All entries of my technical and business blog

A couple months ago, I was out in Rochester MN at IBM for a multi-day meeting about communication between the iSeries(system-i, i5, as400, whatever its called now a days) and PHP/Apache. One of the things we talked about was our use of ODBC at (“the triangle”) currently to which they asked a good question - Are we using prepared statements over odbc? Well, right now, we’re not, but I think we should. As always, its up to me to show why we should be doing this. Let’s explore:

While working on some code at (“the triangle”), I run into some issues with the mail() function. On our linux qa and production servers, we can use the mail() function no problem - but on my development platform, windowsxp, I cannot with our current configuration.

For the setup I use at (“the triangle”), I have a lot of filesystem links - and these are made on win32 with the linkd.exe command (the version I got is from the windows 2000 resource cd). When I start a new project, I pass into my script which directories I’d like to make and checkout code into, and which I’d rather just linkd to. Well, every once in a while a link’d folder needs to be a real folder. So, since I’m a lazy programmer, I made a script called convertd which will unlink the folder and then make the folder. Chalk one up to efficiency by batch programming? Let’s see:

I use Subversion (SVN) for source control and deployment both for JEMDiary and at (“the triangle”). While working on my local copy of one of the websites, I got to thinking about the .svn folder and all of its files. The .svn folder is a local cache/db of the file changes in order to support diffs, reverts, and to give cues about file changes and the need to commit. I started poking around inside of the folder - and discovered the text-base folder. Inside of there, every one of my recently changed files were in there with an extension of .svn-base. Could this be a security issue - was I showing my code to the whole world? Let’s figure this out:

On one of the sites at (“the triangle”), one of the programmers on my team had this strange attachment to the Yahoo User Interface libraries - but I guess that was a good thing as it taught him a good lesson - use unobtrusive javascript.

Well, maybe that’s a little bit misleading - actually you’re running PHPdocumentor as an external tool.

I’ve found the need to do some installs from some PEAR channels, so I jumped into my command line on PHP 5.2.0 and windows xp sp2. I ran c:\php5.2\go-pear - which would launch the go-pear.bat file. I executed with standard permissions, and the unthinkable happened - a PEAR FAILURE. OH NO! But I was able to get around it. Let’s look at the details of the error, and my lazy man’s way of solving it:

Having been a shell, perl and batch scripter before I started PHP, I’ve always enjoyed the CLI life. I actually created an interactive or script-able interface for creating all of my projects at (“the triangle”) - hopefully I’ll link to it here when I write it up.

As you can probably tell from my last test of a firefox extension here, I am constantly looking for new tools to make my job more accurate (see: lazier). Another similar extension I came across is Tamper Data. Tamper Data doesn’t have a lot of documentation, so it was harder to actually determine the uses of this extension. Some tutorials on the net claimed that Tamper Data was better than Live Headers or Load Time Analyzer. So, I decided to try it out. Let’s see some screenshots and some features:

As you can probably remember, lately I’ve been writing about PHP’s object handling, patterns, etc. - but I lightly glossed over the whole forced typing of variables - and how that relates to PHP. While looking for solutions to this, I came across a great blog entry that solves this issue - but I’ve gotta bring up some counter points… Finally, I’m also wondering: what are the real important needs for a strongly typed object?