Towards an Alpha Release
May 22, 2008 09:48 UTC, by Chris Simmons, Senior Journalist
From the smooth-cycles department...

The Haiku Project is making great strides these days, with work done this past weekend paving the way for a sooner than expected Alpha 1 Release. Stephan published a fairly extensive report about recent bug fixes, and the development work that makes an Alpha 1 release almost a reality.
 
Not satisfied with his report, we naturally asked Stephan to clarify a few points, both to serve our own curiosity and to satiate our readerships' personal questions too.
 
 
Hi Chris,
 
Will including GCC 4.x make the road to 64bit any easier?
 
Unfortunately, no. 64 bit will be a _lot_ of work. I know that some people in the community are currently very focused on that, but they don't realize that 64 bit does not give them as many immediate benefits as they are hoping for. In other words, 64 bit would be much more of a buzzword than something useful for Haiku at this point. For example, the 64 bit mode has more registers, so some code can be compiled to be more efficient when calculating something, because there are more registers that can hold intermediate results. On the other hand, it means that more memory needs to be stored when saving the state of a thread and restored when switching to another one, so there is probably a slight performance drop again. 64 bit allows you to address much more memory than 32 bit. But until very recently, I was using BeOS with 512 MB of memory, now I have 1 GB.
 
For Haiku, these amounts of memory are plentyful. Sure, it will be nice to use all of 4 GB or even 8 GB if Haiku was 64 bit. Maybe the OS can use it for caching or something. But I don't see any big applications at the moment that would make use of it. All I am trying to say is that this feature is much less important _right now_ compared to how much attention it gets from some of the community at the moment. We have a lot of stuff to do and we need to make careful decisions on what we want to spend our very limited resources on. These decisions do need to balance technical merrit and buzzword factor. So while I agree that it will get more and more important, it isn't right now.
 
Will the specific flag for GCC 4.x be turned off by default?
 
At the moment the GCC 4 build has problems that the GCC 2 build does not have. I guess it depends on whether we can fix those problems first. For example, some applications crash when they quit, while they work fine on GCC 2 builds.
 
What specifically can testers focus on at this point in time that would be helpful to a smoother Alpha release?
 
Just playing around in Haiku and running various applications on it and reporting bugs if they find any, of course making sure that the bug isn't reported already. In case it is, they may be able to add some new insights. Trying to reproduce old bugs and reporting on them in case they are no longer valid would help, or in case they find more reliable ways to reproducing them is all a great help.
 
 
So there you have it, straight from one of the developers close to the matter. Personally, I agree with Stephans sentiment regarding memory, resource usage, and related. BeOS was always very lightweight when it came to resource management, in that even with 64MB of ram back in the days of Pentium I, it was fast. It was nimble. Haiku will be no different; in fact it will likely run better in comparison to BeOS on the same hardware, once Haiku R1 is released. You can be assured we will put them head-to-head when the time comes. Should be fun.
 
So, until then, the moral of this story is, be patient, let the developers concentrate their best efforts on filling in the remaining gaps and chunks of code remaining, pitch in what you can in terms of bug reporting, and hang on for the ride!