June 2005 - Darwin-dev - lists.apple.com

Re: ICMP Router Discovery
by Justin C. Walker 28 Jun '05

28 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Jun 28, 2005, at 08:58 , Chase wrote: I'm trying to learn about ICMP Router Discovery... Works fine for me (try pinging something, as a test). Regards, Justin -- Justin C. Walker, Curmudgeon-At-Large Institute for General Semantics -------- It's not whether you win or lose... It's whether *I* win or lose. -------- _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… [This is a cross-post. It was originally posted on the macnetworkprog list, but it's been VERY quiet for a few days, so I'm hoping for a little more activity on this list.] But, this little code snippet I found in an online tutorial isn't catching any ICMP packets: [snip] It compiles and runs fine, but never catches anything. It just sits on the read() line, waiting indefinitely. I tried unplugging/replugging routers to force them to send ICMP router advertisements, but still nothing shows up. Are you sure that the router advertisements are being sent? Use 'tcpdump' to look at traffic on your subnet, if you haven't already. This email sent to site_archiver(a)lists.apple.com

1 0

ICMP Router Discovery
by Chase 28 Jun '05

28 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com I'm trying to learn about ICMP Router Discovery... int main() { char buffer[8192]; int fd = socket (PF_INET, SOCK_RAW, IPPROTO_ICMP); printf("\nHere we go...\n"); while (read (fd, buffer, 8192) > 0) { printf("Caught ICMP Packet\n"); } return 0; } Any ideas? Thanks. - Chase _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… [This is a cross-post. It was originally posted on the macnetworkprog list, but it's been VERY quiet for a few days, so I'm hoping for a little more activity on this list.] But, this little code snippet I found in an online tutorial isn't catching any ICMP packets: It compiles and runs fine, but never catches anything. It just sits on the read() line, waiting indefinitely. I tried unplugging/replugging routers to force them to send ICMP router advertisements, but still nothing shows up. This email sent to site_archiver(a)lists.apple.com

1 0

Re: we were opensource, once
by Andrew Gallatin 28 Jun '05

28 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Ernest Prabhakar writes: > Hello everyone, > > Michael wrote: > > But I think it could be so much better. > > For the record, I (and many others at Apple) agree that it _could_ be > "so much better." Unfortunately, the solutions aren't always as > simple as they might appear from the outside. > There are probably many lawyers involved, which tends to complicate things greatly. Even with all the limitations, publishing the majority of the xnu source, and the source for many of the IOKit families is very helpful to my work as a driver developer. Many times, the source is much more helpful that the documentation, and can go a long way towards making up for poor or missing documentation. My main wish is that Apple would open their development process a little more, and make the Darwin source available for development versions of MacOSX. Darwin source snapshots (and they wouldn't even have to be buildable) roughly corresponding to the various tiger seeds would have helped me a great deal. Also, the ability to discuss some of the problems I was encountering on the darwin lists would have probably saved Apple employee time, and definitely would have saved my time. Just to name one concrete example, when testing a piece of 64-bit code in a tiger seed, I discovered a problem in a 64-bit area of libSystem which was fixed by the release. (This actually happened 3 times for 2 different issues: Bug #3944194, #3973045 and #3994249.) To verify that it wasn't my code, I downloaded some example code and verified it crashed the same way, yet ran in 32-bit mode. I then opened a bugreporter ticket. Between debugging my specific code, finding the & building the example code, and filing the radar request, I wasted roughly 8-10 hours for 3944194, and longer for 3973045. An Apple tech support person had to look at each bug, and I was still left with the nagging feeling that it may be my own problem because developers generally get no feedback from bug reports. Had the Darwin 8.0 codebase been open and discussable on these lists, I'm sure I either would have seen a note that "foo doesn't work in 64-bit mode", or I could have gotten a quick answer to a "I'm seeing foo failing in 64-bit mode" query on darwin-dev@, or I'd have been able to look at the source to verify the bug was in the Darwin code and would have been able to make a more intelligent bug report. In any case, Apple and I would have both saved time and money. Cheers, Drew _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: we were opensource, once
by Markus Hitter 28 Jun '05

28 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Am 28.06.2005 um 04:02 schrieb Derick Centeno: I really find this discussion very, very odd. Cheers, Markus P.S.: Darwin and even Mac OS X Client suits as a server quite well ... - - - - - - - - - - - - - - - - - - - Dipl. Ing. Markus Hitter http://www.jump-ing.de/ _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… If anyone wants the wild roaming freedom of true open source why not participate as a professional in the Yellow Dog Linux support mailing lists, Board or elsewhere where Linux for the PowerPC is utilized heavily! Because any Linux on the planet is a completely different beast. How would you run a commercial Carbon or Cocoa app on top of YDL? Doing open source develoment isn't about getting rid of that plenty amounts of personal free time, it's about enhancing user experience. Apple, has made it's decision to change CPU's for many reasons. I don't see the relation between the openess of Darwin and Apple's current choice of processor. And I do view that Apple's position is very reasonable otherwise it's code would be as easy to crack as Microsoft and development would be more chaotic and harder to control; it would be also extremely difficult to maintain the quality item that the Mac OS actually is. Huh? Following this argumentation, Linux would be a complete mess and always open to intruders. As we all know, it isn't. Apple never had a problem to defend Darwin / Mac OS X against insecurity, poorly designed code and just ugly hackery. Sometimes even against featuritis. This one of the things I love about Darwin. I can switch to YDL and have a server on the laptop! This email sent to site_archiver(a)lists.apple.com

