gtkmm

I applied a bunch of patches that had accumulated and put out a new version. This one fixes the build on gcc 2.96, and is the first version of gtkmm to build with Microsoft's MSVC++, using the latest version of their compiler. It seems like the “.Net 2003” version still needs some extra workarounds, but it's far more standards-compliant than ever before. Sorry, Microsoft, too late, and still catching up.

julian also confirmed that gtkmm now builds without too much trouble on MacOs X. It just needs a little workaround for a gcc 3 compiler bug. SuSE 8.2 needs this too. It's already fixed in the latest gcc 3.3.1, I think, but Apple and SuSE are shipping strange versions.

Now we just need someone to deal with the simple Solaris/Forte problem that happened sometime – probably since 2.2.0.

Stalking Nat

I stopped for an evening in Munich on the way from Linz to Nuernberg. The day was getting pretty bad after being trapped on a delayed train with Salzburg-grade brainless yank tourists and germans who express their customer disatisfaction by sighing very loudly and then sighing very loudly again, in unison. Then I saw celebrity hacker Nat outside the Munich train station, conspicuously not being in Boston. I think he was on a secret mission. I confused him greatly by saying hello.

Hard Drive

I finally tracked down a new 2.5″ hard drive for my laptop. Austria is such a backwater sometimes. I’m starting the big reinstall. I’ll be working in Nuremberg from Tuesday to Saturday. They have a normal internet connection so I should be able to process the various patches that have built up.

I have to find a new short-term apartment in Linz at the end of the month. I’m losing my pretty wee wooden Haueschen. I shall miss reading in the hammock in the garden.

Ximian Connector

It doesn’t look like I can get the Ximian Connector to work with Infineon’s Exchange server because they have some stupid 2-stage login procedure. Why must large companies be so slow to do anything, yet somehow so quick to do the difficult stupid things? It’s a pity – I was impressed with the simplicity of the Connector’s setup process and I had a good feeling that it would work well on a normal network.

So now people must still endure my Outlook-generated non-threading emails and I must endure the pain of using it. It’s particularly annoying that the preview pane insists on downloading all the dodgy images from my junk email, thus confirming that my email address works.

Yikes

I killed my hard disk again. I do this often. Time for a reinstall.

JFDI

The ongoing GNOME website changes produced some unpleasantly unconstructive reactions. This finally got to me when I saw someone insisting that nobody should try to help. Although this was just a lot of bizarrely vicious nonsense from a small group of unusually bitter people, it does give me an excuse to make a couple of useful points.

Power

Like most open source projects there is nobody in GNOME who has any real power or who is able to make decisions just because of their position. Some names are more familiar than others and some peoples' opinions are more valued than others, because they are active or known for their good sense. They are still just opinions without force. And one characterisic of the best contributors is that they are ready to be persuaded by better arugments.

We have some official groups such as the GNOME foundation and the GNOME release team, but both are clearly subservient to the GNOME community. Anything else would be foolish because there is no way for them to force anybody to do anything. It turns out that you can get a lot more done when you accept that fact and work with it. I wrote some documentation for the release team recently that stresses that point.

So nobody should be blaming an imagined leadership instead of fixing what they care about. Apathy is so 80s.

JFDI

Because nobody has any real power, stuff only gets done when somebody decides to do it. Anybody is free to do stuff as long as they aren't complete muppets.

Some random examples:

  • jdub has been reworking the GNOME web sites. He happens to be on the foundation board and release team but he was the one who did this work because he was the one who did this work. It could have been organised a little better, but nobody made any great effort to do do that.
  • I write a lot of emails and poke at bugzilla, trying to encourage decisions and activity when it looks like nobody else is going to do it. (For instance, The 2.4 new modules threads on desktop-devel-list, the ui-review status, trying to get agreement on GTK+ 2.4 changes for the HIG). I don't pretend to be an expert but I don't need to be. Being on the release-team makes me feel pressure to do this kind of thing, but you don't need to be on the release-team to do it. Now that I'm in the swing of it, I'll probably continue to do this kind of thing even when I'm not on the release team.

