Jared's rambling thoughts

No longer posting to livejournal now

Wed, 19 Jan 2005 20:31:45 GMT

Now that I've completed my migration to a new blogging service which I believe better suits my needs and taste, I will be no longer posting to livejournal. You will need to update your bookmarks or RSS aggregator program.

My new blog is now at:
http://jarednevans.typepad.com/blog/

or you can subscribe to the RSS feed of my new blog at:
http://jarednevans.typepad.com/blog/index.rdf

Using Livejournal for one year was a good learning experience for a new blogger like myself but now I'm ready to move on to a more professional blogging service. With the features on my new blog I really like being able to categorize my posts and I should be adding a search engine link soon.

I won't remove the Livejournal blog anytime soon and leave it up for future reference.

Be seeing you all over at typepad.com!

livejournal to ???

Mon, 17 Jan 2005 00:35:53 GMT

My paid livejournal account is due to expire in the next few days. I am now looking around for a new blogging service. I'll keep you appraised on which one I will move to.

I'm looking at www.typepad.com and am liking what I see so far.

FreeBSD

Fri, 14 Jan 2005 17:57:52 GMT

After a short hiatus from UNIX systems when I focused on Microsoft Windows Server 2003 / IIS 6.0 / ASP.NET, I returned to the open-source scene where I installled two different Unix-like systems on my new laptop.

One of the most popular Linux distro out there is known as Debian 3.0 which purports to be an Linux system that is easy to install and upgrade as new and updated packages are checked into the system.

I got a very cool software called VMWare Workstation which is nothing short of a miracle. VMWare allows you to virtualize operating systems on top of the operating system that you primarily use for your daily stuff. It allows you to run entire operating system inside a window. You are also able take a snapshot of the operating system and mess around with it and then revert back to where it was before you made the modifications!

I was able to get Debian installed and even the X Window System up and running with ease (if you follow the instructions carefully!) under VMWare! After playing with it for several weeks and experimenting with updating packages among other tasks, I wasn't too enlightened by how it all was supposed to click together. While it was easier to do stuff compared to my days with Slackware 0.9 back in 1993, there were some limitations and dependencies problems with package installation/upgrading. Once, i went as far as completely destroying the system when the wrong version of kernel was installed.

Debian agrees with those who are extremely conservative and are willing to stick with older packages. There are several releases that happen at the same time with Debian called stable, testing, unstable. There were versions of several packages I wanted in the testing release but they make sure you know that there is no guarantee that everything in the testing release has been fully tested yet. This caused a bit of disillusion for me, since it was not a easy thing to do to mix in different applications from different releases.

However, in recent days, Doug Sampson at Dawn Sign Press, where I also work, brought up the idea of using FreeBSD. I've heard of this operating system which has experienced many forks in the years past and the cabal of super-developers who have the final say about anything that enters or exits the BSD operating systems.

I decided to give it a whirl (under VMWare of course!) and all I can say is WOW! I now believe that FreeBSD is what Linux aspires to be. FreeBSD is different from Linux in that it is one entire coherent system, rather than having applications slapped on with the kernel and hope it all works together. (...I can hear the hordes of rabid Debian users descending on me now)

It is a normal thing for a BSD administrator to compile the ENTIRE system or upgrade many applications by recursively compiling everything to resolve any dependencies problems. At the first breath, this seems dangerous to the new-comer... "what if there is one teensy-bitsy bug that causes the system compilation to fail?" The only way you will learn is to jump feet first into the water and get wet. It's only through the process of working through problems that you will gain the skills necessary to be able to handle these kind of systems when the inevitable comes sooner or later.

I was apprehensive at first when I faced the necessity of upgrading many applications at one go to fix various security holes that were discovered. You can imagine my relief after 12 hours of compiling when it was successful with zero errors and I was running a system that had all known holes patched up. It was a very good feeling afterwards.

The philosophy of FreeBSD being one entire system is in fact one of its strengths since you can be very confident that whatever the developers introduce into the system is guaranteed to work with the system and will not cause any conflicts. This is why I highly, highly, HIGHLY recommend that you use what is known as the ports collection on FreeBSD.

