Book: Working Effectively with Legacy Code

Working Effectively with Legacy Code by Michael Feathers
ISBN: 0-13-11705-2

"...legacy code is simply code without tests... Code without tests is bad code."
From those statements, it doesn't take much to figure out what this book is about - how to write unit tests for code without tests. Of course, if you've ever tried to do that, you know that it's easier said than done. Many/most programmers who inherit a big ball of mud of code without tests (legacy code) just punt; the existing code has no tests, I can't see how to get any of it under test, so I'll just hack and pray - just like the original author(s) did.

That's where this book comes in. It's a primarily large collection of recipes about how to write unit tests for legacy code. That said, the focus is not really on how to write the tests but rather how to get chunks of the legacy code into a test harness so that you can write unit tests to characterize the existing functionality before adding or modifying functionality. It also contains techniques to add new functionality in such a way that you can test it immediately and possibly execute the "clean up" that the original author(s) promised would happen as soon as that next deadline was reached - all those years and deadlines ago.

The bulk of the book (Part 2) is organized as a series of complaints or excuses and how to deal with them. These include such topics as "My application has no structure," "Dependencies on libraries are killing me," and "This class is too big, and I don't want it to get any bigger." In each chapter, the author provides examples (in multiple languages - Java, C++, C, etc.) of these problems and specific techniques that can be used to address them. The last chapter (Part 3) is an encyclopedia of the techniques for easy reference.

If you're lucky enough to do only green-field development, you might think this book would be useless. However, one interpretation of this books is that it is a list of sins to avoid while your playing in the green field. And, many of the techniques can be interpreted as best practices for how to write you code to ensure it's testable. (Of course, you're following test driven development and achieving near 100% coverage, so that would never be a problem with your new code, would it? :-)

My one disappointment with this book was that I was hoping it would provide ideas about how to create higher-level (e.g., functional) tests. Of course, high-level tests are no substitute for unit tests. It's just that I was tasked with creating "some tests" quickly for an entire application, and unit tests are not practical in this particular case, which is my problem, not the author's.

This book is an excellent resource and cookbook for how to add unit tests to an existing code base that lacks tests, and it also provides design and implementation templates to ensure that new code is testable as it's created.


enjoy,
Charles.

vmware-cmd: SystemError=HASH(0x95d8d70)

More tales from the dark side. Trying to start a VM out on an ESXi host:

vmware-cmd  [conn options]  '[datastore] path-to-vm/conf.vmx' start hard

Returns this output:

Fault string: A general system error occurred: Internal error
Fault detail: SystemError=HASH(0x95d8d70)

I spent hours permuting the parameters. And vmware-cmd's -v option doesn't dump out all of the SOAP guts, so I had no view on what was going on. Finally I ran a vifs --dir out on the VM dir and found a bunch of vmware.log files. I fetched one that showed a normal startup and then this:

Nov 05 05:51:51.883: vmx| [msg.License.product.expired] This product has expired.
Nov 05 05:51:51.883: vmx| Be sure that your host machine's date and time are set correctly.

There was a botched version of ESXi that we installed that contained a time bomb, and it had come home to roost. Fair enough - our bad, but certainly there could be a better error message than SystemError=HASH(0x95d8d70).


Charles.

Vmware ESXi - Unable to clone virtual disk

To import a VM from VMware Server to ESXi one must convert the disk by cloning the old one using vmkfstools. Fair enough. Doing this using the ESXi RCLI tools should look something like (based on what I've seen for non-RCLI for ESX):
vmkfstools [conn-options] -i old-disk new-disk -d thin

However, when you do this with RCLI, you get the following useless error message:
Unable to clone virtual disk : A general system error occurred: Internal error

Using the --verbose flag to look at the gory SOAP details I could see that the thin option wasn't getting sent. Looking through the Perl code it looks like one needs to specify both the -d and -a (adapter type) options. Once I added -a lsilogic it worked like a charm.

My big complaint (and I have other examples of this up my sleeve) is if the user makes a simple, bonehead parameter error, the command should point it out rather than saying "general system error...internal error." I hope this post will help someone else out if s/he is unlucky enough to encounter this error message.

Charles.

WebTest Groovy example thows NoSuchMethodError

I've been working on a project to do functional testing on a web application using WebTest. I've coded up some prototypes using the standard XML/Ant format/language/DSL, but I wanted to try doing it in Groovy, since I'm not a big fan of XML. I copied the sample application from the manual, but when I ran it I was getting and exception:

Caught: : java.lang.NoSuchMethodError: org.apache.xpath.compiler.FunctionTable.installFunction(Ljava/lang/String;Ljava/lang/Class;)I
at test.run(test.groovy:33)
at test.main(test.groovy)

(If you're following along at home, you'll note that that line number is too large for the example - I've already hacked some stuff into it.)

