Given my too short stay in Australia, it was deemed necessary I check out Melbourne’s festival scene as Canada’s state of affairs has led to attending headlining tours-only. With the geographical isolation of Oz, a rarity of international acts coming through leads to excessive admission charges to make up for visa and travel costs. Next month Welsh band Future of the Left is $47AUS at The Corner. I paid $15CDN in Toronto at ElMo. $10US at Austin’s Beauty Bar in 2009. So dropping $150 for Harvest Festival of more than a handful of desirable bands hooks into my opportunity cost appreciation. And the lineup for this festival was a doozy for overseas acts Down Under.

My prospective schedule for the day:

• Kevin Devine 12:30 – 1:15PM
• The Walkmen 3:00 – 4:00PM
• TV on the Radio 4:00 – 5:00PM
• Mercury Rev 5:00PM – 5:30PM
• Death in Vegas 6:15 – 7:00PM
• The National 7:00 – 7:45PM
• Mogwai 7:45 – 8:15PM
• Portishead 8:30 – 9:45PM
• The Flaming Lips 9:45 – 10:45PM

Actual happenings:

• The Walkmen 3:40 – 4:00PM
• TV on the Radio 4:00 – 5:00PM
• Mercury Rev 5:00PM – 5:20PM
• stand in line for Brazilian Barbecue 5:20 – 7:00PM
• The National 7:00 – 7:20PM
• Mogwai 7:30 – 8:15PM
• Portishead 8:30 – 10:00PM
• Holy Fuck 10:00 – 10:30PM

Working with the Seam Framework and RichFaces for over a year, I found their respective components work only for the simplest of HTML forms. Nested AJAX views cause all sorts of difficulties once you get into EL expressions for iterated components (e.g., from <rich:datatable>) within <f:subview>-named scopes that require usage of rich:element.

Even after I move to Australia, Rogers manages to get my ire. Emailed below to their support, likelihood of a proper response…?

My most recent wireless invoice contains billing detail:

Oct 02 Early Cancellation Fee 500MB Data Option 100.00

I was under contract with Rogers wireless May 2008 to May 2011. I purchased an unlocked iPhone4 direct from apple.com and activated Mt. Sinai’s iPhone corporate plan in September 2010. At no point was another time-sensitive service agreement signed or even communicated to me by a CSR.

On top of this assumed contract, the most current billing cycle ends Oct 7 yet this ETF appears Oct 2 while I’m also being invoiced for another period into November due to Rogers’ draconian 30 day notice for cancellation policy. If the iPhone plan is for one year, the cancellation date would fall outside my service time.

When I called in to cancel, at no point did the retentions rep indicate an iPhone plan ETF of $100 would be applied. And I’m traveling in Australia now and won’t be calling again using international long distance to clear this up.

I sent the below letter to TSN’s Audience Relations, where relations means you get an auto-reply indicating a cog will read the message but not personally respond. If there’s enough volume on a topic, someone in power at Bell Media may hear about it at the end of the day so you may be better off calling them at 416-384-7660. Also, come to my club’s AFL Grand Final party in Toronto. We’ll figure out a way to show it on TV whether TSN is involved or not.

I’ve come to learn that Bell Media doesn’t plan to air the AFL Grand Final on TSN2, instead showing Rugby World Cup games of Australia v Russia and France v Tonga. The March 18th press release stated,

“The top two teams face off in the AFL GRAND FINAL – the Aussie Super Bowl – airing live on TSN2.”

Football clubs across the country have planned Grand Final parties with the expectation of a live broadcast, just as many have done for more than five years now when it was once annually broadcast by Setanta Sport. The AFL’s party tracker has 11 registered events in Canada that will have hundreds of attendees each.

Last minute scheduling news of a tape-delayed broadcast on TSN this Saturday has lead to a non-announcement through social media of it now being streamed live on tsn.ca. Fans of AFL in Canada already had an ad-free web site to stream the Grand Final and such a haphazard handling of the AFL license will tarnish good will toward the TSN brand while also diminishing the opportunity to attract future viewership to the sport.

At this point, the best case scenario would be for Bell Media to offer a pay-per-view option to all the pubs holding event nights this Friday. AFL Grand Final parties have a social atmosphere akin to Canada’s run at Olympic hockey gold and that setting certainly isn’t attained hunched over a laptop.

I thought about posting this on Craigslist but it’s seriously not even worth the effort meeting people to off-loading such junk. To storage the lot goes!

