[17:00:14] «Mode» Ken changes channel modes to moderated [17:01:23] nice & quiet :-) [17:01:31] «Voice» Ken gives voice to jmcmullan [17:01:36] Hello, this is Jason McMullan of the AROS Team [17:01:36] I'm one of the primary maintainers for the amiga-m68k port of AROS, [17:01:36] and have dabbled in other sections of AROS, with a goal of [17:01:38] improving the AmigaOS 3.1 compatability of the m68k port. [17:02:29] I await your questions. [17:03:00] «Topic» Ken sets topic to "Home of the Sunday AshaChat | No Religion, Politics, or Coding | /msg [operator nick] [question]" no quotes" [17:03:17] Will newer APIs beyond AmigaOS 3.1 be implemented ? [17:04:11] Some AmigaOS 3.9 APIs may be implemented, to support popular m68k programs, but in general AmigaOS 3.1 is the primary API goal for the AROS ABIv1 branch [17:04:54] After ABIv1, who knows. But I do not believe the AmigaOS 4.x Interfaces API will be supported by AROS in the near future. [17:04:56] next [17:05:04] Is AROS just an implementation of official Amiga API's or is it expanded with non-Amiga APIs ? [17:05:58] AROS is currently a combination of the base AmigaOS 3.1 APIs, with some MorphOS compatible APIs, and some of our own, where it makes coding easier. [17:06:25] But mostly AmigaOS 3.1 [17:06:26] next [17:06:37] Will AROS have native video support in the form of an API like quicktime ? [17:07:55] If you mean layered on top of graphics.library, sure, I don't see why someone couldn't develop a quicktime or Canvas style API. [17:08:51] But if you mean a new lower level instead of graphics.library, I don't see that happening without a significant amount of additional development effort (and more developers interested in implementing that feature) [17:08:57] next [17:09:03] Will there be an ARM version, as i hear efforts are made porting it to raspberry pi. And will it run on android tablets. [17:10:16] There are several ARM versions - ARM Hosted (on Linux), ARM Hosted (on Android), ARM on Pi, etc. It's all a matter of packaging , distribution, and maintenance now. [17:10:18] next [17:10:42] How many people are working on AROS. [17:11:20] I would say that there is no precise number. We all do this as a hobby - no one I know of if paid full-time to develop AROS. [17:11:48] But on average, I see a core group of about 5 to 8 people that make commits over the course of a week. [17:12:30] And that 'core' changes over time. For example, my commit rate was near zero for several months this year, due to other (paying) priorities. [17:12:33] next [17:12:40] Is there a development plan for the future, what still needs to be done ? [17:13:42] Ah, that assumes that AROS has an end-state we are trying to achieve. I don't think that's the case for me, at least. I find things that interest me, scratch that itch, make changes, and push them back. [17:14:14] I can't speak for the motivations of the other developers, but that's my drive. [17:14:18] next [17:14:21] Will AROS support systems with multiple CPUs ? [17:15:29] That is in the experimental phase right now. Michal Schulz and I have been toying with the question "What are the minimal changes needed to the AmigaOS 3.1 API to support SMP"? [17:16:13] The answer so far seems to be "few, but subtle". For example, SysBase->ThisTask is no longer meaningful on SMP, but FindTask(NULL) is. [17:17:09] Disable() and Forbid() are shockingly bad on performance, but adding a spinlock semaphore mode to SignalSemaphore will help new code on SMP. [17:17:40] It's still a toy experiment - heck, I can't even get it to boot 75% of the time - but it's fun playing with the API and learning about SMP. [17:17:42] next [17:17:51] Does aros have or will it have some emulation for ppc or 680x0 on x86 or ARM ? [17:19:43] For m68k, we have UAE-on-AROS, but I don't think that's the answer to your question. I believe you are not asking for a "run in a box" mode like the *UAE family, but a "run alongside" mode, similar to the m68k-on-PPC modes on MorhphOS and AmigaOS 4.x [17:21:30] In that case, the simple answer is no. The amount of effort needed to handle all the endianness issues that run-alongside m68k mode would need to keep track of for either a ARM or x86 little-endian AROS host.. well.. We'd need about 20 full-time developers to implement that in the course of a year, and about 10 maintainers to track changes in that for the rest of time. [17:21:42] So, tl;dr: no [17:21:43] next [17:21:44] Related perhaps will 68k apps support multi-core? [17:23:22] If someone