Re: network routing malfunction
On Wednesday, 23. May 2012 at 18:25, Pierre Abbat wrote: > when I type "route show", I get lots of routes for active > connections. "route" on Linux just shows the routing table. I use "netstat -r" (or -rn to not wait for DNS) to see the routing table – it works everywhere and does what you need. /Matthias
Re: Graphics Setup Help!
Hi, can you tell us a little about what doesn't work when you try to use the NVidia card? For starts, there is the "vesa" driver which basically works with every video card out there but severely lacks in features (can't remember exactly, dual-monitor use may among those missing). The Nouveau driver is available in pkgsrc-wip[1][2] although I don't know the state of it, if it is up to date and if it builds on DragonFly. Give it a shot… Best, Matthias [1] http://pkgsrc.se/wip/xf86-video-nouveau [2] http://pkgsrc-wip.sourceforge.net/
Re: pkgsrc2012Q1-firefox build fails
Hi, the problem is that some packages try to use the installed versions of their own shared libraries or tools during the build process. In the past I've worked around this by - linking (ln -s) from the old library file to the new one - 'make replace'ing the package in question to be sure: - delete the symlink - 'make replace' again then: - continue with pkg_rolling-replace I think you can, if pressed on time and this happens a lot for a particular .so, just create the symlink, let pkg_rolling-replace run through, then delete it. Not sure if this whole thing is a bug and if so, where. Best, Matthias
Re: Mount Ext3 under DragonFlyBSD
On Saturday, 28. April 2012 at 12:27, v...@ukr.net wrote: > > it has the 'sparse_super' flag turned on. It looks like it is > impossible to turn it off once the file system is created, and > DragonFlyBSD driver for Ext2 file system does not support this flag for > some reason. Is that so? Yes, sadly the ext2fs driver in DragonFly is a bit ancient. /M
Reply-To / was: Re: specifying "-s" flag for openntpd in rc.conf does not work
"You are holding it wrong"™ This is not a question of mail client features and your particular favourite being superior to everything else. The question is: what should be _the default_? Should a reply to a public question be public as well unless explicitly chosen not to be or should it be the other way round? I do have a preference there which is of absolutely no consequence to the content of this rant. Matthias -- Matthias Rampke Sent with Sparrow (http://www.sparrowmailapp.com/?sig) (because I can't stand clients that force me to press my mail into folders) On Wednesday, 4. April 2012 at 22:50, Max Herrgard wrote: > David Crosswell wrote: > > On 05/04/12 03:54, McLone wrote: > > > 2012/4/4 Andrey N. Oktyabrski: > > > > P.S. To the mail list administrator: DragonFly mail lists have not the > > > > Reply-To header. Is it possible to add it in the mail lists manager > > > > settings? I forget to "Reply all" regularly and press "Reply" instead > > > > :-( > > > > > > > > > > seconding > > > > > > > Get a better client. > > This was sent with the 'Reply List' feature in Ice-Dove. > > This app. may not be available in DFly, but there must be others. > > > > > mail/thunderbird (hit the reply list button) and mail/mutt (hit 'L') > have this feature and both run on dragonfly. > > cheers, > max > >
Re: well wireless card for dragonfly
2012/4/2 Andrey N. Oktyabrski : > Atheros 2413 cardbus wireless card. The machine must be rebooted at least > once per week because wifi AP stops. For me sudo rcrun restart netif usually does the trick, no full reboot necessary. Best, Matthias
Re: running vkernels in production
On Fri, Feb 3, 2012 at 12:52, n0g0013 wrote: > the solution would > appear to be, use `/etc/ttys` to start and manage the process. Another solution would be to use tmux or screen in the host system to start the vkernels in detached sessions (I've found tmux to be better suited; screen gets confused if started outside an interactive environment, e.g. startup scripts) I use separate tmux sessions to run semi-interactive $stuff automatically; since I want them to be associated with a specific user I have them started from that user's crontab with entries like @reboot tmux -u new-session -d -s name1 "/path/to/command1 arguments1" @reboot sleep 2; tmux -u new-session -d -s name2 "/path/to/command2 arguments2" Note that I found a race-condition in tmux where one of the sessions won't be reachable if they're started to quickly; thus the "sleep 2;". This delay can probably be the same for more subsequent sessions as this only happens on startup of the tmux server, but I haven't tried that in practice. You can then attach to the vserver's console by ssh -t user@server "tmux attach -t name1" etc. This procedure works in general for any program that you want to run unattended but may need to reach interactively. Hoping this helps, Matthias
