‘Walter’ asked about dealing with endless ssh port dictionary attacks. Several suggested changing ports and removing passwords, with opie(4) for when keyfiles aren’t available. Also, there’s various auto-blocking scripts, though they aren’t as useful nowadays.
Alex Hornung has taken on a very overdue and very necessary project: an update of linux binary support. His code is available for anyone who wants to try it. Testing so far is working, but it could really use something complex, like Java with OpenOffice or tomcat, or perhaps Firefox/Flash. Will it make it into the 2.6 release, which is potentially a week away? Maybe – testing like the above would help.
p.s. we would all individually owe Alex a beer for this.
I gots a Summer of Code org application to write up, so you get a linkdump:
An mp3 of “PFSense II, Rocking The Datacenter” at NYCBUG, via Dru Lavigne. Also, her “BSD for Linux Users” at the SCALE website.
A 5-part series about Scripting Vim, written by a terribly smart guy? Interesting! (Yeah, I did link to a part of it before…)
Have you ever wondered if building more than one pkgsrc package at a time can be a problem? Others have too, and apparently there’s a fix. If you don’t want to have to get to a command line to find the answer, it’s:
PKGSRC_LOCKTYPE?= none
The type of locking that will be done if competing processes
attempt to do work on one package directory simultaneously.
Possible values:
* none: No locking takes place at all.
* once: When the lock has already been aquired by another
process, the current process is terminated.
* sleep: When the lock has already been acquired by another
process, the current process will sleep for PKGSRC_SLEEPSECS
seconds and then try again.You should also set OBJHOSTNAME when you are using the same
copy of pkgsrc on different hosts, maybe via NFS. This is because the locking process writes its process ID into the lockfile, and process IDs on different hosts are unrelated.See also: LOCALBASE_LOCKTYPE, WRKDIR_LOCKTYPE.
We have several potential Summer of Code mentors already, but if you want to get in on the action, let me know. Org applications start Monday, and I’d like a count before then…
Michael Neumann has added his port of the e1000 driver from FreeBSD, though he doesn’t recommend using it yet. He’s looking for testers who have this hardware.
The ‘freeze’ for pkgsrc-2010Q1 is scheduled to start March 16th, which will be right around the same time of the DragonFly 2.6 release. The freeze lasts 2 weeks, usually, so new packages for 2.6 will be built probably about mid-April, based on this info…
A new, free issue of BSD Magazine is out, with the theme of “BSD as a Desktop“. The next issue is “Hosting BSD”, and if that sounds interesting, Dru Lavigne is a contact for writing an article on that theme.
clang, which many people look to as a gcc replacement, is now able to build itself. (Thanks John Marino for the heads-up, some time ago) It can also build world and kernel on DragonFly, going on the work of Sascha Wildner!
Using the pkgsrc package, put
clang_CC=/usr/pkg/bin/clang
in /etc/compilers.conf and then set $CCVER to “clang” when building:
env WORLD_CCVER=clang make -DNO_GCC44 buildworld
I haven’t tried this, so any errors in description are mine, not Sascha’s – can someone verify? I don’t have a test system to run it on right now.
Edit: see Sascha’s comment for the definitive method.
I think I’ve almost caught up on my backlog of Things To Post:
The March issue of the Open Source Business Resource is up, with the theme of “Mobile”. The BeagleBoard and OpenBTS articles are going to appeal to some specific people.
More links from Dru Lavigne.
“It’s sinking in that Sun is gone“, the vi Complete Key Binding List, and Post-quantum cryptography, all ganked from Trivium.
OpenSSL version 0.9.8m has been imported by Peter Avalos; this version contains a bugfix for a security issue.
The iPad is BSD-based, which is neat. I like the notion of covering more BSD devices and products here on the Digest.
But: do you like even the vague concept of open source? The iPad is expressly designed to limit your choices, especially with media consumption and the programs you run. It’s sort of like owning a TV circa 1975 – you get what large media groups and the network (Apple) want you to see. I don’t want to come across as someone who’s complaining because it’s different – I’m complaining because it’s not different.
There’s other people arguing the same idea, with even more reasons. This is one of the most specific. Plus, there’s the natural progression…
(reproduced from my email to users@/kernel@)
The application period for Google Summer of Code 2010 starts in about a
week. We were able to enter in 2008 and 2009, so I’m optimistic that we
will get in for 2010 too.
Saying “I’m willing to mentor” doesn’t force you to commit yet; you don’t
have to work with a student on a project you don’t find interesting.
However, mentoring is a multi-week commitment to support a student who may
or may not have the best planning skills – please be ready to help.
I need to know soon how many potential mentors we have since we have to
ask for a given number of slots from Google as part of the application
process. If you are interested in mentoring, speak up here or by email, please. If
there’s a particular project on the GSoC 2010 page that looks interesting,
put your name by it.
Joerg Sonnenberger announced new behavior in pkgsrc: Performing “bmake install” in pkgsrc with a package that supports DESTDIR will build a binary package and then install from that package. This means a package will be successfully build before the installation process is started, and I assume is to assist further work down the road.
Details: The old behavior was to build and install directly, which “bmake stage-install” can reproduce. DESTDIR support means that the software can be installed as non-root.
There’s several projects in the works:
- Rui Paulo is working on an update of wireless device support in DragonFly; Matthew Dillon added libbsdxml to DragonFly as part of the prep work.
- Joe Talbott is working on unionfs; Nicolas Thery supplied some links, while Matthew Dillon wants deduplication.
For those wanting to try swapcache(8): Newegg has two deals running: a Corsair P128 128GB SATA II SSD for $355 and a Patriot PS-100 2.5″ 64GB SATA I/II SSD for $170. These are probably bigger than what’s needed for using swapcache, but if you’re itching to spend…
Welcome Samuel J. Greear as the newest DragonFly committer; he’s been around the project for a while, but recently became more active.
Michael Neumann has fixed the ability to stream Hammer data between 32 and 64 bit systems. However, this is a change to 64-bit systems that requires them to match; make sure that you are not mixing 64-bit systems built before and after this commit on the 21st.
I can’t find the commit message in the mail archive, so I’ll quote it here:
As part of a report on the status of swapcache and tmpfs, Matthew Dillon noted that a side effect of using a SSD for swapcache means that disk activity stays efficient, and the wear meter on the SSD is reduced much more slowly than for regular disk use.