|
|||||||||||||||||||||||||||||||||||||||||||
|
Current pace of AROS' development can be perceived on www.ohloh.net/p/aros. Though as can be seen here, the Russian developer Pavel "Sonic" Fedin is alone responsible for more than 3400 commits in eleven years, while the other four most important AROS contributors are inactive or almost since five years (apart of Georg Steger who was still active one year ago)... One can then see how much Pavel is important in current AROS evolution. ![]() I'm 32 and I have a wife. I like everything technical. My first interest from the early childhood was electrics and electronics. I have engineer degree (it's complete russian 5-years high education, between bachelor and master), in electronics design. However I work as a software developer, because there's actually no hardware design in Russia. ![]() I always liked alternative things. Don't know why. Perhaps because i like knowledge, and I like investigating. It's neither interesting nor exciting to deal with Windows PCs, because they are dedicated to trivial tasks like office word processing and Internet browsing. ![]() Wish to upgrade. ![]() No. ![]() No. Since PowerPC architecture actually ceased development, I lost interest. I still keep my Pegasos just because I feel sorry to drop it. FireWire is also dead by itself, superseded by USB. And I have no hardware. I had DV camera donated for the project, but it failed, and I was never able to repair it. ![]() When I first leaned about AROS, I liked the idea. First, I had a possibility to run Amiga-compatible OS on both machines (I had a Pentium I PC built from srap parts back then, to run MS Office). Second, I had m68k Mac laptop, and it seemed a nice idea to have a portable Amiga system. ![]() Of course. I added ShowViewPorts method to graphics driver, which gets a singly-linked list of bitmaps. This list reprsents z-ordered bitmaps to display. A bitmap also has XOffset and YOffset attributes, which a driver-side representation of ViewPort offsets. It's up to the driver to handle all this. A software implementation is a graphics.library plug-in. It intercepts three points:
![]() Yes. No time for now, but there are plans. For example I want to implement a framework for building a mirrored display driver. ![]() It's not in front. It's just the only bitmap. Display drivers that don't support hardware composition, have only one bitmap on display. ![]() It could, but:
![]() Yes, there could. One can try to expand the composition plug-in into subsystem allowing different representations. ![]() They are rendered in real time. ![]() I thought about this. However I lack time and knowledge. I have never been an artist, so graphics (including programming it) is completely alien to me. I very poorly understand GL. ![]() The main of benefits is reduced difference between kernels, and high percent of code reuse. This means that every change is well tested, since it's emerged into all platforms. Like new alerts and stack backtrace. ![]() Phase 1 is a small step towards all this, however, ACPI lets to:
![]() Yes it could. First of all it could be extended with exec-alike library which would allow to run dedicated tasks on secondary CPU cores (PowerUP-alike approach). This is much easier to implement than full SMP (because the new library is legacy-free), but still useful (for example video players and coders could support this). ![]() Yes, I can. Exec pools are already MP-ready. They are rewritten to use chunks of fixed size. There's no more 'thresh size', and puddle size is always a multiple of CPU page size. If some large block is allocated (larger than puddle size), an enlarged puddle is allocated. This is still a puddle, so its tail is reused for small allocations. This allows to use pools to break up memory pages (obtained from kernel's allocator) into variable-length blocks. AllocMem() with memory protection active is supposed to be just AllocPooled() from system default pool. Kernel's allocator exists in draft version and is not used yet. I have no time to test it well. Additionally, with current default pool implementation AllocMem() would be slow, because the pool would be mixed. The same pool would contain chunks with different flags. Pool manager supports this, but it will be slow. I am thinking on how to improve it. ![]() Don't know.:) ![]() Don't know, really. I lack time and phantasy... ![]() I have no relations with MorphOS devteam. And have seen only public (or leaked) code. If you remember, there was some code leak as a result of internal conflict in MorphOS team. I released some unofficial MorphOS updates back then, based on leaked sources. ![]() Yes, I will. Don't know when, but I want to. In fact I am using this version myself, to install newly built files to my laptop's native AROS partitions. I already know what to do with keypresses. I also sometimes experience freezes, but haven't found origin yet. ![]() The same as on any other platform. "configure" then "make". MinGW is needed. ![]() Yes, there is a single sort of ARM executables.:) ARMv6 CPU, VFP FPU, softfloat ABI. Other flavors are proposed (like ARMv7 + Neon hardfloat), but currently this is not in production. ![]() iOS is a very restrictive system. For example it's impossible to allocate a memory with read-write-execute access. It can be either read-execute or read-write. This way, memory protection support is mandatory for AROS to work on iOS. I came back to this port now, and implementation of display driver is in progress. I test it on simulator, which uses i386 architecture and is free from memory restrictions. When I'm done with this, i'll complete memory protection.
|