Angus Thinks

Category: Macintosh

  • Running Oracle – The Solution

    This follows on from my previous posting. Searching for an Oracle

    This is kind of old news now, because it’s been in use since SQLEditor 1.2 was released. However I though I would explain how I solved the problem that I was having with installing Oracle.

    The problem was that I didn’t have a linux machine with enough memory and at the time, qemu wasn’t really fast enough to run Oracle in a virtual machine.

    The first strategy I used was to upgrade the memory in my linux box (Cetaganda) to 512 MB. This solved the memory problem and meant that the testing could continue.

    Then Parallels Workstation was released, which was able to run Oracle XE in a virtual machine fast enough for regular use. It also uses a different IP address than the host machine, which is useful in itself for testing.

    Parallels is definitely the answer to this problem, I don’t need to switch on another machine, there is no additional noise and the performance is excellent for my purposes. I’m even considering adding the loading of the VM to the unit test setup so it will load automatically before the test cycle starts.

    I’ve also got MySQL and Postgres running inside virtual machines.

  • Change console keyboard layout in debian

    I’ve been using Parallels Workstation virtualization a lot recently for testing databases and it works amazingly well.

    However I did come across a weird problem. I’d installed Debian linux and the console keyboard layout wasn’t correct for my Apple USB keyboard.
    The solution was to use this command

    dpkg-reconfigure console-data

    Which walked me through selecting the right keyboard layout.

    [link]

    Update

    Djamu posted a comment suggesting that the following command might work better for server installations:

    dpkg-reconfigure keyboard-configuration

  • Finding and fixing StackOverFlowError in java

    I was having a problem with some java one day, I’d made a change to some code and I kept getting a StackOverflowError. Unfortunately the stack trace didn’t help and some time inspecting the code didn’t offer any quick solutions. So I turned obviously to Google. The answer was simple, although it took some work to find it.

    If you get a StackOverflowError in java it can sometimes be difficult to identify where the problem is occurring. This is because in some versions of the JVM the relevant stack trace isn’t automatically provided. This is the case in 1.4.2 when using the JIT compiler. (which I was)

    One work around that worked for me was to add the -Xint option to the java call. This forces the JVM to use interpret the code rather than attempt to compile it. The interpreter will provide a stack trace and the problem can be easily found and fixed.

    The bug report that gave the answer is at Sun:
    Bug Report.

  • Searching for an Oracle

    Recently I’ve been trying to set up a box to run Oracle again. There are enough SQLEditor users that Oracle support is important.

    Fortunately Oracle have several downloads that can be used, there is a developer license for 10g and various chargeable options. However the option that is most interesting is Oracle Express. This is a cut down version of Oracle that has some limitations in maximum performance, but otherwise works the same as the regular version. For my purposes it’s perfect, since I’m not actually doing any data processing at all, maximum performance is irrelevant.

    So the next step was to download and install a copy.

    First I had to select windows or linux. Obviously a difficult choice, so I avoided it and downloaded both just in case.

    Next I looked around for a suitable machine to run this database on. Oracle 10g does run on macs, but no word of Intel mac support, so my Imac is out (directly at least).

    My next thought was a virtual machine on my Intel Imac ‘aslund’. Qemu runs windows really quite well and it runs linux perfectly well too. Unfortunately despite several hours playing around with settings I couldn’t get anything that would run fast enough and I couldn’t seem to get Oracle to run properly at all. I suspect that either I didn’t get one of the settings right, or there is some other problem somewhere.

    Next I thought about ‘cetaganda’, which is my windows box. This meets the minimum requirements of 256MB ram, and has both windows and linux. No worries there.

    Unfortunately it was debian linux and this requires Red Hat Enterprise Linux. After some thought I realized that Centos is a clone of Red Hat and so should work just as well. Which is probably would, if only my machine had enough memory. Unfortunately this machine was built to a (small) budget and has integrated graphics. The integrated graphics use memory from the main system for graphics, which reduces it from a nominal 256MB to only about 218 MB. 218MB isn’t enough for Oracle apparently and it complained.

    Next step, the windows download (lucky I got them both before).

    Windows XP sees the installer, unfortunately the same problem: not enough memory. (Although oddly the release notes mention this being a problem that has been fixed).

    Next I may consider my iBook ‘Komarr’, however that will be annoying, because when I tried it before, it was slow.

    The best plan may be to add more memory to cetaganda and run it that way.

