Why browsers are bloated

https://github.com/WebKit/webkit/blob/master/Source/WebCore/platform/Scrollbar.cpp
https://github.com/WebKit/webkit/blob/master/Source/WebCore/platform/win/ScrollbarThemeWin.cpp
Let's reinvent the fucking scrollbar, which every goddamn platform with a UI already has, and make it behave subtly different from the native one!

Right-click a native scrollbar in some other app:
- Scroll Here
- Top
- Bottom
- Page Up
- Page Down
- Scroll Up
- Scroll Down

Right-click a scrollbar in Chrome:
- Back
- Forward
- Reload
- Save As...
...

Right-click a scrollbar in Firefox and Opera:
Absolutely fucking nothing happens!

What the fuck!? How did these terminally retarded idiots get involved in creating one of the most important pieces of software to the average user?

>>720
But Scheme with

(define-syntax def
(syntax-rules ()
((def a ...)
(define a ...))))

(define-syntax lam
(syntax-rules ()
((lam a ...)
(lambda a ...)))))

is.

>>716
I remember everything I write. It's like a photographic memory for concepts.

What do you think will be the hardest part on the browser?
The rendering engine, I guess?
Could you make a release of what you have so far once the CSS layout is done?

>>723
The rendering engine isn't necessarily hard, just tedious with all the tiny bits of configurability that it needs to support.

What's hard is making sure it's secure, and trying to optimize it.

>>720
Clojure is a Lisp.

>>723
Don't count on it.
*Cudder is all talk and no action*

>>724

What's hard is making sure it's secure, and trying to optimize it.

*optimise
And no, it isn't.

>>727
Yes it is. Cudder's solution is incomplete as he decided to ignore the full range of the specification. Feel free to prove me wrong by writing a full spec, fully secure and fully optimized web renderer.

>>728
Feel free to prove yourself right.

>>723

What do you think will be the hardest part on the browser?

JavaScript

Could you make a release of what you have so far once the CSS layout is done?

Once Acid2 passes.

>>724
What do you mean exactly by "secure"? It seems everyone is throwing this word around like some kind of holy grail when it could mean anything from DRM and other user-hostile shit (fuck no, it's not going to be "secure" if you mean that) to buffer overflows (won't happen, there's very little dynamic allocation and what there is will have lengths checked very thoroughly...)

just tedious with all the tiny bits of configurability that it needs to support.

Calculating widths/heights is the most annoying part. One of the reasons why it's taking so long is because I'm trying to simplify it --- the spec is almost always more verbose and complex than it needs to be. Earlier in this thread I shrunk the HTML tokeniser from the absolutely bloated spec into a <1KB binary, so I expect some level of savings for the CSS parser and renderer... optimisation comes naturally from design.

>>730
Alrigiht bitch. Forget showing us the web browser. How about you just show us ANYTHING you have written. You see, we aren't even quite sure you know how to program.

>>731
RTFT, I've posted code and screenshots already.

>>732
40 lines of asm, lol

>>730

JavaScript

I don't think many people care about that.
Good part of the web works without it. Of course it'd be a nice addition to your browser, but I don't think it's essential.

Once Acid2 passes.

How far is it into passing Acid2?

Mr. Cudder, do you think your Asm/C performant browser will allow users who have an Intel DX4 to browse the modern web? What do you expect will be the requirements?

Shank those who bully Cudder!

>>734
The JS comes later. I'll definitely release a version without any JS support. I don't know how far, since I'm still designing the layout algorithm.

Cudder is a male jew roleplaying as a girl.

can we get a roadmap with what's done so far and what's missing before the first release (acid2 pass)?

>>739
Done:
- HTML tokeniser
- HTML parser
- DOM tree

In progress:
- CSS tokeniser
- CSS parser
- CSS box generation
- layout

Planned:
- Network
- Rendering UI
- Configurator

>>740
Of those, which won't be portable?
Network, of course.
For rendering UI, have you looked into Vulkan? Or do you still plan on doing Win32 and then Xlib, separately?

Vulkan

No. No fucking "cross platform" bullshit, especially something that isn't even really out yet. We don't need another bloody layer of bloat or a dozen multi-MB libraries. Windows version will be single binary expected to be <1MB and usable on Win98SE up to (probably, won't test) 10.

I don't really know what to do with *nix yet. Horribly fragmented ecosystem with tons of unstable dependencies. AFAIK there's nothing like a MSVCRT.DLL that I can just link with and get all the standard C library functions on any version of Linux. Instead there is this whole fucked-up symbol-versioning GNU bloatstrosity with all the different libc versions. And that's not even getting into the retarded ELF symbol export/import system which is more like "throw everything together randomly". The PLT/GOT/etc. crap is headache-inducing every time I have to deal with it.

If I was making a hybrid OS with the best features of Windows and *nix, it would definitely have dynamic linking more like the Windows system.

>>742
Forget loonix, it's too shitty to use anymore. Stick with real UNIX® and VMS.

Instead there is this whole fucked-up symbol-versioning GNU bloatstrosity with all the different libc versions

Just static link against musl libc and forget GNU anything.
http://www.etalabs.net/compare_libcs.html

Static linking against musl will make it work on any linux, regardless of the installed libc.

>>740
Lol, just parsing the text into the proper data structures can be done in an afternoon, you fucking n00b.

>>742
Static link you fucking ifiot

If cudder ever releases anything (I know, let me dream for a second), I'd put money on it being slower than modern browsers, with a similar footprint as browsers of similar limited functionality.

CUDDLER why aren't you implementing your own PCB traces and transistors?

>>748
Real world considered harmful.

>>749
Did somebody just say "Haskell"?

>>749-750
Stop making it look like I'm replying to myself. I'd like a real reply, preferably from CLUTTER shirmself.
Anyway, CUTLERY lives in a world with neither deadlines nor abstract thinking, so >>749 is right and >>750 is wrong.

>>751
But >>750 didn't claim anything, how can he be wrong?

>>752
You know when you show your sister your hentai collection, and she says "that's just wrong"? It's like that.

>>753
Cudder should rewrite Windows in Haskell, though.

I'll show all you cocksuckers when I make a web browser out of vacuum tubes!

>>755
I'll show all you cocksuckers when I make a hydraulic computer!

I'll show all you cocksuckers when I make a penis pump computer!

>>744,746

static link

That's even worse.

>>745
Yes, a shitty inefficient version can be done in almost no time. I could probably have an acid2-passing horribly bloated browser-ish thing if I just used C++ and banged out classes all day long for a week. That's not the point. It's taking so long because [1] I have tons of other stuff to do and [2] I'm thinking carefully, very carefully, about how to do it. Amortised over the past few months I've probably been spending less than an hour a day on this.

>>747 start reading around >>123. Parser is overall twice as fast as Chrome's, less than 1/10th the size, and that's with a stupid C part I haven't done much optimisation on yet.

Couldn't CUDDER, you know, fork Webkit and add his own transvestite assembly code instead of reinventing the wheel?

>>758
namefags need to go