• 8-port 10mbps Ethernet Soho Hub
  The manufacturer seemingly wants to 100% hide its brand from the packaging, manual, and equipment. Ethernet 8/16-Port 10Base-T Stackable Hub.
• Linksys Wireless-G USB Network Adapter with SpeedBooster
  Ideal for your desktop to avoid whinging about too many cords tangling up precious domestic space. Go wireless and give your missus one less reason to complain.
• Logitech Cordless Desktop 967461-0403
  Wireless keyboard and mouse combo. See above.
• Tekram Ultra SCSI DC-315U
  50 pin. What up.
• Intel Core 2 Duo Processor E6300 1.86GHz 1066MHz FSB 2 MB L2 Cache
  Melt it into an ashtray maybe?
• StarTech ST100S Ethernet card
  Designed for Windows 98!
• TMS418169DZ-60 SIMM DRAM 2x8mb
  Show the ladies your VGA graphics in Ultima! Guaranteed panty dropper!

Also four CaseLogic flip books for storing CD/DVDs. 30x, 100x, and two 200x.

Are landlines relevant anymore? No. But I have a GE 26920GE2 cordless phone.

When using RichFaces’ <a4j:support> to reRender components in the view upon an event, I’ve come across an XML parsing error in the AJAX response which prevents the refresh from completing. The AJAX progress indicator icon continues into infinity despite the request completing its cycle, albeit in an error state. Firebug’s panel indicates the issue is “duplicate attribute”, where I track down the id=”” attribute as being added twice to at least one of the components being redrawn.

Swear I’ve been meaning to post more on this blog but every item is in a scattered draft form. It doesn’t help that last week my 1TB Seagate hard drive decided to stop powering up so I had to rebuild my computer while losing over a year’s worth of music along with 10 hours+ of Doom editing which was the tedious details-oriented sort. Otherwise, I tend to keep all my files online somewhere so the incident wasn’t as disastrous as it’s been in the past. But I’m going to try finding a circuit board to hopefully repair the hard drive because my meticulously organized music collection is a labour of love (and I cry myself to sleep at night.)

I’m still Tumblring and tweeting about.

Like last year, I tracked my plate appearances for my softball team.

Spring: Multitards, 7-1 for 1st
May 11 won 23-17, 7-7, 2B, 3B
May 18 won 21-8, 5-5, 2B, HR
May 25 won 33-14, didn’t play (sick)
June 1 won 28-13, didn’t play (TTC subway delay due to St. George Stn. police investigation)
June 8 lost 21-8, 2-5 (warning track FO with the bases juiced was disappointing)
June 15 won 15-14, didn’t play (Thrice concert!)
June 12 won 8-7, 1-2, SF
June 29 won 16-15, 5-5, 3B (pitched my teammate the hit that was the winning RBI in top 8th)
AVG 0.833 (20-24), SLG 1.208

I didn’t end up playing the summer session because games were Monday nights which interfered with my Aussie Rules Football team’s training at Cherry Beach. Then for autumn, our team (named Catch the Taste!) forfeited half our games due to lack of numbers and for the other half I was either out of town or too damned hungover to make it. Mid-afternoon Sunday leagues played in seasonally-deteriorating weather are just not a cool idea. But I did fill in a for a couple friends with spot starts:

June 13 for Kieno’s team, won 10-9, 2-3, 2B, GW 1B
June 17 for Zeff’s team, lost 15-9, 3-4, 2B
Aug 8 for Kieno’s team, lost 16-6, 1-3, BB

My commitment to hard hit grounders between SS/3B or liners to shallow outfield paid dividends in keeping my on-base percentage on the up-and-up. Here’s to hoping a free evening schedule in 2011 will allow me to keep this momentum going against my age’s grain.

Creating TestNG casesI was having issues creating proper data sets for our test suite since the data model is so complex. I was manually instantiation each class’ entity, populating the object, persisting, then committing the transaction manually. It became obscene for certain objects that referenced up to ten different classes with multiple constraints each. We already have SQL scripts to pre-populate our local JBoss environment so it was a possible solution to use that same data for our tests. I posed the question of a better solution on the interweb and it was suggested I try another DBMS that supports DB2 used by our site and also operates in-memory to increase speed. We already were using Hypersonic but it doesn’t support SQL import from DB2. So I tried H2, using this JDBC connect string to enable the DB2 compatibility mode and trigger our create + populate scripts on the test suite’s startup.

jdbc:h2:mem:localDB;MODE=DB2;INIT=RUNSCRIPT FROM ‘~/workspace/database/create.sql’\;RUNSCRIPT FROM ‘~/workspace/database/populate.sql’

As the H2 features page indicates, “for certain features, this database can emulate the behavior of specific databases. Not all features or differences of those databases are implemented.” Enter what I found out was in fact DB2 incompatibility mode.