1 0

access to IOFrameBuffer
by john 28 Jun '05

28 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hi, Thanks! -- John _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… Is it possible to gain access to the IOFrameBuffer for the main display? What I'm needing to do is to be able to call the ::setTransform() method. Is there perhaps another way, or perhaps something equivalent? This email sent to site_archiver(a)lists.apple.com

1 0

Re: we were opensource, once
by Derick Centeno 28 Jun '05

28 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com I really find this discussion very, very odd. This as usual is for individuals to decide. On Jun 27, 2005, at 2:35 PM, Michael L Torrie wrote: Michael L Torrie wrote: Red Hat is just as guilty of hiding their dirty laundry. Take this for example: [PATCH] read() sometimes crashes in the second pthread_set_canceltype call (https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=130926) Tis true. However none of Apple's bug system is open at all to my account. RedHat does need to get their act together on this also. Apple is not the only guilty one, just more guilty. Have you contacted any redhat developers about this issue? Under the terms of the GPL, you have legal access to the patch that fixes this bug. You are forgetting that Apple, HP, IBM, SGI, Sun Microsystems, Intel, nVidia, ATI, VIA, Red Hat, Novell, etc. are all businesses that sell highly competitive products and will continue to do anything to stay in business and make money for themselves and shareholders -- even if it is under the guise of allegedly contributing to the open source community. I can't deny that some of these companies have contributed significiantly to the open source community -- some more than others -- but one shouldn't be surprised when they hold their cards close to their chest. I would imagine that Sun is going to face similar challenges with Open Solaris. Who's going to want to give freely to a project that is unlikely to give much back? Best wishes... _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… If anyone wants the wild roaming freedom of true open source why not participate as a professional in the Yellow Dog Linux support mailing lists, Board or elsewhere where Linux for the PowerPC is utilized heavily! The key advantage for participating in Yellow Dog Linux (www.yellowdoglinux.com) is that because it follows the Red Hat tree very tightly one is essentially running Red Hat on a Power PC! Keeping in mind that Apple is shifting to Intel, doesn't change the fact that YDL code will run without any change on IBM PowerPC Boxes, as well as other PowerPC systems! Apple, has made it's decision to change CPU's for many reasons. It also decided to handle what it considers as open source and what it considers proprietary in unique ways which it controls. If you want to see how different or similar things are with Sun I refer you to Openoffice.org and ask you to compare it to StarOffice. Or how Java runs as an open source compiler (available within YDL) vs. the proprietary Sun engine. Companies, have rights, akin to individuals to determine what they will share and what they will not. The licenses, Apple's included, are very clear what one may or may not do as a programmer/developer. And I do view that Apple's position is very reasonable otherwise it's code would be as easy to crack as Microsoft and development would be more chaotic and harder to control; it would be also extremely difficult to maintain the quality item that the Mac OS actually is. Regarding suggestions or recommendations to make to the Apple development team there are all sorts of means of communicating one's views; the issue is just how far is one willing to share or contribute. Just as anything else there are rules to becoming and being an Apple developer; there are expectations of trust, commitments and so on which protect each party. For my part, the fees were a bit stiff, but there really is nothing like a cash investment to keep someone's attention riveted to what is essential, as well as what one is doing or intends to do as a software product. Again with YDL one doesn't have to worry about any of this and one can jump in at any level of participation at any time. BUT as everything is voluntary by nearly everyone, chaos is very much a part of what goes on in Linux in all it's distributions. In many ways, Linux is it's own explanation of how a wheel gets reworked and improved again and again. The only difference, in Linux, is that the recognition one receives is not quite tied to the effort one actually contributes or participates with. Which is just as well, as I've found errors in coding various kernels and some drivers which I'm sure neither a freshman nor a junior programmer would have made but because the project is "free", the programmer is much less careful. This kind of thing happens A LOT. Human error is one thing, what I keep finding is sheer carelessness, literally a "lack of caring" about what is being coded. When it becomes time for Linus to retire or "give up the ghost", I doubt very much whether anyone of his survivors will be able to collect the value of his contributive efforts from the whole planet. I also doubt that whoever follows him will be as conscientious as he has proven himself to be. But this is the nature of open source; it is not the nature of AIX, Sun, Unix or even the Mac OS. We each choose where we "play" or "work" for "free", and when we expect to get "compensation" or "pay". This is the very thing that communists, socialists and the like have never quite understood. There is a great deal of difference between a contribution, a gift or a freebie or other intended benefit as opposed to the practice of theft, unwelcome intrusion or rape. This stuff is the very essence of comprehending the distinctions which individuals, (human or corporate) participate in producing as products for sale or compensation or contributions for society's good. The author Ayn Rand covered the fundamentals nicely in her work, "The Fountainhead"; valuable reading even today. Or one can see the movie with of the same title starring Gary Cooper. Her other works are also compelling covering these issues... On Mon, 2005-06-27 at 11:15 -0700, Gary wrote: This is a major 2.4 linux kernel patch that fixes a serious flaw in the POSIX threading library in RH Enterprise Linux 2.1 -- an OS that's been around for a few years now. Strangely enough, even when logged in to my Red Hat bugzilla account, I'm told that I am not authorized to access that particular bug ID. Wait a minute. YDL users/programmers are way beyond 2.6, isn't everyone else on the Red Hat tree? Why should anyone be dealing with a deprecated kernel? Not sure what you mean by that last statement. If Apple doesn't owe the open source community anything, then does the community owe apple anything? Ultimately, like I said, Apple is a cathedral, and I expect cathedral-like behavior from them. I had hoped, based on experiences with other OSS vendors, that it would be different. That was all I was really trying to say. Sun and Apple have contributed to open source communities by allowing their systems to be able to run open source products with either a recompile from source or a basic rebuild. Speaking for myself, I am still mystified that I can now run gnuplot, xemacs, gimp, openoffice.org, xine and MORE and STILL be within Mac OS X! What Apple did to xine in incorporating it (or the Goom plug-in or both) into iTunes is a wonderful example of how Apple embellishes open source... that is not a contribution to the open source community??? Or did you think that iTune's music visualization engine was all Apple's doing? Just what do you have in mind as a contribution to the open source community, if you believe that this is not good enough???? I can use Xcode or not or I can switch to YDL and have a server on the laptop! And although Apple is abandoning the PowerPC chip, because of YDL I can run on a PowerPC for as long as the run of my natural human life (which if I'm really lucky is projected to be another 40 years if I don't suffer from Alzheimer's or anything else first) AND have that code run on any PowerPC from IBM or anywhere else. If anyone here is looking to do PowerPC stuff still, YDL is very much around and active as a path forward which seems to me to be really good news for all those Apple PowerPC clusters which various universities use. Implementing YDL into Unix or other networks or research or work environments is not the problem; gathering the expertise to weed out all the careless errors out there however is. And like it or not, there is a limit to how much anyone will work or endure for "free" or even for "love". Sooner or later it comes back to paying for expertise as opposed to the current trend in how to avoid paying for expertise. This email sent to site_archiver(a)lists.apple.com