Re: kernel version in git repo
On Wednesday, 25. January 2012 at 18:34, Pierre Abbat wrote: > Matthias answered by private email Ooops, forgot to add users@ again. For the record, here's the core bit: > to mark your currently checked out version use "git tag bug_foo"; if you just > want to know the commit ID use e.g. "git show" or "git log" (top entry) – in > the end both ways are equivalent only that with git tag the note is within > the git repo. /Matthias
Re: wifi AP setup howto
2012/1/3 Andrey N. Oktyabrski : > ifconfig_wlan0="create wlandev ath0 wlanmode hostap inet 192.168.2.234 you need wlans_ath0="wlan0" ifconfig_wlan0="wlanmode hostap inet 192.168.2.234 netmask 255.255.255.0 ssid thinkpot nwkey topsecretpswd authmode shared mode 11g pureg hidessid" in /etc/rc.conf – not 100% sure about the latter, but in any case creating wlan0 and configuring it are separate steps. without the wlans_ath0 line the rc scripts don't even know they should look for *_wlan0 variables. Also, check out the docs about hostapd, I think it should (needs to?) be involved in creating an AP. Again, not sure. :matthiasr
Re: w3m / boehm-gc issue (really is getcontext in x86_64)
On Friday, 16. December 2011 at 13:28, John Marino wrote: > > FYI, Per bugs #2108, Alex created the getContext.S function a few months ago: > http://leaf.dragonflybsd.org/~alexh/0001-getcontext-x86_64-implementation.patch > FWIW, here's my (failed) stab at taking Alex' getcontext.S and the old x86_64 makecontext.c and signalcontext.c from before the removal: http://2p.grade.so/dfly/getmakesignalcontext.patch It does not compile, because there is no PIC_PROLOGUE and PIC_EPILOGUE for x86_64 – I don't know why (not needed?) and my last and only steps with assembler were a long long time ago and involved only a quarter of bits so I really don't know where to go from that. But maybe the {make,signal}context implementations are of any use? /matthiasr
Re: sharing files between DF guest and Win host in VirtualBox
On Freitag, 9. Dezember 2011 at 00:48, sweepslate wrote: > > However, in the meantime, I just noticed that DF can't access the > Internet. This is an issue I need to take in a VBox forum. > Be aware of the different VirtualBox network types – from the top of my head the most important ones: - Bridge – as if your VM was connected physically to the same Ethernet (or WLAN, but to the VM it will appear as regular ethernet). This case is IMHO the easiest to use and configure and you can reach the VM from any computer on the network. Internet access will work with this just as it works for everything else. Downside is that you have to decide on a host interface to use (e.g. if I use my MacBook's Ethernet it won't work via WiFi and vice versa). - NAT – VBox will create a virtual, internal network and act as a router for it. Gives internet access to the VM, but the VM can't be accessed from outside the host. Also, in usual home network configurations, this ends up in a double-NAT situation (address being translated once on the VBox host and then on the network router) which may or may not sometimes break stuff. - Internal – the same, but without any routing. VM can only be accessed by host and can only access host (unless you do the routing manually). Maybe you chose the Internal option? try switching to NAT and/or Bridging. /matthiasr
Re: How to Correctly map sernos to device names
On Donnerstag, 24. November 2011 at 11:14, Siju George wrote: > Hi, > > Is there a way to correctly map /dev/serno to device names. I don't think there is a 1:1 correspondence that is stable across reboots (e.g. when I had two external HDs they sometimes switched places between da8 and da9 – which is why you should use sernos as much as possible). You can look at the minor and major device numbers though: % ls -l /dev/da* /dev/serno/WMAVU2232228* crw-r- 1 root operator 31, 0x1e110047 Nov 24 17:00 /dev/da8 crw-r- 1 root operator 31, 0x1e120047 Nov 24 17:20 /dev/da8s1 crw-r- 1 root operator 31, 0x1e110047 Nov 24 17:00 /dev/serno/WMAVU2232228 crw-r- 1 root operator 31, 0x1e120047 Nov 24 17:20 /dev/serno/WMAVU2232228.s1 with matching devices they will be the same. //matthiasr
Re: cant change font size within gedit
On Samstag, 19. November 2011 at 02:22, Pierre Abbat wrote: > > If you run it from the command line, do you get warnings about not being able > > to contact dbus? I had the same problem and never did figure out how to fix > it (I went back to kwrite/kate). I had the same problem with avahi – rebuilding dbus somehow fixed it.
