Press "Enter" to skip to content

Month: October 2001

Avoiding the Perl problem+Recruitment

[Note: this is an old blog entry which I used to have as a Slashdot journal entry. I thought I had lost it, but thanks to the Wayback machine, you can now see what I was blogging about all those many years ago.]

Entry date: Sunday October 21 @06:36PM [2001]

I found a work around for the Perl problem I reported on last week, it seems if you add a line similar to:

DBI->trace(3,”nul:ignore.txt”);

To your source code, DBI no longer causes your Windows 2000 machine to crash. If you increase or decrease the trace value though, expect the crashes to come back.

Only problem is, is that no body else seems to have encountered this problem.

Oh, Mitch Collis (the Operations Director at Cradley Print allowed me to hunt, find and assimilate (in other words – employ somebody) a ‘junior perl developer’ to help me as my workload has been booked up for several years already. I received around a dozen CVs from various agencies (Connex Recruitment [who have an annoying habit of sending outgoing Emails in MS Word format], Computer Futures, IT Executive and Forman Computer Staff). After applying my ‘criteria’ selection (staff between 18 and 30 with experience of Perl, HTML and similar technologies who are willing to work for less than 18,000UKP per year), I came out with 3 likelies. One of them, Tony Fox, was quite impressive on paper – despite only having a one-page CV. No other candidate supplied URLs, Tony supplied 8+, and his CV was ‘interesting’. He was my favourite candidate, but he was the second one to be interviewed.

Anyway, we (Mitch and myself) interviewed the first candidate (from Forman Computer Staff – although he was also available via Computer Futures) but found him a bit too ‘formal’ – plus he had well documentated and designed code with him – IMHO if people can afford to spend time documenting their code, they didn’t code to their full potential.

Tony came along and I was instantly impressed – Mitch came along and did his usual thing of offering Tony the job within a few minutes of meeting him (although I know Mitch was hanging outside the ‘Swallow Tail’ customer suite eavesdropping). However, Tony has a couple of other interviews to go to, so we’ve got to wait until Monday to hear. The third interview isn’t due until Monday afternoon, so we might have a ‘back-up’ candidate.

Personal life wise, although I’ve paid up to be listed on AmIHotOrNot and UDate.com, I still haven’t found anybody yet, but I’m still deciding whether or not to try and make it up with my ex (Gaynor). I went around to see her flat today (for the first time since our breakup) and it wasn’t bad. We got along quite well, but wouldn’t let anything ‘too serious’ happen between us (kiss on the cheek was the limit).

She’s not sure whether or not she is in love with somebody called Mark – I certainly hope she isn’t. It’s bad enough with both of my sisters going out with two seperate lads called Mark, but my ex as well. It’s going be a nightmare!

Anyway, it’s 00:37 and I’ve got to be up at 5:00 to get ready for work. ALLO!

Problems with production sections

[Note: this is an old blog entry which I used to have as a Slashdot journal entry. I thought I had lost it, but thanks to the Wayback machine, you can now see what I was blogging about all those many years ago.]

Entry date: Monday October 15 @06:05AM [2001]

Kak! I’m experiencing major problemos with the ‘section checking’ code for the production side of the Cradley Group Intranet. It seems as though I’m either hitting a problem with the Activestate Perl executable (v5.6.1 build 629, Aug 20 2001) or with DBI (1.14) or DBD::ODBC (0.28).

Basically, when I certain parts of the code run, Perl.exe throws up a Dr.Watson error on my Win2k Professional work’s laptop, and only half of the page displays on the browser. The problem is is that the script doesn’t send anything to the browser until it’s practically finished – therefore it would seem that the ‘output to browser’ code is buggy. Problem is, is that the same code is being used on Beebware.com, Adlive.co.uk, PDFCheck.com and the intranet site all without problems. Adlive/PDFCheck/Intranet are running on Win2k Advanced Server with Apache, Beebware.com is on a –nix box with Apache and my laptop (beebware.demon.co.uk if I’m online and disabled my personal firewall) is Win2k Professional with Apache. Therefore I don’t think it’s a ‘sending output’ problem.

I’m going to pull all the code that it could be running apart in the hope of narrowing the bug down – it’s going to be a lot of fun (coffee pot standing by – I think I’m going to need it)….

[Added]

Looks like it’s a reported bug with DBD::ODBC when you’ve executed a number of SQL statements. Activestate’s bug database has it as 15682: DBD::ODBC crashing perl on exit, 17417: Perl triggers Dr.Watson on exit with access violation msg, 16568: Access Violation using DBD::ODBC, 16493: Scripts using DBD Mysql or DBD ODBC crashing, 17439: DBD::ODBC crashes on exit and I’m sure theres another couple of reports in their database.

I’ve now had to add ‘cache-facilities’ to some of my ‘SQL-blocks’ to try and avoid the problem.
Ho hum – such is life.