delivers a m68k SMP system to port AROS onto (with MMU support, of course), I don't see why not. There would be a performance penalty when an app tries to directly access SysBase->ThisTask (and a few other fields), but it shouldn't affect most well-behaved apps. [17:23:32] next [17:24:01] Do you consider yourselfs competition or Hyperion, or is aros benificial for the community. Is AROS some sort of insurance if there will be no more official amiga development ? [17:26:09] I consider AROS to be a bit of a trailblazer. We don't compete with Hyperion on PPC (yes, we have a PPC port, but it is not our main focus and only lightly maintained), but we do provide the AmigaOS 3.1 API to CPUs that Hyperion does not sell AmigaOS 4.x on (i386, x86_64, ARM, m68k, etc) [17:26:50] So, not so much "Hyperion, follow us!" ,as going 90 degrees from Hyperion's path. [17:27:16] And again, we don't try to complete on the AmigaOS 4.x Interfaces API. [17:27:20] next [17:27:22] if you have any questions, /msg [17:27:36] Initially there was an AROS port to ACube's Sam. Will there be other ports to systems like the AmigaOne X1000? And updates to the Sam version? [17:28:44] I did the Sam 460ex port for ACube, but I received little user interest, and have since passed on my development platform to another AROS team member. So, I don't think there will be any Sam460ex updates, at least from me. [17:29:27] As for the AmigaOne X1000 platform, if AmigaOne would like to sponsor a bounty, and provide sufficient documentation for the X1000 chipsets, I'm sure someone would take up the challenge. [17:30:17] It may not be me, however. I do not feel that the PowerPC architecture is commercially viable in the long term (personal opinion here!) [17:30:19] next [17:30:32] Is it true that pieces of AROS code are used in the AmigaOS3.5 and 3.9 ? (If so - mod added) Was that because those pieces where in assembler or was source code missing. [17:32:01] If it is true, then someone in Hyperion would have to verify that. So long as they conform to our license, I don't have a problem with Hyperion or anyone else using AROS sources in their product. [17:33:14] We (recently) noticed that our default cdrom filesystem was encumbered by a licensing clause that prevented commercial use of that filesystem, and this last spring (on request of a AROS m68k commerical interest) I wrote an alternate CDROM filesystem driver that was not encumbered, but was under the standard AROS license. [17:33:22] next [17:33:32] Memory Protection is often wanted. How is the state and what is planned in regards on ABI1 [17:34:49] Full memory protection will require a great divergence from the AmigaOS 3.1 ABI, so that will have to wait until ABIv2. [17:36:05] That being said, we already have some ports (for example, Linux Hosted) that support faults on undefined memory regions, so that is a bit of a start. (NULL pointer dereferences can actually be caught on that port! Woo!) [17:36:12] next [17:36:25] Will AROS for the 460ex work on older Sam's like the 440ep? [17:37:17] I did try to make sure that the Sam460ex would work on the 440ep, but I did not have a 440ep for testing when I was working on that port, so I cannot guarantee that the 440ep functions with AROS. [17:37:40] next [17:37:52] And any time frame known when ABI1 will be finished ? [17:38:01] HA! [17:38:37] In all seriousness, I don't think ABIv1 will be "finished" until 2017, at the earliest. [17:38:50] next [17:39:11] Why does AROS need 256MB ram on (a >700Mhz) x86, isnt amiga os an os with small footprint. [17:41:05] Mainly because a lot of our distribution packagers like to use lots of truecolor icons and giant 24-bit wallpapers. The core OS itself (even Wanderer) runs on a 4Mb Amiga 1200, assuming planar icons and no background [17:41:18] next [17:41:27] Do you plan to attend this year's AmiWest and if so, what will you be presenting? [17:43:01] No, I won't be making it this year. The company I work for (a semiconductor design firm) is releasing a new network processor in the next year, and I am very busy with the preparations for that release. Let me put it this way - I'm hoping to be able to take off Christmas day. [17:43:03] next [17:43:21] Sort of a dummies question... what is the advantage of using AROS on the 68k over the stock OS 3.x? [17:43:48] Simple answer: None. No advantage at all on Amiga hardware. [17:44:48] Long answer: If you are making your own hardware (or FPGA), AROS is a very cost effective way of delivering a working