Slashdot Mirror

Sounds like you want http://freshmeat.net/projects/http-replicator:

"HTTP Replicator is a general purpose, replicating HTTP proxy server. All downloads through the proxy are checked against a private cache, which is an exact copy of the remote file structure. If the requested file is in the cache, replicator sends it out at LAN speeds. If not in the cache, it will simultaneously download the file and stream it to multiple clients. No matter how many machines request the same file, only one copy comes down the Internet pipe. This is very useful for maintaining a cache of Debian or Gentoo packages."

I strongly suggest apt-proxy for Debian. It's drop-dead easy and Just Works great!

The tool's homepage http://gertjan.freezope.org/replicator/ has a nice list of similar tools, but is out of date with respect to apt-proxy, which is not a bash script anymore but is Python now (IIRC).

The only problem I can find with it is there's no RPM for it in CentOS. I'll get around to building one Real Soon Now...

There are quite a few regular expression tools available, with different capabilities and purposes. For the novice who doesn't want to learn more or doesn't have time, the best is probably txt2regex, which walks you through the construction of the regexp and generates output for 20 different programs and languages. It is one of the few tools that I know of that isn't specialized for a particular language or program. My own tool, Redet, provides an interface to 29 regular expression implementations. It is aimed at people who know something about regular expressions or are willing to spend some time learning but helps out by providing palettes showing the notation for each program and a history system, so that you can first construct the pieces of a complex regexp, then assemble them. It also has features aimed at providing a search environment that may be useful for people who need no help constructing their regular expressions.

regex-coach uses PERL-style regular expressions. Its particular virtue is that it can single-step through the match and show the parse tree, so it is useful if you want to understand the matching process in detail. Similar in that it helps to understand the implementation of regular expressions is re_graph, which given a regular expression draws the corresponding finite state automaton.

A couple of nice tools aimed at Python users are Kiki and Kodos.

These and some other tools and libraries are listed on this page.

Who wants the javascript before gmail use obfuscate in code

use save as

There's also a QuotaFolder product for zope, which together with Zope offers exactly what you want. You can see a working setup of all of this at FreeZope

Now, in many companies, your resume will just get thrown out because you don't match some HR monkey's checklist

This is a bit frightening. I haven't been officially employeed (only doing contract work) for a few years now. I would really hate to see what most HR companies make of my resume.

Would I even qualify for an H1B, if I (theoretically) applied to go to the US?

It's widely known that in the dot-com era (mainly the late 90s) technology workers were well paid, and were snapped up pretty fast by startups and big companies alike, often with little more than certification or degree.

What kind of CV (resume) would be acceptable in today's market? Would someone like me, for example, stand any chance of getting a job in the US these days?

I'm sure that all the programmers who worked 20 hours a day to fix many of the problems, and the governments worldwide who allocated billions to fund plans to fix these problems would be thrilled to hear you say that Y2K "just wasn't that big a deal".

Yes, the problems were blown out of proportion by some people (Gary North, for example), but ignoring the real issues that did exist at the time is just as stupid. I don't think it would have been apocalyptic if these issues weren't addressed as they were, but it would certainly be a huge headache, especially in the financial and business sectors. The amount of private contracts that I and a lot of people that I know got for converting legacy DOS apps to Windows or Linux (the two most common) would probably surprise you. A lot of them didn't even have the source code, so complete re-writes were neccessary.

First, there's tvision, a port of Borland's turbovision to gcc. Unfortunately turbovision is not in the public domain as could be deduced from that page, so tvision is not free software and on shaky legal grounds, but borland doesn't seem to mind.
Another interesting project is twin, which is a text-mode windowing environment, something like screen with a TUI.
nstti, the not so tiny text interface, might make a good starting point if you decide to write your own in python.
And there was this butt-ugly GUI that worked directly on vga hardware that would've been fine for POSes, but I can't find it right now.

The solution to this is to compile Python with the right options against the correct libc.

The FreeZope.org server we run has a ZODB of on average about 8 GB and occasionally grows to over 10 GB.