I couldn't find any examples online of people having exactly the same problem, but I'd seen other errors blaming an old version of xalan.jar for this problem, especially under Mac OS 10.5 - my platform. (BTW, this happened under both Java5 and Java6.) Although I didn't have CLASSPATH set to anything, I figured I'd try setting it to have the version of xalan.jar that ships with WebTest, and it worked. So, my complete invocation line is:

groovy -cp /usr/local/java/webtest_1720/lib/xalan-2.7.0.jar -Dwebtest.home=/usr/local/java/webtest_1720 test.groovy

enjoy,
Charles.

VMware ESXi default resource pool name

I've been playing around with ESXi trying to figure out how to use their Remote Command Line Interface (RCLI) to import and run a VM. This has been a major PITA, which I suppose I could have been documenting as I go but I didn't. (Full disclosure - I'm doing this without any formal training, without their snazzy Virtual Infrastructure tools, or anything else.)

According to VMware's RCLI Installation and Reference Guide, the data center and resource pool parameters to vmware-cmd -s register are optional, but nonetheless, it kept telling me "Must specify resource pool".

Eventually, I found a forum thread that contained the answer: Resources. (Kinda like "plastics" only different.) I don't have a data center set up, but I found that using almost any word for the data center name seemed to work.

So, in the end, this is what worked for me (your milage may vary):

vmware-cmd  [conn options] -s register '[datastore-name] vm-dir/conf.vmx' root Resources

enjoy,
Charles.

Increasing Memory for Ant

I had an issue today with the javac Ant task running out of memory when compiling a project (~800 files) under Java 6 on the Mac. It was fine under the default Java 5 compiler on the Mac, and it was fine under Java 6 on the PC. Initially, I set memoryMaximumSize on the javac task, which also required that fork be set.
This was OK, but since this was specific to compiling on the Mac, I figured the solution should be specific to my environment on the Mac. After looking at the source of the ant script, I realized that the ANT_OPTS variable could be set to contain an option (-Xmx512m) for the JRE running Ant, and that ANT_OPTS could be set from either ~/.antrc or ~/.ant/ant.conf. So, I created ~/.antrc with one line: ANT_OPTS=-DXmx512m.
(Update: oops, that should be ANT_OPTS=-Xmx512m)
And it worked perfectly, of course - so simple it had to work.

enjoy,
Charles.

Ant, JavaDoc, and CreateProcess error=87

I've been developing an Ant build file for a large-ish base of existing code. I've been using my Mac as the primary devleopment platform, but the ultimate target (for the other developers) is Windows. My task to build the javadocs runs fine on OS X, but the first time I rant it on Windows, I got the following:

c:\demo\build.xml:180: Javadoc failed: java.io.IOException: Cannot run program "c:\Program Files\Java\jdk1.6.0\bin\javadoc.exe": CreateProcess error=87, The parameter is incorrect

After about 10 minutes of Googling, I found a/the simple solution: add the useexternalfile attribute to the javadoc task in Ant. Voila , back in business, on the PC at least, and it didn't seem to break the build on the Mac.

Charles.

Book: Essential SQLAlchemy

Essential SQLAlchemy by Rick Copeland
O'Reilly Media
ISBN 10: 0-596-51614-2 / ISBN 13: 9780596516147

This is a great book describing how to use SQLAlchemy to connect Python programs to databases. In fact, at the moment (mid-summer 2008), it is the book, since there are no other books on the subject, yet. Athough I am not (yet) a SQLAlchemy user, this book seems to cover all of the core topics in SQLAlchemy. The text includes many straightforward examples of how to use various facilities in SQLAlchemy and how to map various database programming problems into Python code via SQLAlchemy. Copeland also provides a whirlwind tour of some extensions to SQLAlchemy.

I heard about SQLAlchemy project on the This Week in Django podcast. Django doesn’t use SQLAlchmey, but it does use a similar object-relational mapper (ORM). As I mentioned, I haven’t used SQLAlchemy so I came into this book with a somewhat blank slate. I have, however, been programming in Python since before 1.0, and I’ve worked with database APIs and ORMs since the early 90s in C++, Java, and Python. So, I was familiar with the basic landscape of database programming, even if I hadn’t used SQLAlchemy. And, I’m currently working on a large Python project that is coded using the Python database API directly, which is very tedious. So, the whole time I was reading this book, I was looking at how to fit SQLAlchemy into this existing code base.

To be honest, the first chapter (the proverbial introduction) almost turned me off. The author starts out slowly enough, but then he starts touching on a huge number details, which were glazing my eyes over. However, the second chapter (getting started) started back at ground zero and stepped through everything in a nice clear fashion, and the rest of the book continued in that vein. He covers all the topics you would expect in a database programming book: queries, updates, joins, the built-in types, and how to hook in to provide support for your own types.