  • People keep complaining about developer documentation. They do neither the easy things (submit bugzilla bugs when they find
    undocumented or badly-documented functions) or the difficult things (submit patches, or try to organise the whole API documentation effort). They should do something. If you have time to whine on gnomedesktop.org then you have time to help yourself by using bugzilla. If you think the problem is bigger then you should find out how you can improve it by posting on mailing lists.

  • There is a hell of a lot of activity in bugzilla – people are analysing problems and finding solutions. Lots of these intermediate contributions don't even show up in ChangeLogs. Recently I noticed that lots of people are making lots of simple but significant usability changes, making applications more HIG-compliant. (Hopefully this bugzilla query shows the HIG and usability bugs closed in the last month.) They are just getting on with it, because they care.

It can take a little while to get acclimatised but it's just a question of being polite and constructive. one of my first emails to a GNOME list shows that I took a while to learn. Get involved. Don't rant. Be concise.

Take a look at the How To Help page if you want to make a difference.

Another week of long hours fighting against ClearCase

Another week of long hours fighting against ClearCase for the right to read and edit our own source code. Some companies just can’t pay enough for awful products. Many CEOs and CTOs would pay fortunes to be kicked repeatedly in the gonads, and many do.

I’ll be in Nuernberg again for the first half of next week.

Blogging is better when you are as mad as an owl.

gtkmm

gcc 3.3 seems to have various bugs that are making life difficult for the Debian package maintainers. But I guess someone has to experience bugs first.

XP, schmechschpee

A while I ago I promised myself never to pay for any microsoft product ever again, what with the quality being so abysmal, but I have wanted to take a look at Windows XP for a while. I finally got hold of a copy and achieved blue screen in an as yet unprecedented five seconds after installation. The Windows XP installation is significantly less user-friendly than RedHat's, and it then leaves you with a fairly nonsensical desktop. After I fixed the bluescreenness (XP's NVidea driver) I then became outraged at how, every time you make an internet connection, it tells you that you need an MS passport account to do that. That's such a blatant anti-trust violation. Those bastards.

I discovered that you can restore the grub bootloader after Windows destroys it by first making a boot diskette (/sbin/mkbootdisk –device /dev/fd0) and then reinstalling grub (/usr/sbin/grub-install /dev/hda, I think) after booting from that disk.

gtkmm

I think I'm getting back on top of the bugs now. I don't like having too long a list of things to do.

Open Source Management

I wrote this little <a href=”https://www.murrayc.com/murray/subjective/management_open_source.shtml”>What Managers Can Learn From Open Source</a> document to make me feel better. That which inspires it most lately is getting fantastically worse.

gtkmm

A think I finally fixed the latest little gtkmm lifetime bug – an occasional leak. Harold Hopfes really dug deep into GTK+ and gtkmm with startling energy to help analyse the bug. He seems to be one of a team of gtkmm developers at EADS. As the bugzilla bug shows, we went backwards and forwards a few times trying to find the problem and solution.

orbitcpp

Another great hacker who seems to have exploded, fully-formed out of freak lightning strikes and nuclear waste, Bowie Owens is doing ever more fantastic work on orbitcpp, fixing all kinds of stuff that I should be doing for him. This guy seems to know his CORBA.

GNOME

I did a very superficial ui-review on gnome-meeting. I am suddenly rather concerned about its user interface but I’m sure that everything can be fixed very easily.

In Nuremberg

I’ll be in Nuremberg next week and apparently I’m to work 80 hours or so in 5 days. I expect to lose all sanity and hope by the weekend. There are murmurs of an open network connection there, but I don’t think I’ll have a chance to do any wortwhile stuff with it.

I have a new toy – a copy of Ximian‘s Connector so the GNOME lists don’t have to suffer the hopeless Outlook-ness of my emails, but I won’t be able to try it out until the week after next. Outlook (well, actually the hopeless corporate Exchange setup) is one of the few reasons that we use Windows desktops in this department, so I’m looking forward to showing that it’s not necessary.

The other day as I was sitting in a doorway with my laptop using an open wi-fi, a small child asked me if I was homeless.