" folder there guys!

Also, isn't it that NTFS is case sensitive but explorer refuses to admit it? I'm pretty sure it quite happily saved both "a.txt" and "A.txt" as two different files when accessed from Linux, but threw a hissy fit when I tried to rename those files in explorer.