The Gnu Public License has reached version 3. Generally, everyone seems indifferent or, in Jem Matzan’s case, apathetic. (Links via hubertf) Everyone’s more interested in the Jesusphone. While we’re being mean about GPLv3… that’s one ugly logo.
Month: June 2007
Here’s two slightly tangential things that involve DragonFly: The first is a thread about large (64-bit) file support in Apache, and how it is treated as a special case because of poor planning under Linux – it’s not a problem in BSD. This led to an excellent quote from an excellent book, “The Cuckoo’s Egg“, by Clifford Stoll:
“We’re watching someone who’s never used Berkeley Unix.” He sucked in his breath and whispered, “A heathen.”
Also, “_why” posted a question about checkpointing to the users@ list, for an issue that Matthew Dillon later fixed. I recall that this _why is the same fellow who wrote “Why’s (Poignant) Guide to Ruby“, a programming book that is unlike any other.
Theo De Raadt’s description of bugs in recent Intel processors has made it to Slashdot, where in the comments, Matthew Dillon went through each bug and listed his opinion on each.Â (two comment entries, starting here)Â In contrast, Linus Torvalds’ general response was much more subdued.Â (Thanks, Wiger Van Houten, for links)
According to these reports from the OpenBSD-misc mailing list, Intel’s Core Duo is buggy, and upcoming features on Intel motherboards create a second running environment accessible even when your computer is off, both of which create security risks. (Thanks, Hasso Tepper)
Before anyone starts to hyperventilate, keep in mind: 1: this is a warning of potential problems, not an assessment of existing problems. 2: It’s an OpenBSD mailing list, which can be described as ‘adversarial’.
Sascha Wildner posted a patch that allows direct use of WPA and DHCP commands within ifconfig for wireless connections, along with some other changes.
It is indeed possible to watch every file access made on your system, everywhere, by attaching ktrace to the init process.Â It’ll generate a lot of data, though.
Here’s an entertaining possibility: Brian de Alwis, on the port-i386 NetBSD mailing list,Â noted that he was able to get ATI’s proprietary drivers for XFree86/Xorg working on NetBSD, at least for Xinerama support.Â Â I daresay this would work for DragonFly too – anyone have a recent ATI card to try out?Â (Original link from Greg Troxel)
Because my timing is awesome, I wrote an entry about Joerg Sonnenberger’s new pbulk system just as the computer hosting it went down. It’s back up, so, go back to the original entry if you missed the original output from that bulk build.
A recent post here on the Digest attacted a lot of comments – some trolling, some useful.Â Read at your leisure.
This week, UnixReview.com has a full arrangement of the usual topics: “Certification Changes and Updates“, a Regular Expressions column on “Python’s Mechanization“, a Shell Corner column continuing “Littera Delenda Est“, along with “Elements of Efficient and Secure Service Provisioning with Solaris“, and “Test Your Knowledge of Users and Groups“. Strangely, no book reviews this time.
Seen linked on Blue’s News: Kahvipapu articles on Linux gaming with first person shooters, and strategy games parts one and two.Â There’s more sections promised, and it covers some games I’ve never heard of.
I link it here because some subsection of these games run on DragonFly; they can be found in pkgsrc or may compile directly.Â DragonFly’s biggest hurdle for many games is the lack of 3D support through DRI.Â Now that we have modular xorg, it’s probably not too complex a project.Â Admittedly, I’d mostly be using it for fancy screensavers, but it’s still a project I’d like to see.
If someone wanted to fill a niche site need, there’s no site that exists for BSD games.Â Admittedly, it’s a subset of a relatively small audience, with a limited quantity of games, but that just means that such a site could be built with sheer willpower, rather than funding.Â Â Kind of like this one!
Joerg Sonnenberger has posted the results of his new pbulk system, for bulk builds of pkgsrc, on the email@example.com mailing list.Â Â His test bed is building DragonFly, so the results show just how many packages build on DragonFly.Â The report comes in text and graphical (warning: big!) versions too.
An interesting corollary to this, from reading the reports, is that out of 7,213 packages in pkgsrc, only 167 actually fail to build on DragonFly – that’s only 2% broken.Â There are other packages that fail due to dependencies on those broken packages, but it’s still a remarkably good percentage.
Chris Turner wrote up an interesting summary of what he’s seen in terms of the need for ‘realtime’ audio and how it’s been dealt with in the Linux world as well as BSD.Â There’s some mailing list links in there that can be used to eat up an hour or two of reading on a weekend…
Welcome Joe Talbott, who by this change appears to be our newest committer.
pkgsrc has a temporary freeze coming up, where only fixes will be committed in preparation for the 2007Q2 branch, for release this Saturday, June 16th.Â (No link, cause netbsd.org is apparently unreachable for me right now.)
Matthew Dillon asks that 1.9 users test using USB memory devices; he’s recently committed a large number of fixes related to physically removing mounted USB drives.Â Also, automatically mounting reconnected drives is a small, easy project enabled by this recent work.Â (See linked article for details.)
NATA, the ‘new ATA’ disk system, will be in the next release of DragonFly, but it will still be called ‘NATA’, not renamed to ‘ATA’.Â Keep this in mind when eventually updating with a custom kernel file.
Do you have a leaf.dragonflybsd.org account?Â Now is a good time to clean it up.
Matthew Dillon, while investigating a separate problem, ended up improving the separation between CPUs in a multiprocessor system.Â The Big Giant Lock is still there, but it’s a move in the right direction.
OpenSound was previously available for DragonFly, though support for it was quietly dropped probably around the same time 4Front stopped supporting FreeBSD 4. In any case, it is possible it could go into contrib/ now, if it has benefits – hopefully they will make it available under a more BSD-style license.
Network driver code has been shared between the BSDs a great deal lately, with a flowering of available drivers and support.Â Having a shared sound model too would also lead to benefits greater than the sum of its parts.
Matthew Dillon wrote a long message on how things are progressing with DragonFly; some projects like improved SMP support and 64-bit processing are almost ready for prime time, and just need someone to step up and complete them. The track record so far for DragonFly has been astoundingly stable; major changes in threading and process management have gone into the tree and it’s happened completely without destabilizing the system – e.g. it’s been safe even to run bleeding-edge code.
Also: the upcoming release will be 1.10, and hopefully GCC4.x can be made the default by the time 2.0 arrives.
From the DragonFly mailing lists: Matthew Dillon posted a list of what will and won’t be in the next release.Â Rahul Siddharthan pointed out that there hasn’t been much user-visible improvement since FreeBSD-4, speaking specifically about 64-bit processors and SMP.Â Steve O’Hara-Smith added some less well known benefits we already have, while Michael TalonÂ described the speed boost a 64-bit operating system gives.Â Matthew Dillon said “someone just needs to do it“.Â I daresay the conversation is not over.
Alexander Motin wrote to the kernel@ mailing list about netgraph in DragonFly; it was getting difficult for him to maintain compatibility with mpd4 in the FreeBSD-4 style netgraph in DragonFly and also support netgraph in FreeBSD.Â Matthew Dillon said “Don’t wait up“.Â Anyone feel like updating netgraph?Â It’s probably not easy.
Preview, the version of DragonFly that is not-quite-bleeding-edge, has been updated so that everyone can try the latest version of NATA and virtual kernels.