Slashdot Mirror

Gatech Sunsite.dk Norweigen Mirror Another Norweigen Mirror

While that tutorial gets you quite a bit for XUL coding, the overall documentation for Mozilla is sparse. I've been working on a bug for a couple weeks now, and in the process I've learned a lot of how Mozilla works, but I've had to do it the hard way. I use a lot of find and grep to trace conceptual maps of data flow and how Moz keeps track of certain things.

That's the hard way. Try a source indexing program like Source Navigator, or my personal favorite, LXR. The latter requires that you figure out how to install and configure a CGI script to run under Apache (or whatever httpd), so it takes some effort to get it going, but being able to surf into localhost and browse all your source, hyperlinked, witht he browser of your choice, is a decadent luxury that just can't be beat. Get it here if you're interested. The source is on the site, and there is also an ambitious lxr2 project out there, using PostGres as the back end.

Again, this has been performed by the Bergen LUG in Norway. Overseen by Alan Cox. Here's a ping:

64 bytes from 10.0.3.1: icmp_seq=0 ttl=255 time=6165731.1 ms

See their homepage here.

This has been tried by the Bergen LUG in Norway with Alan Cox watching. See this page http://www.blug.linux.no/rfc1149/

train fruit flies (instead of carrier pigeon) to carry the messages

I don't know of an RFC on ICMP packets via fruit flies. Perhaps one needs to be written.

Also see the World's First RFC 1149 implementation

D

Always think they know better. This is a much better protocol for transmiting wirelessly. CPIP (carrier pigeon internet protocol).

Not only does each node cost a heck of a lot less, but they're self-replicating!

IIRC it happened a year ago, upon the implementation of RFC 1149.

Together from TogetherSoft is an excellent reverse-engineering package for C++ and Java. There is a free eval that does most everything, but will not print or export diagrams. The full-version license is pretty pricy (around $3k) from what I hear, the website doesn't list a price.

Also, to find references to particular methods and subroutines, LXR is very useful. LXR is designed for use with the Linux codebase, but it is generic enough to be used with any C/C++ project. It takes a couple of hours to get running, but it is free and very cool. I just wish it would support Java.

Speed of Carrier Pigeon Measurement Using Ping

I would say yes, since the 2.2 kernel also uses these Pentium big pages:
http://lxr.linux.no/source/arch/i386/mm/init.c?v=2 .2.19#L302

We also meet systematic freezing with XFree86 on our Athlon+Ati/AGP.

I tried to read some Linux (the kernel) code a long while ago ... it gave me a headache

Funny, when I compare two source files from FreeBSD and other from Linux, I find the Linux version to be much easier to read. Oh well, I guess I just prefer Linus' coding conventions better.

Maybe it's another implementation of RFC1149?

For those that don't remember RFC 1149, it basically specifies a protocol for IP over avian carriers, CPIP (carrier pigeon internet protocol).

There are details of of the first implentation of this protocol (including pictures) here...

Funny Stuff.

Buddy system. Hairy. You really aren't expected to understand this.

Wait, the Kernel uses GCC extensions? I thought the Kernel was written in real C, not that bastard GCC version. I've never look at Kernel code, so I'm not sure. Is this really true?

Here's some kernel code. Now you've seen it.

If it's true, I think that's a huge mistake. The Kernel should not be at the mercy of one compiler.

Why not? The major goal of operating system design is to extract as much performance as possible with as little overhead as possible. Portable code by definition is rarely as efficient as code targetting a specific platform or compiler.

Oh yeah, Linux uses the BSD socket API too, doesn't it? Surely that means they took BSD's freely-usable code, rather than reimplementing it, right? Well... maybe not. Despite those files having the same name as their BSD counterpart, they're nothing alike.

Oh yeah, Linux uses the BSD socket API too, doesn't it? Surely that means they took BSD's freely-usable code, rather than reimplementing it, right? Well... maybe not. Despite those files having the same name as their BSD counterpart, they're nothing alike.

The kids don't need to Play Quake or Half Life @ school(well I do but thats another subject :-))

I think you need to take a closer look at RFC 1149 the Carrier Pigeon Internet Protocol (CPIP). According to this site Pigeons can fly 40-50 Miles per hour, so that would give you about 2 hours per webpage. I would reccomend some major caching servers on the school end. Also it would be a good Idea to invest in a Pigeon farm, this would also give your students great expereince in farm managment.