Re: partition invalid or corrupt
Hello, I see two potential problems here: On Freitag, 11. November 2011 at 07:36, william opensource4you wrote: > Disk 2 has > pcbsd on primary partition (p1), and I've prepared a 15GB partition > (p2) for dfbsd.. I'm not sure about this, but do by any chance PC-BSD and DragonFly use the same partition type? If so, something somewhere might get confused and look for the DragonFly disklabel (= BSD subpartitioning) in the PC-BSD partition. If so, I don't really know what to do about that. > > In my grub (managed by my linux: last fedora) I've put (same as for > netbsd and freebsd): > Root (hd1,1,a) > Makeactive > Chainloader +1 Maybe (but this is more of a blind guess either) try (hd1,1)? --matthiasr
Re: Can someone upgrade tor in Q3?
On Sonntag, 6. November 2011 at 22:17, Pierre Abbat wrote: > > I'm using From git://git.dragonflybsd.org/pkgsrcv2 > (http://git.dragonflybsd.org/pkgsrcv2); how often is it synced? about every 6 hours, unless it's broken … consider trying https://github.com/jsonn/pkgsrc – it's a git mirror maintained by the NetBSD/pkgsrc people. (It's fairly new. Should we consider deprecating pkgsrcv2.git?) --matthiasr
Re: onboard sound
On Donnerstag, 3. November 2011 at 03:58, Pierre Abbat wrote: > Okay, can you recommend some programs to use the sound? I tried alsa play, > and > got this: > > ALSA lib pcm.c:2215:(snd_pcm_open_noupdate) Unknown PCM default > Playback open error: -2,No such file or directory ALSA is Linux-only. There is supposedly a compatibility library, but I've never gotten it to work and it seems to be unmaintained. The BSD soundsystem is OSS compatible; so you want either something with OSS output or something that uses libao; IIRC you also need to install audio/libao-oss for those. A quick test to check if sound works is "cat < /dev/urandom > /dev/pcm0" (expected result: noise :)) --matthiasr
Re: onboard sound
Have a look at dmesg again – you just loaded plenty of drivers, but most of them don't do anything and thus don't write anything to dmesg. Yours should speak up though. --matthiasr On Montag, 31. Oktober 2011 at 22:01, Pierre Abbat wrote: > On Monday 31 October 2011 12:37:45 Samuel J. Greear wrote: > > kldload snd_driver for the pseudo-driver. > > I did that and dev/dsp appeared. How do I find out which module is > responsible? There's no snd_audio810. > > Pierre > -- > sei do'anai mi'a djuno puze'e noroi nalselganse srera
Re: Gconf error
On Mittwoch, 12. Oktober 2011 at 23:35, Pierre Abbat wrote: > I edit a file with gedit and get this line repeatedly on the terminal: > > GConf Error: Failed to contact configuration server; the most common cause is > a missing or misconfigured D-Bus session bus daemon. See > http://projects.gnome.org/gconf/ for information. (Details - 1: Failed to > get connection to session: Unexpected lack of content trying to read a line) > > There is a dbus daemon running. I even ran "dbus-launch gedit &" and got the > same error. What's wrong? I've had a similar issue* with dbus refusing to speak to avahi. For me, re-building and reinstalling (bmake clean && bmake replace clean) sysutils/dbus resolved this; maybe you could try that? --matthiasr *avahi-daemon[1211]: dbus_bus_get_private(): Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.
Re: How can I make a movie or animation?
On Donnerstag, 8. September 2011 at 06:38, Pierre Abbat wrote: > Is > this the right tool, or should I try something else? I don't know about mencoder, but ffmpeg can definitely do what you want: http://en.wikibooks.org/wiki/FFMPEG_An_Intermediate_Guide/image_sequence#Making_a_video_from_an_Image_Sequence basically it's ffmpeg -f image2 -i --matthiasr
Re: What happened to branches other than dragonfly-2010Q3
On Freitag, 19. August 2011 at 18:40, Siju George wrote: > > Also why not consider providing a pkgsrc.tgz snapshot for download > once a week or so? > It would be easier to pull through http/ftp since it can be resumed? That would be redundant now that there's https://github.com/jsonn/pkgsrc – Github already offers automatic .tar.gz and .zip snapshots of all branches. -m.