  • Crash with WSMethodInvocationInvoke and malformed XML

    I’ve been doing some stuff recently with XMLRPC for both SQLEditor and an unannounced new product.

    One of the key parts of XMLRPC in cocoa is the Apple web services Core. It provides almost everything you need to use web services.

    However I did discover one oddity that led to some puzzelment

    As part of my development I’d written a test server in php which operates a few simple functions that the client can try. This means that I can simulate various commands and cause intentional failures to see what happens. I also have a php client to check the server independently.

    I had been making alterations to check a particular function in the client worked correctly. It was a new function that provided an array of values that could be placed into a menu on the client. The idea seemed sounds, the design had been worked out and new functionality had been added to both client and server. The server seemed to work fine with its test client and I was ready to try my main client application.

    Unfortunately it either crashed or generated an exception on every call to WSMethodInvocationInvoke. Initally I assumed that it was a bug with my client application, perhaps I’d passed in the wrong arguments or made a mistake in setting up the context for the call.

    However after some testing of various possibilities I discovered what seems to be the actual cause.

    One of the included php files contained an extra new line character at the top before the doctype.

    Removing this extra new line solved the problem
    (presumably because it was now valid XML).

    Moral of this story:

    XMLRPC implementations and XML parsers are not always very forgiving of whitespace at the top.

    Interestingly, the xml parser in camino doesn’t like xml files with extra lines at the top either.

    I’m still looking to see if there is some setting that can be changed to make the parser more forgiving, or to pre-process the xml first.

    (I’ll be reporting this issue as soon as I can produce a suitable test case)

  • PHP, trim and the non breaking space

    An odd thing occurred today, I was trying to do trim($string) on a string in php, but I kept getting strings back that appeared to have a space at the beginning, the exact thing that trim is supposed to remove. I carefully checked and the space was definitely found when I copied the result out to TextWrangler. Finally I tried substr($string,1) and then I got the answer, the new string started nbsp; the string I was trimming had  , a non-breaking space at the beginning, which trim doesn’t remove. A quick switch to str_replace and now my string doesn’t have a space at the beginning. 🙂

  • plutil

    STANDARDS
    The plutil command obeys no one’s rules but its own.

    from the plutil man page

  • SQLEditor built for Mac On Intel

    SQLEditor has now been built for Mac OS X/Intel as well as Mac OS X/ppc. I was able to download the Xcode 2.1 release last night and worked on the port this morning. SQLEditor at the moment still has bits of Java as well as Objective C so I wasn’t entirely confident of success, but I converted the build target to the xcode native type and moved the java code into a separate target. Then I ran the build process and after a bit of fiddling I was able to produce what appears to be a working fat binary:

    file SQLEditor
    SQLEditor: Mach-O fat file with 2 architectures
    SQLEditor (for architecture ppc): Mach-O executable ppc
    SQLEditor (for architecture i386): Mach-O executable i386

    The new version of the application seems to run as well as the old on PPC, although it is currently set up to use the 10.4 sdk as opposed to the 10.2.8 sdk of release SQLEditor versions. (Although SQLEditor is not officially supported on less than 10.3)

    If I had a Mac On Intel machine to test it on I would be very curious to find out what happens. 🙂

    But as far as I can see, and with the assumption that the ported version will not have unexpected bugs, the process is simple.

    But will that assumption will hold? I have absolutely no idea.

  • Mac OS X Intel

    It has been officially announced that the Mac platform will be moving to Intel processors.

    I’m still thinking about what this means for MalcolmHardie Software and SQLEditor and I have been reading about the transition to find out more.

    Some useful points culled from publicly available documention:

    • Mac OS X on Intel doesn’t use Open Firmware (at least not currently). Does this mean that it uses a conventional BIOS or something entirely different?
    • Rosetta (the translation environment) does not run Applications built for Mac OS 8 or 9. Does this mean that the Classic environment will run them, or are these applications dead?
    • Rosetta doesn’t handle code written specifically for AltiVec
    • Rosetta must run the whole application, you apparently can’t mix some bits of native code and other bits of emulated code, even with plugins. (I suspect inter-application communication might work here though)

  • Apples & Intel?

    Maybe it’s going to be hardware based dynamic recompilation?

    Or some kind of microcode based emulator?

    Modern microprocessors are usually doing conversion of the assembly anyway, so why not from one instruction set to another.

    I guess we’ll all know in a few hours …