You will be able to easily keep the ports up to date and ensure that you install applications that will be compatible with your system and at the same time be able to ensure that discovered security holes are rapidly closed. This is a cool thing because after the developers implement a recent version of an application that you highly desire, they can check it into the system to be rigorously tested. After the tests are passed, it is approved to be entered into the ports. Once it's available, you can go ahead and install it on your system and be fairly confident it will not break anything.

If you are curious and want to know more about the different design philosophies between FreeBSD and Linux, read this excellent link!

I have set up a nice FreeBSD configuration with several must-have applications. I look forward to using this system and getting more comfortable with it. I'm going off now to install my favorite open source database called PostgreSQL!

tip on using your SKI,II pagers

Wed, 12 Jan 2005 23:11:05 GMT

If you have several folders in your SideKick mail application, you can send email directly to that folder!

For example, if you have a mail folder called "Important",

Send the emails to:

Important#yourusername@tmail.com and yes, be sure that it matches the capitalization of your folder Important, not important

and you should see the email show up in your "Important" folder!

Hacker penetrates T-Mobile systems

Wed, 12 Jan 2005 18:46:16 GMT

This had to happen sooner or later....

Update:
cnn.com has an online article about this too:
http://www.cnn.com/2005/LAW/01/12/cellular.hacker.ap/index.html

------------

Hacker penetrates T-Mobile systems

By Kevin Poulsen, SecurityFocus Jan 11 2005 7:43PM

A sophisticated computer hacker had access to servers at wireless giant T-Mobile for at least a year, which he used to monitor U.S. Secret Service e-mail, obtain customers' passwords and Social Security numbers, and download candid photos taken by Sidekick users, including Hollywood celebrities, SecurityFocus has learned.

Twenty-one year-old Nicolas Jacobsen was quietly charged with the intrusions last October, after a Secret Service informant helped investigators link him to sensitive agency documents that were circulating in underground IRC chat rooms. The informant also produced evidence that Jacobsen was behind an offer to provide T-Mobile customers' personal information to identity thieves through an Internet bulletin board, according to court records.

Jacobsen could access information on any of the Bellevue, Washington-based company's 16.3 million customers, including many customers' Social Security numbers and dates of birth, according to government filings in the case. He could also obtain voicemail PINs, and the passwords providing customers with Web access to their T-Mobile e-mail accounts. He did not have access to credit card numbers.

The case arose as part of the Secret Service's "Operation Firewall" crackdown on Internet fraud rings last October, in which 19 men were indicted for trafficking in stolen identity information and documents, and stolen credit and debit card numbers. But Jacobsen was not charged with the others. Instead he faces two felony counts of computer intrusion and unauthorized impairment of a protected computer in a separate, unheralded federal case in Los Angeles, currently set for a February 14th status conference.
On July 28th the informant gave his handlers proof that their own sensitive documents were circulating in the underground marketplace they'd been striving to destroy.
The government is handling the case well away from the spotlight. The U.S. Secret Service, which played the dual role of investigator and victim in the drama, said Tuesday it couldn't comment on Jacobsen because the agency doesn't discuss ongoing cases-- a claim that's perhaps undermined by the 19 other Operation Firewall defendants discussed in a Secret Service press release last fall. Jacobsen's prosecutor, assistant U.S. attorney Wesley Hsu, also declined to comment. "I can't talk about it," Hsu said simply. Jacobsen's lawyer didn't return a phone call.

T-Mobile, which apparently knew of the intrusions by July of last year, has not issued any public warning. Under California's anti-identity theft law "SB1386," the company is obliged to notify any California customers of a security breach in which their personally identifiable information is "reasonably believed to have been" compromised. That notification must be made in "the most expedient time possible and without unreasonable delay," but may be postponed if a law enforcement agency determines that the disclosure would compromise an investigation.

Company spokesman Peter Dobrow said Tuesday that nobody at T-Mobile was available to comment on the matter.

Cat and Mouse Game
According to court records the massive T-Mobile breach first came to the government's attention in March 2004, when a hacker using the online moniker "Ethics" posted a provocative offer on muzzfuzz.com, one of the crime-facilitating online marketplaces being monitored by the Secret Service as part of Operation Firewall.