Re: What happened to branches other than dragonfly-2010Q3
On Freitag, 19. August 2011 at 15:02, Justin Sherrill wrote: > > I'm curious to see if git chokes when downloading from a different source... In this case it definitely will – the repos were created by completely different scripts and ways (they convert CVS -> Fossil -> git AFAIK), so the commit IDs won't match. In general, adding another clone of the *same* repo as a git remote works fine (e.g. you have a clone of dragonfly.git on your home server, clone that on your workstation, then add the git.dragonflybsd.org or leaf repos). But the NetBSD pkgsrc git mirror and the DragonFly git mirror have to be considered completely separate. -m.
Re: radvd
>From what I remember, it's more-or-less-ish the case that end-user systems (OS >X, desktop Linux, Windows(?)) will accept address configuration (be it v4 or >v6) from the network by default, "server" systems won't. In-betweens like >Debian require you to decide on this during installation. IMHO basic network configuration for both IPv4 and IPv6 during installation, offering a choice between DHCP and static for the v4 stack and DHCPv6, router advertisements and static configuration for v6, would cover 99% of all cases. On Donnerstag, 18. August 2011 at 01:53, Justin Sherrill wrote: > If there's other operating systems that run it by default, I'd rather > stick with what people will expect (just working) than turning it off > because of a security problem that is, at this point, hypothetical. > > I haven't used IPv6 very often, so there may be arguments I don't know about. >
Re: radvd
On Aug 17, 2011 11:05 PM, "Justin Sherrill" wrote: > > Would it be worth setting this by default, since (someday, somehow) > IPv6 is becoming more common? I'd be in favour, but I'm not sure about the security implications of accepting anything from the network by default upon plugging in. (How is it with DHCPv4 now? It would be best to just do the analog thing for IPv6, I think) -m.
Re: radvd
DragonFly doesn't accept router advertisements by default. You have to set something like ip6mode="autohost" in /etc/rc.conf and some sysctl (ends in accept_rtadv). -m.
Re: rebuilding pkg_install fails
On Donnerstag, 11. August 2011 at 07:15, Pierre Abbat wrote: > > > > I assume there's something out of place on your machine that's > > confusing the build. Maybe try doing a 'bmake clean'? That > > suggestion sounds kinda weak now that I said it. > > Tried that. Didn't fix it. I've seen this before … or some reason the build does not pick up the libfetch built during the pkgsrc bootstrap but DragonFly's system libfetch. They are incompatible. Dunno why though, this shouldn't happen. Sorry I can't help you more, I don't have access to my DFly machine for the rest of the week, but take this as a pointer what to look for. -m.
Re: screen problem
On Montag, 13. Juni 2011 at 22:21, Tim Darby wrote: > kernel: pid 18254 (screen-4.0.3), uid 0: exited on signal 11 (core dumped) I remember seeing that too. I don't exactly know what made it go away, probably recompiling the package (and possibly some library it depends on) from pkgsrc. hope that helps, m.
Re: system update question
On Montag, 6. Juni 2011 at 19:54, Justin Sherrill wrote: > Wouldn't he need to do this once, to establish the local copy of the branch? > > cd /usr/src && git branch DragonFly_RELEASE_2_10 > origin/DragonFly_RELEASE_2_10 > git does this automagically, at least in recent versions, if you try to switch to a non-existent branch that is in origin/ -m.
Re: system update question
Hi, to switch to the stable branch cd /usr/src && git checkout DragonFly_RELEASE_2_10 from that point on git pull in /usr/src or make src-update in /usr will keep you up to date on the stable branch. For more info, have a look at /usr/src/{README,UPDATING} and section 7 of the man pages, esp. build(7). Regards, matthiasr
Re: Chromium Browser on DragonFly
Hello all, just a quick heads up that a Chromium 11 port is now available from the chromium11 branch of the repo[1]. Completely untested on x86_64 for now, i386 more or less works. Preliminary i386 package (again, against *my* setup) at [2]. Again, credit goes to rxg for updating the port to Chromium 11 in the first place, I just fix up the build errors. -matthiasr [1] http://gitweb.dragonflybsd.org/~matthiasr/chromium.git/tree/chromium11 [2] http://leaf.dragonflybsd.org/~matthiasr/chromium11_i386.tgz
Re: where'd kde4 go?