org.h2.jdbc.JdbcSQLException: Syntax error in SQL statement

DB2 allows FROM as a column name. H2 does not. And no, I didn’t originally write this schema so don’t blame me for its poor naming.

org.h2.jdbc.JdbcSQLException: Column “CURRENT_ROOM.ROOM_NAME” not found

Aliases defined in an INNER JOIN clause cannot be referenced in the SELECT expression (isn’t that how Hibernate builds all its join queries?)

org.h2.jdbc.JdbcSQLException: Table “DATABASE” not found

Uh…

org.h2.jdbc.JdbcSQLException: Syntax error in SQL statement “…”; expected “AFTER”

Well it looks like a DB2 trigger doesn’t need that AFTER keyword.

So we’ve switched back to Hypersonic and I’m now creating an XML-based data set using DBUnit.

Since I prefer to dilute my online brand, you can now also follow me on Tumblr. Longer items that are less microbloggy will still be appearing on this domain.

For some reason Seam Framework doesn’t include form validation to ensure at least one checkbox is selected out of a group of checkboxes. Like a radio button, but more than one may be selected. One custom code sample was dodgy because validation takes place in JSF lifecycle before the model is updated, so you must wait for the component tree to be built before such a validation can take place.

You may use jQuery validation but what if the user has JavaScript disabled in their web browser? Or the JS payload hasn’t yet arrived? You still require backend assertion for the correct entity state.

The solution I found wasn’t optimal but hey, I’m not the only one wanting this simple form functionality.

Define a custom <mytaglib:atLeastOneValidator> tag

/resources/WEB-INF/web.xml

/resources/WEB-INF/compositions.taglib.xml

Create AtLeastOneValidator validator specific for that tag

/src/main/validator/AtLeastOneValidator.java

Test out <mytaglib:atLeastOneValidator>

/view/atLeastOneValidatorTest.xhtml

DEBUG [SequenceGenerator] Sequence identifier generated: 1025
DEBUG [AbstractBatcher] about to close PreparedStatement (open PreparedStatements: 1, globally: 1)
DEBUG [ConnectionManager] aggressively releasing JDBC connection
DEBUG [ConnectionManager] releasing JDBC connection [ (open PreparedStatements: 0, globally: 0) (open ResultSets: 0, globally: 0)]
DEBUG [AbstractSaveEventListener] generated identifier: 1025, using strategy: org.hibernate.id.SequenceGenerator
DEBUG [AbstractSaveEventListener] generated identifier: 1025, using strategy: org.hibernate.id.ForeignGenerator
DEBUG [AbstractFlushingEventListener] processing flush-time cascades
DEBUG [AbstractEntityManagerImpl] mark transaction for rollback

It’s a one-to-one foreign key generated by a DB2 sequence to persist multiple entities in one transaction. It seems to be created and assigned to the two entities mapped to each other but I have no clue why the JPA/Hibernate transaction decides it’s a no go as all the required fields are assigned in the objects being persisted. I’ve been working on this Seam Framework wizard for a week and this is an example of roadblocks in the dozens that have taken me to varying edges of sanity when it comes to troubleshooting this project.

I can only take solace in knowing once these niggles are straightened out, the rest of the system will be easy to implement. But the technological learning curve and unseen issues derived from requirement complexities make me believe Seam is a terrible framework solution. Absolutely terrible.

Update July 2: This error was occurring because a RunTimeException was thrown in the Hibernate library which wasn’t displaying in the server’s log or on the JBoss stack trace because the entrance method was invoked using EL. Wrapping my persist() call with a try {} containing a generic Exception catch revealed it was a constraint violation occurring on a value set as a Hibernate TrueFalseType (char(1) with ‘T’ or ‘F’ values) when it should’ve been a YesNoType (char(1) with ‘Y’ or ‘N’ values).

You can’t. But you can use <h:outputText> for anonymous visitors (read-only) and <rich:inplaceInput> for known logged-in users (write-enabled.) Customize the inplaceInputTest.xhtml s:hasRole() parameter to give the proper boolean value on whether the logged-in user is an administrator or not of the field bound to the input.

Define a custom tag to replace <rich:inplaceInput>

/resources/WEB-INF/web.xml

/resources/WEB-INF/compositions.taglib.xml

Below, evaluation of the rendered="" attribute decides which of the two elements appear at one time.

/view/compositions/inplaceInput.xhtml

Custom <mytaglib:inplaceInput> Usage

/view/inplaceInputTest.xhtml

Yeah, it’s all happening. Migrating a CMS to Java’s Seam framework using an existing DB2 schema/dataset, the Lucene search indexes on a few tables caused DB2 error SQL1584N. The bizarre part was it was only happening locally on my machine, but not for two others at work, who each have the same amount of RAM. It may have been because I was on Ubuntu and the other devs use Red Hat Enterprise Linux.