"[A]m offering reverse lookup of information for a t-mobile cell phone, by phone number at the very least, you get name, ssn, and DOB at the upper end of the information returned, you get web username/password, voicemail password, secret question/answer, sim#, IMEA#, and more," Ethics wrote.

The Secret Service contacted T-Mobile, according to an affidavit filed by cyber crime agent Matthew Ferrante, and by late July the company had confirmed that the offer was genuine: a hacker had indeed breached their customer database,

At the same time, agents received disturbing news from a prized snitch embedded in the identity theft and credit card fraud underground. Unnamed in court documents, the informant was an administrator and moderator on the Shadowcrew site who'd been secretly cooperating with the government since August 2003 in exchange for leniency. By all accounts he was a key government asset in Operation Firewall.

On July 28th the informant gave his handlers proof that their own sensitive documents were circulating in the underground marketplace they'd been striving to destroy. He'd obtained a log of an IRC chat session in which a hacker named "Myth" copy-and-pasted excerpts of an internal Secret Service memorandum report, and a Mutual Legal Assistance Treaty from the Russian Federation. Both documents are described in the Secret Service affidavit as "highly sensitive information pertaining to ongoing USSS criminal cases."

At the agency's urging, the informant made contact with Myth, and learned that the documents represented just a few droplets in a full-blown Secret Service data spill. The hacker knew about Secret Service subpoenas relating to government computer crime investigations, and even knew the agency was monitoring his own ICQ chat account.

Myth refused to identify the source of his informational largesse, but agreed to arrange an introduction. The next day Myth, the snitch, and a third person using the nickname "Anonyman" met on an IRC channel. Over the following days, the snitch gained the hacker's trust, and the hacker confirmed that he and Ethics were one and the same. Ethics began sharing Secret Service documents and e-mails with the informant, who passed them back to the agency.

Honeypot Proxy
By August 5th the agents already had a good idea what was going on, when Ethics made a fateful mistake. The hacker asked the Secret Service informant for a proxy server -- a host that would pass through Web connections, making them harder to trace. The informant was happy to oblige. The proxy he provided, of course, was a Secret Service machine specially configured for monitoring, and agents watched as the hacker surfed to "My T-Mobile," and entered a username and password belonging to Peter Cavicchia, a Secret Service cyber crime agent in New York.

Cavicchia was the agent who last year spearheaded the investigation of Jason Smathers, a former AOL employee accused of stealing 92 million customer e-mail addresses from the company to sell to a spammer. The agent was also an adopter of mobile technology, and he did a lot of work through his T-Mobile Sidekick -- an all-in-one cellphone, camera, digital organizer and e-mail terminal. The Sidekick uses T-Mobile servers for e-mail and file storage, and the stolen documents had all been lifted from Cavicchia's T-Mobile account, according to the affidavit. (Cavicchia didn't respond to an e-mail query from SecurityFocus Tuesday.)

By that time the Secret Service already had a line on Ethic's true identity. Agents had the hacker's ICQ number, which he'd used to chat with the informant. A Web search on the number turned up a 2001 resume for the then-teenaged Jacobsen, who'd been looking for a job in computer security. The e-mail address was listed as ethics@netzero.net.

The trick with the proxy honeypot provided more proof of the hacker's identity: the server's logs showed that Ethics had connected from an IP address belonging to the Residence Inn Hotel in Buffalo, New York. When the Secret Service checked the Shadowcrew logs through a backdoor set up for their use -- presumably by the informant -- they found that Ethics had logged in from the same address. A phone call to the hotel confirmed that Nicolas Jacobsen was a guest.

Snapshots Compromised
Eight days later, on October 27th, law enforcement agencies dropped the hammer on Operation Firewall, and descended on fraud and computer crime suspects across eight states and six foreign countries, arresting 28 of them. Jacobsen, then living in an apartment in Santa Ana in Southern California, was taken into custody by the Secret Service. He was later released on bail with computer use restrictions.

Jacobsen lost his job at Pfastship Logistics, an Irvine, California company where he worked as a network administrator, and he now lives in Oregon.