In short, recompile Python like this:

export CC="gcc -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64"
./configure --prefix=/some/directory --with-thread; make; make install

Warning, this probably is very linux specific :)

When I was first doing the main layout for my web site, I tried using fixed positioning for the sidebar. Unfortunately, it doesn't work in IE, so I had to scrap it. Similar things have happened when I used other perfectly valid HTML constructs that weren't supported by the web browser that delivers "The Web the Way You Want It".

A while ago, I wrote down some thoughts about Bayesian web filtering. The main point that I had was this: with web filtering, the people on both sides of the HTTP connection are on the same side, rather than having a greedy asshat spamming millions of decidedly unappreciative victims. You can make, say, porn sites that are indistinguishable from a discussion of the W3C recommendation process. How is a bayesian filter to know the difference? With email, the HTML tags are very suspicious; the IMG tag gets a very high spam probability. But with web pages, what can you do?

Imagine the mess that would cause. There are already far too many "professionally" designed web sites that refuse to work without Internet Explorer, such as Amano's World. Can you imagine the nightmare for users of other browsers if IE became actually required? What about proxy servers? I currently use bfilter, which blocks out the worst of the web before it gets to me. I imagine that it wouldn't be allowed. So we limit proxies to... whatever microsoft provides? That and approved versions of squid if these "control the way you look at my web site even if it is completely anathema to the web's philosophy" people are feeling generous. And what about plugins for proxies that have such an architecture? Would it be necessary to disallow third party plugins/modules/whatever, therefore destroying the hard-earned flexibility that programmers are always striving for?

Actually, I think that alone would pretty much kill those W33 0wn j00 web sites. Still, I much prefer to give power to the users, since that's what I am. Control freaks give me the willies.

So I cooked up this logo for my local linux users group, and they nominated me for Sun's regional delegate program. Wow! Being flown across Australia to attend LCA2003, what a blast.

<impressions>
...everyone knows linux, conversations start from that point,
...Saying hi to 5 people before my first coffee of the day;
...the kernel, is it too big? Can these guys pull it off?
...Where are the gcc people? Is this the GNU/Linux divide? It seems the kernel is in bed with GCC.
...Okay, some things are simpler in 2.5, phew!
...Another nametag, another famous person, another nerd, another one i've seen on my favourite mail list.
...in a room with 200 others (blechh) going over the SMP scheduler with Rusty... Do these people really know what this is about, or are they just groupies?
...Finding Linus and fam at a cafe; like royalty...
...So, GUI toolchain hackers get all the babes... i'm workin on it, i am. yes.
...Being heard, asking questions, we all geeky here, no need to be shy
...Meeting sun guy in toilet, thankyou thankyou, yes send more people next year!
</impressions>

The free account is somewhat limited - you don't get full control, but it should be a nice first start.

Yes, these methods can all be easily circumvented to somebody that really wants to get in.

1. Enable WEP...and enjoy the 20% bandwidth loss. Airsnort.

2. Change the default SSID. SSIDs are not needed to zero down on the AP. Triangulation and GPS are effective enough.

3. Disable "broadcast SSID". The beacon frames can easily be captured otherwise. Attack by enabling your cards monitor mode (not to be confused with promiscious mode which only captures packets on the current network), sniffing all air traffic.

4. Change the default password of your access point. However, I'd like to point out even changing ones password can be insecure. My access point, and I'm sure others as well, send the admin password in a urlencoded form, unencrypted, in plaintext for anyone with a monitor-mode NIC to sniff.

5. Control access via MAC addresses. Spoofing as trivial as ifconfig eth0 down; ifconfig eth0 hw ether 00:00:00:00:00:01; ifconfig eth0 up. On OpenBSD use sea.c. Use arping to sniff MACs.

6. Disable DHCP in your wireless router. Static IP addressing, subnet range determined from arping. Private addressing:

• 10.0.0.0 - 10.255.255.255 (10/8 prefix)
• 172.16.0.0 - 172.31.255.255 (172.16/12 prefix)
• 192.168.0.0 - 192.168.255.255 (192.168/16 prefix)

7. Change your IP subnet. See above.

8. Move your access point away from windows. No match for high-gain antennae.

Hope this helps an aspiring black hat! Remember, no network is completely secure. WaReZ anyone?

Slashdot uses mod_gzip (watch your connections sometime if you don't believe me), so apparently they decided CPU time was cheaper than bandwidth, at least on very large scales.

That should be endorsment enough, right?