operating system on your product without needing to sublicense AmigaOS ROMs. [17:44:52] next [17:45:00] When people want to develop for AROS 68k they use 68k targets (what often works but not always). Is a SDK or similar planned for AROS 68k? [17:46:28] Other than the recent icon.library issues (an incompatability I would like to address, mind you), one should always be able to use native m68k tools (like VBCC or SASC or DICE) to generate working m68k binraries for both AmigaOS 3.1 and AROS m68k [17:47:27] However, my personal build system just a Linux x86 laptop running cross-compilation tools generated by AROS's build system. [17:47:47] Easy enough to do, I can look up the URLs after the Q&A if people are interested [17:47:48] next [17:47:56] What would you like to see implemented in AROS? [17:49:28] ABIv1 completed, SMP (x86_64), SendMsg()/GetMsg() to support memory protection between target and destination [17:49:31] In that order. [17:49:33] next [17:49:49] we are out of questions, if you have any, /msg no waiting :-) [17:50:19] oops to soon [17:50:44] Are you looking for people to work on AROS ? [17:51:02] YES! ALWAYS! You don't have to ask for our permission, either. [17:51:04] http://www.evillabs.net/wiki/index.php/AROS_m68k-amiga [17:51:28] Checkout the git repos, hack away, and start yelling at us to merge your fixes! [17:52:35] The git repo on repo.or.cz is the 'official' one, but feel free to fork off my work tree on gitorious also: [17:52:38] https://gitorious.org/aros/aros [17:52:42] next [17:52:51] Have there been interest in AROS from people working on Amiga hardware (FPGA or other)? [17:54:08] Some. MikeJ of FPGAArcade wanted AROS as one of the 'ready to run' OSes for the FPGAArcade Replay (which I demoed at AmiWest 2012), and ACube bountied a port for the Sam460ex. [17:54:47] Again, get me hardware (once I have time again!) and a bounty, and I'd be happy to see what I can do if the project interests me. [17:54:48] next [17:54:56] Will there be 3D support in Aros 68k? [17:56:00] We have Mesa and Gallium, but can you point me to a 3D card for amiga-m68k that has full specifications and documentation on the 3D subsystem? [17:57:02] Of course, PCI is possible, but to be truthful, I think 3D on amiga-m68k would be more interesting as a FPGA extension (for, say, the DE2 or FPGAArcade devices) than a commercial PCI card [17:57:04] next [17:57:16] any more questions ? /msg [17:59:06] Seems to be a duplicate, but I'll post it if Jason doesn't mind Do you plan a *special* AROS hardware like the SAM or X1000 beside the normal PC-version? [18:00:43] No, we have no hardware team. AROS is a purely software effort. But if you want to build a special machine, and have AROS ported to it, it shouldn't be too difficult. [18:00:46] next [18:01:05] Do you think, a hybrid of x86/ARM and a FPGA with the traditional Amiga chipset would be the way to go (for AROS)? [18:02:59] It would be... interesting. But I don't think I could see where the market for that would be in today's world. It wouldn't have the m68k software library (Amiga games and applications), and it would have a terrible chipset (compared to an FPGA that simply had a USB2.0 controller, and a 1024x768 VGA 24 bit framebuffer) [18:03:05] next [18:03:24] At this time we announce "last call" for questions - /msg [18:05:13] I would like to ask Jason more about his idea to add Amiga chipset emu to Qemu, I realise it was just a tough experiment, but what would the [18:05:14] practicalities actually be for that to work? Would we be able to have unusual Virtuals AmigaOS with odd CPUs etc [18:05:49] kind of a mess [18:06:00] Yes, that should be possible. Qemu (as of today) has a relatively clean isolation between the CPU and hardware. [18:06:22] The main effort would be porting the WinUAE style hardware emulation to QEMU. [18:06:34] It's a Small Matter of Programming. [18:06:47] (but, not one that interests me at the moment) [18:06:48] next [18:07:27] It is after the hour time frame, so I think we'll close this down now. [18:07:36] «Quit» crasbe (~crasbe@p5B325B48.dip0.t-ipconnect.de) has signed off (Ping timeout). [18:07:37] Ok, it's been fun! [18:07:40] Our thanks to Jason McMullan and all the guest who attended! [18:07:55] * jmcmullan goes back to his paying gig... [18:07:55] Channel will return to unmoderated status in a sec. [18:08:01] <-- jmcmullan (~jmcmullan@c-98-219-185-181.hsd1.pa.comcast.net) has left this channel (Leaving). [18:08:05] Thanks, Jason for stopping by