The hacker's access to the T-Mobile gave him more than just Secret Service documents. A friend of Jacobsen's says that prior to his arrest, Jacobsen provided him with digital photos that he claimed celebrities had snapped with their cell phone cameras. "He basically just said there was flaw in the way the cell phone servers were set up," says William Genovese, a 27-year-old hacker facing unrelated charges for allegedly selling a copy of Microsoft's leaked source code for $20.00. Genovese provided SecurityFocus with an address on his website featuring what appears to be grainy candid shots of Demi Moore, Ashton Kutcher, Nicole Richie, and Paris Hilton.

The swiped images are not mention in court records, but a source close to the defense confirmed Genovese's account, and says Jacobsen amused himself and others by obtaining the passwords of Sidekick-toting celebrities from the hacked database, then entering their T-Mobile accounts and downloading photos they'd taken with the wireless communicator's built-in camera.

The same source also offers an explanation for the secrecy surrounding the case: the Secret Service, the source says, has offered to put the hacker to work, pleading him out to a single felony, then enlisting him to catch other computer criminals in the same manner in which he himself was caught. The source says that Jacobsen, facing the prospect of prison time, is favorably considering the offer.

Banda Aceh

Tue, 11 Jan 2005 21:08:23 GMT

Banda Aceh on the northern tip of Indonesia was one of the hardest hit area of all. The place was very close to the epicenter of the undersea earthquake. Complete and utter destruction when compared to the other areas that were hit.

Now I don't have to wonder anymore about that!

Sat, 08 Jan 2005 03:28:32 GMT

Since cats are always supposed to land on their feet when they are tossed, what happens when there is no gravity?

Watch a kitty tossed in zero gravity

102" plasma TV display!

Sat, 08 Jan 2005 01:57:30 GMT

This would be a dream come true if I was able to get one at home!

This one is the 102" plasma TV display! How much would it be? Get this: the older 80" plasma TV display was only $45,000! Make a wild guess on the 102" model! Would the Asian women come with the TV at that price? :)

making a Tsunami donation

Mon, 03 Jan 2005 20:40:19 GMT

One of my favorite authors, Arthur Clarke (2001: A Space Odyssey) is safe in Sri Lanka (where he now lives). After reading his remarks over at his foundation, I donated some money via paypal to Sarvodaya. You can make a donation to them too.

You can view their photos at
http://www.flickr.com/photos/sarvodaya/sets/69733/

hardest hit area from the Tsunami

Mon, 03 Jan 2005 17:52:07 GMT

Indonesian province of Aceh

Before the tsunami hit

Afterwards:

claiming with technorati

Mon, 03 Jan 2005 03:16:53 GMT

Technorati Profile

Grabbing information about local Deaf Movies into a RSS file

Thu, 30 Dec 2004 00:40:06 GMT

This is the perl script that will automatically screen scrape the website at ncam.wgbh.org to extract the movie you are interested in.

Jared

------------

use HTML::TokeParser;
use XML::RSS;
use LWP::Simple;

use XML::RSS;
my $rss = XML::RSS->new( version => '2.0' );

$rss->channel(
                title        => "Deaf Movies at AMC Fashion Valley RSS",
                link         => "http://ncam.wgbh.org/mopix/nowshowing.html#CA",
                description => "Quick checkup on weekly Deaf movies",
                );

my $url2parse;
# Setup argument to your Netflix RSS
my $arg = "http://ncam.wgbh.org/mopix/nowshowing.html";
# Get the URL, assign it to url2parse, and then parse the RSS content
$url2parse = get($arg);
die "Could not retrieve $arg" unless $url2parse;

$p = HTML::TokeParser->new(\$url2parse);

while (my $token = $p->get_tag("a")) {
         my $text = $p->get_trimmed_text("/a");
         if ($text eq 'AMC Fashion Valley 18') {
         #if ($text eq 'AMC 30 At The Block (near Los Angeles)') {
        #print "text: $text\n" ;
        #print "San Diego\n";
        my $td1token = $p->get_tag("td");
         my $td1text = $p->get_trimmed_text("/td");
        #print "td1: $td1text\n";
        my $td2token = $p->get_tag("td");
         my $td2text = $p->get_trimmed_text("/td");
        #print "td2: $td2text\n" ;
        my $td3token = $p->get_tag("td");
         my $td3text = $p->get_trimmed_text("/td");
        #print "td3: $td3text\n" ;
                if ( $td2text eq "* No Information Available at Time of Posting") {
                    $td1text = "no info at this time";
                    $td3text = "no info at this time";
                }