Another idea is to break some standards and perhaps use an UDP tunnel betweent he sites on your pigeons to make sure you don't have to do all the tcp handshake stuff. O yeah.. .perhaps using larger packet sizes would be a good idea too.

2001: Privacy advocates scream Big Brother up to no good and bemoan loss of civil liberties. GOVNET building commences.

2002: Microsoft proposes "GOV.NET". U.S. Government is impressed with plan to hand control of U.S. Government Information Technology Infrastrucure over to private company in return for promises that Microsoft Operating Systems on exisiting Infrastructure will function more reliably.

2003: GOVNET protocol figured out by shadowy hacker or group of hackers known as DarthBilbo and disseminated in Spam Usenet posts: 14 year old turns Department of Interior server into Gnutella node, 22 year old uses California Homeland Defense Office server to stash 100 gigs of porn and pirate music. Chinese hackers splash "F*** Poisonbox" on homepage of over 100 government servers.

2004: FARC and IRA Terrorist sympathizers, despondent at the decimation of their ranks in the past 3 years by U.S. Anti-Terror activity, launch crippling DoS attacks with Code Red, White, and Blue worm. Unpatched versions of IIS9 installed on 98% of GOVNET Servers. Entire GOVNET shut down. Worm works via social engineering exploit whereby all GOV.NET Server Admins get email saying: "Hi! How are you? I send you this file in order to have your advice. See you later. Thanks". IIS9 patch would have prevented this exploit by denying the Admins access to Server rooms by revoking the revolving monthly license for their Microsoft Passport service on their National Identification Cards. Nation awaits crippling of U.S. Infrastructure and mass confusion.
Nothing happens. An office worker in a State Department office building is quoted on the evening news as saying "Those GOVNET boxes? They never worked anyway. We just use them to play Solitaire and FreeCell. We all use our personal AOL email accounts and chat rooms to conduct important government business."

2005: President Rudolph Giuliani announces plans for GOVNET2. Based on earlier work on the RFC 1149 implementation. Privacy advocates scream Big Brother up to no good and bemoan loss of civil liberties. Building commences.

How many other things started out as an April Fool's day joke and then actually got implemented?

Well, the classic example has to be RFC1149, A Standard for the Transmission of IP Datagrams on Avian Carriers

Where is that TCP/IP carrier pigeon again?

IP over Avian Carriers is specified by RFC 1149, updated by RFC 2549. The Bergen LUG made a prototype implementation, which /. covered last April. Ping time was about an hour, with 50% packet loss.

So the real question is, will this make an RFC 3091 protocol happen more rapidly?

Sig: Tell all your friends NOT to download the Advanced Ebook Processor:

"NOTE! This copyright does *not* cover user programs that use kernelservices by normal system calls - this is merely considered normal use of the kernel, and does *not* fall under the heading of "derived work".

> Stackable vnodes, which are supposed to address this same need, are ancient history

I think I gave up on the idea that Linux would ever know what it was doing with filesystems when I saw this abomination. That's right, Linux not only has no concept of vnodes, it actually uses a union of every single filesystem type for its inode data structure. Search lxr for "vnode" if you don't believe me. So much for generic, to say nothing of modular.
--

Wow! Pneumatic tubes must be exactly what The RFC 1149 project needs. Imagine using these for LAN and corporate networks - add a few hundred pigeons, a barrel of grease and voila! High-speed CPIP communication.

More pictures, this time mostly from the other site, is at http://www.blug.linux.no/rfc1149/bjoff_bilder/

- Vegard

Eric S. Raymond has decided to include this event into The Jargon File version 4.3.0. He salutes the RFC fetishists of BLUG and calls them "wonderfully insane hackers", in addition to describing the event as "what may just be the funniest, cleverest hack of the last fifteen years" in a Linux Today article.

The Jargon entry can be found here.

See for yourself...

Pic here.

"Hey Alan, stop bogarting. Pass it!"

I propose that once testing has been completed on a fully standards complient version of RFC 1149, testing on the implementation of RFC 2549, or "IP over Avian Carriers with Quality of Service", should begin. This extention of RFC 1149 adds many important features, such as quality of service, security, and traffic shaping.

