+ 6 - 5 | § ¶Planning MonoDevelop work
The next MonoDevelop release is going to be an interesting one. There have been lots of bug fixes since 0.7. One of the main issues was assembly conflicts when the loading of GTK# 1.0 in the MD process. Now all project references are loaded out-of-process, so this is not a problem any more. A sanity check has been added at MD startup, which ensures that MD and all addins are compiled with the same assembly versions. Code completion works also much better, and the documentation tool tip is now properly resized to fit the screen size, even for large method signatures. There will also be new interesting features, such as big improvements in MonoQuery and the NUnit addins. I hope all this makes MD 0.8 way more stable and useful than 0.7.I've been working lately in a reorganization of the MD assembly structure. The goal of this reorganization is to make MD more modular, by separating GUI code from non GUI code, and IDE specific code from non IDE specific code. Thanks to this division, MD can now really be seen as a development platform rather than just an IDE, and will have a much wider range of usage scenarios. For example, by installing the MD core it will be possible to build MD projects using a command line tool which doesn't have any IDE or even GTK dependency. I hope we can commit this reorganization shortly after the 0.8 release.
After this reorganization, my next focus will be improving the addin engine. I wrote something about this topic some time ago. One of the topics that is going to be controversial is whether we should put addins in the GAC or not. There are good reasons for doing it, as well as good reasons for not doing it, but we definitely need some way to ensure that addins are not a breeding ground for assembly conflicts.
I think this will be the last big API change before MD 1.0. I'm not saying that we are going to release MD 1.0 after this change, just that the API could be considered mostly stable at this point. Well, there will be one last thing to do: to audit the API and hide classes and methods that don't really need to be public. The MD API is a serious thing, since many (I hope) addins will depend on it, and we won't be able to introduce breaking changes at mid term.
So, I don't know when MD 1.0 is going to be released. It depends on which addins we want to ship in sync with it. It is important to notice that the release cycle of MD doesn't need to be in sync with the release of the addins, although we certainly need some stable addins for 1.0. I think a stable debugger addin is a must for 1.0. The ASP.NET addin looks feasible after the work being done by Michael Hutchinson. The addin for Stetic it's also important to have, but I don't know how stetic is coming along since the build requires GTK 2.6 and I'm still on 2.4.
Of course, we'll need to write documentation, improve the web site, design a better logo (please!), etc etc. Lots of funny work to do.
Holidays!
But nothing of all this will happen before my holidays. I'm leaving on September 1st and I'll be disconnected from the world until at least the 19th. Take care.
+ 2 - 5 | § ¶Namibia
I usually don't plan my holidays long beforehand, but this year I did, since there was a trip I was really interested in: a trip around Namibia, specially designed for people who like nature photography. The travel agency called me one week ago and told me that the expedition has been canceled due to lack of enough people. Sigh. So turns out that reserving in advance can be also useless.Since then, other people interested in the travel have been working on a B plan, which basically consists on renting a 4x4 and go in our own. We lack one more person to fill a car. So, if you have some 20 spare days in September, do you like photography and adventure travels, and want to share it with some Spanish and Portuguese guys, drop me an email *now*. It'll be lots of fun!