                $rss->add_item(
                                title => $td2text,
                                link => "http://ncam.wgbh.org/mopix/nowshowing.html#CA",
                                description => "$text --- $td1text --- $td3text" ,
                                );
         }
}

print $rss->as_string;

My technical blog

Sat, 25 Dec 2004 19:43:00 GMT

I've set up a new blog for techncial stuff only. It's at:
http://jaredevans.blogspot.com/

ASP.NET development under Windows XP Pro / IIS 5.0 vs. Windows Server 2003 / IIS 6.0

Wed, 22 Dec 2004 19:24:52 GMT

Hopefully, someone out there in the blogosphere will find this via feedster or technorati and offer feedback on this...

I have a question about ASP.NET development - I've done development work on Windows 2003 server with Visual Stuido .NET and IIS 6.0 installed.

Recently, I got a laptop with Windows XP Pro installed with Visual Stuido .NET and IIS 5.0. What types of issues can I expect to ecounter while programming/testing on my laptop then moving the files over to the production Windows 2003 server.

I'm concerned that while I may get ASP.NET enabled webpages working on my laptop but when I move the files over to the server, they will break.

Update:
Several people have said it's entirely possible as long as ACLs are set correctly on both machines.

this doesn't seem well-known but does exist

Wed, 22 Dec 2004 06:07:53 GMT

spellbound - spell checker for firefox

windnsea

Mon, 20 Dec 2004 14:58:32 GMT

windnsea1
Originally uploaded by jarednevans.

A photo I look last Friday morning.

Anti-Satellite (ASAT) was an earlier success.

Fri, 17 Dec 2004 20:52:51 GMT

It turns out that we did at least develop the capabilty to hit satelilites with missles from a high-flying jet. I wonder why this wasn't revived lately considering how flagrant the pro-war Bush Administration is.

--------------

F-15 ASAT

In the late 1970s, even before the advent of the Strategic Defense Initiative (SDI), an anti-satellite (ASAT) mission evolved for the F-15 Eagle.

The goal of ASAT weapons is the neutralization of enemy military satellites in the event of war, particularly low-orbiting reconnaissance, ELINT, and ocean surveillance satellites. The Soviets had their own antisatellite program in which a killer satellite would rendezvous with the target satellite and explode. The American equivalent involved the arming of an F-15 Eagle with a missile which would be launched against an orbiting satellite from a zoom climb at an altitude of 80,000 feet.

In 1979, a contract was issued to Vought for an air-launched low Earth-orbit antisatellite vehicle. The Vought ASM-135A that emerged was a two-stage rocket, with a first stage derived from the AGM-69 SRAM-A and a second stage derived from the Altair III rocket. The ASM-135A weighed about 2700 pounds at launch and was 18 feet long. The payload of the ASM-135A consisted of a miniature kinetic kill vehicle which used an infrared seeker to home in on the target satellite, destroying it by impact. No explosive warhead was to be needed.

F-15A 76-0086 was modified for trials with the Vought ASM-135A. The ASM-135A was carried on the centerline station of the F-15. The aircraft had to be specially wired to accommodate the ASM-135A missile, and had to be provided with backup battery, microprocessor, and datalink for midcourse guidance.

Beginning in the early 1980s, captive flights were made with the missile in place, the aircraft zoom climbing to altitudes as high as 80,000 feet. The first actual launch of an ASM-135A from an F-15 took place in January 1984, the missile being aimed at a predetermined point in space. Subsequently, three launches of the ASM-135A were made against celestial infrared sources.