System temporary tablespace with page size of at least pagesize could not be found.
Explanation:
A system temporary table space was required to process the statement. There was no system temporary table space available that had a page size of pagesize or larger.

The statement cannot be processed.

User Response:
Create a system temporary table space with a page size of at least pagesize.

sqlcode: -1584

sqlstate: 57055

Following the instructions for ensuring system temporary table spaces page sizes meet requirements, I ran DB2 query:

…checking for the entry with Type = System managed space, Contents = System Temporary data, I confirmed the page size used was 32k, which is the highest possible with DB2.

After googling a bit, it seemed the solution was to increase the bufferpool size used by the tablespace giving me this error. The instructions for adjusting the bufferpool are to simply run DB2 query:

Confirm the bufferpool size, then increase the number of pages available until the Lucene indexing doesn’t shit the bed. e.g.,

During this process, you may get the error:

SQL20169W The buffer pool is not started.

sqlcode: +20169

sqlstate: 01654

This just means the change won’t take effect until DB2 is restarted, so execute:

Now, if you have the Seam observer Java annotation setup for Lucene indexing, this DB2 error will begone on the next deploy. And yes, this whole post makes me yearn for a non-Java project started from scratch.

When visiting the family last Christmas, I decided to avoid my iPod earbuds on the trip due to their deteriorated state (srsly, do they last more than three months without having their volume cut in half or the wire fraying?) Instead I took my Seinnhesier HD-595 for the plane/car ride for an obviously superior listening experience. I’ve done this before but this had an unexpected consequence: destruction of my iPod Touch’s audio plug.

On the plane’s return trip, the iPod would only play full stereo audio if pressure were applied to the headphone jack. But not left or right pressure like most jack problems, but pushing straight in. It seemed like an adjustment of the iPod’s internal jack would fix it easily enough. Problem was how could I easily open it? The manufacturing process tightly seals iPods to the point that they need a spudger (sharp plastic tool) to pop them open without causing aesthetic damage. One Amazon retailer won’t ship to Canada and I couldn’t find other online computer retailers carrying such items. My anti-eBay/Paypal policy prevented other channels.

So I just threw “fix iPod” onto my never-shrinking to-do list and held my iPod tight for use in the near future. With winter weather, that proved harder as the act left holes in my coat pockets along with the odd cramped hand (so business as usual – heyo, etc.) At work, I even devised a way to weigh down the earbud’s jack using a coffee mug. How these temporary solutions end up as three month-long habits are indicators of my lifelong love with procrastination.

Last week I decided to finally suck it up and fixed the damned thing or buy a new iPod already. I discovered it’s seemingly impossible to find a brick-and-mortar in Toronto that sells spudgers. The few College/Spadina stores I asked seemed offended that I would do such a repair myself without them to price gouge another witless customer. And the pretentious cunt as iRepair that just stood behind the counter staring at his Mac screen adjusting an iTunes playlist without even looking me in the eye while taking a few seconds between sentences while plainly repeating, “that’s how it’s done”… you can go fuck yourself.

Instead I found an eyeglass screwdriver in a kit given as a Christmas gift a few years ago. It ended up causing scratches after it took 10 minutes to pop open the iPod. There’s a decent iPhone jack repair tutorial available but bits smaller than those for eyeglasses work best for dismantling iPod components.

Adjusting the jack’s ground contact didn’t fix the problem, so I took off the circuit board where flash memory is soldered to get better access to secure the jack down. Since it was a two year old iPod Touch, I didn’t feel a need to be delicate in the process. Needless to say, a small circuit board piece broke off by the screw and after reassembly the iPod’s screen went to an iTunes & USB cable icon bootup loop that interrupted each iteration. And I couldn’t get to recovery mode by pressing home+power.

I basically bricked my iPod Touch because I didn’t bother buying the right tools.

Then I made the mistake of jokingly suggesting in a Facebook status update about getting a replacement iPhone, which spawned a 19 comments argument between three friends of Apple vs. Blackberry vs. Android. There’s still more than a year left of my terrible Canadian mobile phone contract; there will be no smartphone purchases. I’ll likely be a bitch by getting another iPod Touch that’ll require a $70 repair in another two years, simply because I like the WiFi+Safari+Facebook+Tweetdeck when I’m too lazy to get out of bed or away at a hotel.

Lessons learned? Get the right tools if you want to DIY it and don’t use huge indoor headphones for an iPod (big things, small holes).