Audit trails are automatically generated, and can often be found on logs and cable trays

Actually, RFC 1149 is being implemented:
http://www.blug.linux.no/rfc1149/

FreeBSD already has something similar and a bit more useful. There is a hypertext, cross referenced presentation of the entire CURRENT branch of the kernel. It is not quite as pretty as the image of the linux kernel: http://lxr.linux.no/freebsd/source

I would then try to browse the source with a hyperbolic web visualization tool, such as Inxight or Webviz from the Geometry Center (RIP). Apparently there are many such visualization tools. Perhaps one will work well with LXR output.

I've seen a lot of systems popping up on the web over the last few years to allow dynamic additions of content to an information system. Things that allow commentary to be added to webpages, which others can then view; sites like Everything or Everything2.

Why not such a system for documenting code? I know that systems for publishing and linking code to itself exists (like LXR). How about such a system that would allow links to be placed in the text to user-contributed documentation? Said documentation could be anything from "this statement is doing such-and-such" to an overview of an entire module.

This documentation would be user-contributed and, of course, user edited. Editing would need to be done based on a voting system... just saying whether a given bit of doc is useful ought to be enough. Attribution is easily done, as well.

The hardest bit would probably be telling the system where you want to place a link. Do you want to doc the line? The function definition? That word? These 3 functions? That bit of code and that one over there in a different file that happen to work together? Where does the link go?

Anyone have an idea on how to do that? I know I'm up for contributing to the development of such a system (playing with Zope has gotten me interested in dynamic web stuff).
---

Of course there is a browsable kernel source tree, and it is wicked cool. You can click on any function/variable/etc and get the definition. The page is http://lxr.linux.no/

One of the early ideas I had was something along these lines, as well. A site where people would post their code, and it could be peer-reviewed in the same way that people post comments on Slashdot now, crossed with an LXR-type display with line numbering, named anchors, etc. It's a great idea, I think, but extremely difficult to implement well -- for example, how do you attach comments to particular lines without completely disrupting the flow of the code? There are (mostly non-portable) hacks, such as using DHTML to make the comments collapsible, but things like that might detract from the content of the site.

Of course, there's very little non-programmer appeal to this sort of a site, and even most programmers would probably only be interested in a few of the postings at any given time. How would you like to log in one day and see only LISP and Applescript stuff? I think I'd get pretty annoyed and stop visiting after a few days of that.... (no offense meant if you happen to use either of those languages; I like LISP myself).

darren

not to nitpick, but lxr was designed to cross-reference the linux kernel.

In FreeBSD, if there is only one program that uses the FPU, there will be no FPU context switches at all. How can Linux be faster than that? :-)

Read some code: switch.s and npxsave

-T

In FreeBSD, if there is only one program that uses the FPU, there will be no FPU context switches at all. How can Linux be faster than that? :-)

Read some code: switch.s and npxsave

-T

Amiga Linux[1] has existed for many years. It was the first port of Linux to non-x86 hardware, and was done by Amiga people who wanted a better[2] OS.

You talk about the essence of the Amiga coming from it's wonderful multimedia hardware and the people around it, and I agree. My first two computers were Amigas and I loved them. However, time has moved on. After Commodore's bankruptcy the Amiga community has been scattered to the winds, leaving only a tiny hard core, and the hardware and OS that made the Amiga amazing in 1989 are now obsolete.

I don't see any reason why we old Amigans should get excited about this, or even any reason to call the new machine an Amiga: it will be entirely different hardware running an entirely different OS. It will presumably have an Amiga-like GUI but such things are already available[3]. If it runs old Amiga binaries it will in effect be running an Amiga emulator but such things are already available [4]

The Amiga was a wonderful machine in its day but I'm afraid it's time to let it die...

[1] see http://lxr.linux.no/source/arch/ m68k/amiga/?a=m68k and http://www.linux-m68k.org/faq/history.ht ml. Although the amiga version wasn't reintegrated into the linux source 'till 2.0 it did exist as a separate project.

[2] for their definition of better, which is not necessarily yours.

[3] http://www.lysator.liu.se/~marcus/amiw m.html

[4] UAE is written in C and runs under many platforms and OSes.
Fellow is written in x86 assembler and runs only in MS-DOS but is damn fast.
I recommend having a look at one of these. Certainly brought back memories for me

Does he mean something like a system call that could directly copy from file descriptor to file descriptor?

Isn't that what sendfile does?