Something I didn’t realize about SQLAlchemy coming into this is that SQLAlchemy is both an ORM (what I expected) as well as a high-level, database-independent API. Which is to say, you can just access the database as tables, columns and rows rather than as classes, attributes, and object instances. Although I’d personally prefer to use the ORM, I can imagine cases where it might not be the right tool for the job, and it’s good to have a choice.

I was also surprised to see the ORM supports two styles of object-relational access: the data mapper pattern (which I had seen in Django and Hibernate) and the active record (used in Ruby). The author does a good job of explaining both of these and how to use them. He even devotes a whole chapter to Exlir, which is an extension that implements the active record pattern.

One thing that many people might consider odd is the fact that although SQLAlchemy is an ORM, the author waits until chapter eight to discuss how to map object inheritance hierarchies onto relational databases. Most books I’ve read on ORMs discuss this topic early, but I applaud Copeland’s decisions to hold off on discussing it. When books bring this up early (e.g., in chapter three), the discussion often gets bogged down in details, which glaze the reader’s eyes. I’ve dealt with the issue of inheritance mapping enough in ORMs I’ve used and those I’ve written enough that I’m not that interested in the topic (assuming the tool provides the typical, reasonable solutions), and was grateful that he held off on it.

One issue I had with the overall structure of the book is that I’m hard pressed to pigeon-hole the book. Books about a single technology such as SQLAlchemy usually occupy one end or the other of a spectrum. Either they’re hard-core references, often times copying-and-pasting API documentation from a web site (I really hate that), or they’re largish tutorials that may or may not contain enough technical meat. This is a short (roughly 200 page) book that contains plenty of technical meat, but it also includes some simple tutorial motivations for using various capabilities of the tool. Although this mix felt odd to me, I’m sure it will be perfect when I go to apply SQLAlchemy to my existing database projects since I don’t really want a hand-holding tutorial, but a pure reference wouldn’t quite work for me either when I’m just starting out.

In conclusion, Essential SQLAlchemy provides a thorough presentation of the SQLAlchemy tool for interfacing Python code to SQL databases. The author covers a number of different methods in which SQLAlchemy can be used to access databases from Python, and he provides plenty of details of the various APIs available to the programmer.

Enjoy,
Charles.

Converting a Django Site to newforms-admin

Right after I learned about Django at OSCON 2006, I created a prototype to replace an existing PHP site. Due to various distractions, I haven't done much with that Django site since then and it never went live, but I've come back to it recently.

My first order of business was to upgrade from 0.95 to something modern. The first stop was to upgrade to trunk. Following the instructions on the Django site, I got it running on trunk and only had to change maxlength to max_length in my models.

Since, newforms-admin will be merging with trunk soon, porting to newforms-admin was my next step. I pointed my django install at my newforms admin svn tree and let it rip - i.e., I didn't bother reading any documentation or anything.
It blew up with a nasty-looking traceback that ended with this:

func(self, *args, **kwargs)
TypeError: __init__() got an unexpected keyword argument 'prepopulate_from'

A quick Google search turned up this thread on the Django Users mailing list. Five minutes of reading the newforms-admin wiki page and refactoring my code, and I was up and running. So, despite the ugly-looking traceback, it just wan't that bad.

Granted, my project is pretty small because I never had much time to work on it. However, all in all, the conversion process from 0.95 to the bleeding edge of newforms-admin was pretty painless.

Another happy Django user,
Charles.

Book: Effective Java

Effective Java, 2nd Edition by Joshua Bloch. ISBN 0321356683

This is the best book that I've read that I didn't know I needed to read. If you are a pretty good Java programmer, and you want to be better, this is a book you should definitely read.

Back in the day (BITD), we had Scott Meyer's book Effective C++, and we needed it. Even without templates and all the "new" things in C++, C++ was a large and complex language. I loved Effective C++ not only because it had lots of tips to keep you out of trouble when coding C++, but also because the organization was brilliant: each tip was about the right length to be read while sitting on the can.

I ignored the first edition of Effective Java book when it came out because I smugly assumed that Java was so superior to C++ (Java is C++ with the pointy bits filed down) that it wasn't really necessary. And that may or may not have been true in the 1.0 or 1.1 days of Java, but with the release of Java 5, the language has certainly grown and is now sufficiently complex that a book like this is a necessity, especially if you've recently moved up to Java 5 or Java 6. Lucky for me, this book landed on my desk against my will.

The book uses that same sitting-on-the-can format from Effective C++; the information is broken down into nice, small bits of information that you can read one or two at a time. These are grouped into 10 sections. There are specific sections for Java 5 features like Generics and Enumerations, which is great for people like me who were stuck on 1.4 for far too long.

What is really amazing about this book is that I learned a lot about topics that I thought I already knew about. For example, I've used serialization in a number of formats (built-in and do-it-yourself) for years, but I realized there's a lot more that I had no idea about. I will certainly think twice before doing any serialization in the future.

Unless you're the sort that knows the Java Language Specification by heart, you need to read this book.

Enjoy,
Charles.