The first and only ASM-135A launch against an actual target satellite took place on September 13, 1985, when F-15A 77-0084 of the 6512th Test Squadron stationed at Edwards AFB took off from Vandenberg AFB and zoom-climbed up to 80,000 feet and then launched the ASAT against the Solwind P78-1, a gamma ray spectroscopy satellite that had been launched in February of 1979. Both the first and second stages fired successfully, and the miniature kinetic kill vehicle separated and homed in on the satellite, destroying it upon impact.

The test was a success in that it demonstrated that the basic concept was feasible. However, it enraged arms control advocates, who saw the test as a violation of a joint US/Soviet treaty forbidding the development and testing of antisatellite weapons. Solar scientists were not happy about the test either, since although the Solwind P78-1 that was killed had officially completed its mission, it was still sending back useful data.

Initial plans were made to modify twenty F-15As for the antisatellite mission and to assign them to the 48th TFS at Langley AFB in Virginia and the 318th TFS at McChord AFB in Washington. These squadrons had each received three or four F-15A/B airframes which had been rewired for ASAT operations. However, Congress was unwilling to permit any further testing of the system, and the ASAT program was officially terminated in 1988.

--------------------------- Another source --------------

Vought ASM-135 ASAT

After that Soviet Union had developed a "killer satellite" to disable other satellites in the 1970s, the U.S. Air Force decided to develop an anti-satellite weapon system. The ASAT (Anti-Satellite Missile) program began around 1977, and in 1979 Vought was awarded a contract to develop an air-launched missile for use agianst low-earth orbit satellites. The ASAT missile, also known as ALMV (Air-Launched Miniature Vehicle), was designed as a multi-stage rocket, which was to be launched by an F-15 Eagle interceptor in a zoom-climb. Captive flight tests with ASAT vehicles on a modified F-15A began in 1982, and the first launch aimed at a predefined point in space occurred in early 1984.

The ASAT missile used the SR75-LP-1 solid-propellant rocket of the AGM-69 SRAM as the first stage and a Vought Altair III (the 4th stage of Vought's Scout B) with a Thiokol FW-4S motor as second stage. It was launched by an F-15 in a high-altitude supersonic climb. The F-15's computer was updated with special guidance algorithms, and the head-up display was also modified to provide additional steering cues to the pilot. This was necessary, because the zoom-climb and missile release had to be flown exactly as calculated to get the missile near the target satellite. The second stage of the ASAT pointed the MHV (Miniature Homing Vehicle) "warhead" in the target's direction, so that the latter's infrared image could be detected by the MHV's telescopic seeker. The MHV was spin-stabilized, used 63 small short-pulse rocket motors for manoeuvering, and destroyed the target by a direct hit at a speed of at least 24000 km/h (15000 mph). The maximum intercept altitude for the ASAT missile was at least 560 km (350 miles), and possibly as high as 1000 km (620 miles).

On 13 September 1985, the first and only destruction of a satellite by an American air-launched missile occurred, when an F-15A launched an ASAT against a retired communications satellite in a 555 km (345 mile) orbit. This was the only full-scale live test of the Vought ASAT missile. Shortly after that, the missile was officially designated ASM-135A. Although "ASM" could be conveniently read as "Anti-Satellite Missile", the designation was non-standard because intercept missiles are normally designated "AIM". Inert test missiles for captive flights were designated CASM-135A.

The USAF planned to procure up to 112 ASM-135A missiles for use by twenty modified F-15A interceptors. However, the program was terminated in 1988, partly for political reasons because the ASAT might violate treaties about the military use of space.

browser share of those who visited my blog

Fri, 17 Dec 2004 19:57:40 GMT

server
Originally uploaded by jarednevans.

A clone of Super Mario Brothers written in DHTML!

Thu, 16 Dec 2004 22:52:29 GMT

Super Maryo World
A clone of Super Mario Brothers written in DHTML!

Firefox AD in New York Times!

Thu, 16 Dec 2004 22:05:27 GMT

Finally the firefox AD made it to the New York Times!

Crazy- Pacman in MS Excel!

Thu, 16 Dec 2004 18:46:49 GMT

Now I've seen it all!

This is an actual implementation of the infamous Pacman video game in Microsoft Excel!!!

I didn't believe it at first until I tested it and it *WORKS* perfectly! wow!