1 0

Re: we were opensource, once
by Ernest Prabhakar 27 Jun '05

27 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: we were opensource, once
by Michael L Torrie 27 Jun '05

27 Jun '05

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com On Mon, 2005-06-27 at 18:43 -0400, Jake Luck wrote: > not true, see: > > http://marc.theaimsgroup.com/?l=osx-darwin-kernel&m=111774187918906&w=2 We are speaking of the Linux kernel, and specifically the RedHat customizations they made to the linux kernel. RedHat does release their source code completely, but the bugzilla entry was hidden and no one knows why. This thought is somewhat tangental to the darwin conversation. Michael > _______________________________________________ > Do not post admin requests to the list. They will be ignored. > Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) > Help/Unsubscribe/Update your Subscription: > http://lists.apple.com/mailman/options/darwin-dev/torriem%40chem.byu.edu > > This email sent to torriem(a)chem.byu.edu -- _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: we were opensource, once
by Jake Luck 27 Jun '05

27 Jun '05

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com On Mon, 27 Jun 2005, Gary <gary(a)wlfnet.org> wrote: not true, see: http://marc.theaimsgroup.com/?l=osx-darwin-kernel&m=111774187918906&w=2 _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… I can download their kernel source for any major releases they make. So that's not the issue, really. I am curious, however, as to why they'd bother to obfuscate it in bugzilla. This email sent to site_archiver(a)lists.apple.com

1 0

Re: we were opensource, once
by Michael L Torrie 27 Jun '05

27 Jun '05

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com On Mon, 2005-06-27 at 23:26 +0200, Markus Hitter wrote: > There's nothing stopping the Darwin community to develop such things. > But how would you build a strong community around a closed, deaf and > mute cathedral? > Precisely. You can't. This is one of the reasons that we are switching back to RHEL from OS X server for our file servers (now that AFS support under plain linux is much improved). I'm sure I'll be flamed for this opinion, but I've had 2 years of sad experience that has soured me. > > For me, Darwin is read-only source and OpenDarwin is fine for > building add-ons at the CLI level: XPostFacto, Darwine, ... you name > it ... Exactly. And if that is what Apple intends, so be it. But I think it could be so much better. Michael > > > > Have fun, > Markus > > - - - - - - - - - - - - - - - - - - - > Dipl. Ing. Markus Hitter > http://www.jump-ing.de/ > > > -- _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0