(zoomed in screenshot of the invididual Excel cells)

How Google's Suggest works

Thu, 16 Dec 2004 17:59:22 GMT

Details behind Google's Suggest:

http://www.whump.com/moreLikeThis/link/04103

Pigs!

Thu, 16 Dec 2004 17:31:51 GMT

This is what happens when you are in China at a salad bar in Pizza Hut and the management there decides to enforce a rule: only one trip to the salad bar!

Rest of the pictures here!

geek talk: a problem with Tree::DAG_Node perl module

Thu, 16 Dec 2004 17:10:31 GMT

Tree structures are inherently complex and more difficult to handle in a program. This is why I was happy to find a perl module called Tree::DAG_Node which helps to simplify the work required to deal with trees. However, before I could fully finish a program dealing with the trees, I ecountered a roadblock (and the solution) which I go into more details below:

I was programming with the Tree::DAG_Node perl module and I was able to get it to work up to a point where it just dropped the ball or maybe I am blind to what I am doing wrong.

I am doing a walkdown using callback down a DAG tree. When I hit a certain node, I wanted to process the data further. The problem is that when I insert a foreach statement, the walkdown stops traversing the tree.

Testing the DAG Tree output:

print map "$_\n" , @{$root->draw_ascii_tree};
produces the following diagram so I know the Tree creation was successful:
              |
<Shipment and Tracking Data>
              |
          <2000886>
              |
          <INV1000>
      /--------------\
      |              |
<TrackUPS1234> <TrackUPS4321>

using the following code:

my $found_customer = 0;
$root->walk_down({
   callback => sub {
        my $node = shift;
        print " " x $_[0]->{_depth};
        if ( $_[0]->{_depth} eq 0 ) {
                print "\n" ;
        }
        elsif ( $_[0]->{_depth} eq 1 ) {
                if ( $found_customer eq 1 ) {
                        print "write footer email\n";
                        print " " x $_[0]->{_depth};
                }
                print "header email - Customer\# " . $node->name . " " . $node-> attributes->{'name'}. " " . $node->attributes->{'email'} . "\n" ;
                $found_customer = 1;
$f0 = "";
$f1 = "";
$f2 = "";
$f3 = "";
$f4 = "";
@customized_letter_lines = ();
$f0 = $node->attributes->{'email'};
$f1 = $node->attributes->{'name'};
$f2 = $node->name;

my @letter_lines = @header_lines;

        }
        elsif ( $_[0]->{_depth} eq 2 ) {
                print "Invoice " . $node->name . "\n";
                $f3 = $node->name;
        }
        elsif ( $_[0]->{_depth} eq 3 ) {
                print " Track " . $node->name . "\n";
                $f4 = $node->name;
        }
        else {
                1
        }

    },
    _depth => 0
});

if ( $found_customer eq 1 ) {
print " write footer email\n";
}

output is:

header email - Customer# 2000886 Happy, Inc. jd@dawnsign.com
    Invoice INV1000
       Track TrackUPS1234
       Track TrackUPS4321
write footer email

so far so good, but when I add the foreach statement:

...
my @letter_lines = @header_lines;
#insert the foreach statement
foreach my $letter_line (@letter_lines) {
}
#end insert the foreach statement

The output is now:

header email - Customer# 2000886 Happy Inc. jd@dawnsign.com
write footer email

As you can see, it did not traverse the Invoice(INV) and Track nodes.

I got an answer back from Roel van der Steen (the developer behind FrameMaker::MifTree module which makes extensive use of Tree::DAG_Node):

From the Tree::DAG_Node documentation:
   "
   This function [the callback function] must
   return true or false -- if false, it will
   block the next step: [...]
   "

Your problem is that the foreach loop is the last statement executed in
the callback function, and the foreach returns undef, which evaluates
to false, so the walk_down procedure stops traversing. Adding a "return
1" corrects this problem. In fact, when using callback, make sure you
always explicitly set its return value with a "return" statement.

Problem solved!

Used the B&W conversion again

Wed, 15 Dec 2004 22:58:10 GMT

Retouched B&W photo from Fallingwater