On Montag, 16. Mai 2011 at 23:27, Pierre Abbat wrote: I ran "pkgin se kdebase4" and it came up blank. The package is > in pkgsrc. What broke it? The latest reports JustinS posted[1] say it's because of heimdal and xine-libs. > Can it be fixed? Probably. Feel encouraged to do so :) -matthiasr [1] http://avalon.dragonflybsd.org/reports/x86_64/2.11/20110508.2107/meta/report.html
Re: Chromium Browser on DragonFly
The i386 package is now available at [1] and it actually works (as in: it does indeed display web pages - I'm writing this through the GMail web interface from it now!) -m. [1] http://leaf.dragonflybsd.org/~matthiasr/chromium-i386.tgz
Chromium Browser on DragonFly
Hello all, following JustinS' call[1] I have done a preliminary (as in: compiles, doesn't really work yet, see below) port of the Chromium browser to DragonFly BSD. A first x86_64 binary package is available at [2] (compiled against my setup, e.g. Python 2.7, so YMMV) and the pkgsrc bits to compile it yourself at [3]: git clone git://leaf.dragonflybsd.org/~matthiasr/chromium.git /usr/pkgsrc/wip/chromium cd /usr/pkgsrc/wip/chromium bmake install chrome (this will take a long time.) I have an i386 build running right now, the package should be available in a few hours. It starts up [4] but doesn't really work yet; it doesn't load any page, whether from the net or internal (such as settings). Bugfixes are very welcome :) Hat tips go to JustinS for keeping us all up to date and to all the original *BSD porters, most of all rxg who brought Chromium to NetBSD and pkgsrc - they have done most of the real porting work. -matthiasr [1] http://www.shiningsilence.com/dbsdlog/2011/04/29/7697.html [2] http://leaf.dragonflybsd.org/~matthiasr/chromium-x86_64.tgz [3] git://leaf.dragonflybsd.org/~matthiasr/chromium.git [4] http://leaf.dragonflybsd.org/~matthiasr/chromium-dfly.png
Re: System on SSD
On Dienstag, 10. Mai 2011 at 21:04, Sven Gaerner wrote: I also want to move /usr/src and > /usr/pkgsrc and the build directories to a normal HDD. Why? Isn't this where a SSD really shines? (the pkgsrc tree doesn't matter much because it doesn't contain the actual source, but I think you really really want the work directories on the SSD) -matthiasr
Re: Some problems installing dragonFly BSD. Can anybody help?
On Montag, 9. Mai 2011 at 09:47, Ivan Uemlianin wrote: > One thing that happened both times: on startx, the laptop screen filled > with semi-random-looking blocks of colour for a split second before > going into the wm. I think I've seen this before on some linux > installs. Presumably X is configured to some default (vesa?) and not > the particular graphics card, screen, etc., I have on my laptop. I don't think that's anything to worry about … I've once seen what I had on screen before the last reboot, so I guess this is just random contents of video RAM being displayed before anything is drawn over them. On Samstag, 7. Mai 2011 at 17:44, Justin Sherrill wrote: > This is a new bug, then, cause I think the original ath(4) problem is fixed. > I don't have the right laptop and wireless combo to test. In any case, you > may want to file a report including the network encryption type. >From what I can tell from [1] this has not been fixed. I can still file a >report if that helps? Also note that I have control over the WLAN router >(other than my flatmates kicking me when I kill their internet), so if I can >provide any additional info / try anything out let me know. Re: IO APIC Disabling IO APIC didn't help with the SMP kernel, it complains about lapic initialization before dropping to the debugger (is there a LAPIC loader tunable? I can't find any documentation on this or hw.apic_io_enable). Are SMP kernels on UP machines *supposed* to work anyway? -matthiasr [1] http://gitweb.dragonflybsd.org/dragonfly.git/history/HEAD:/sys/dev/netif/ath
Re: Some problems installing dragonFly BSD. Can anybody help?
On Fri, 6 May 2011, Justin Sherrill wrote: sysctl net.wlan.force_swcrypto=1 may help. http://leaf.dragonflybsd.org/mailarchive/commits/2010-11/msg00169.html Works for me. Thank you! -matthiasr
Re: Some problems installing dragonFly BSD. Can anybody help?
On Freitag, 6. Mai 2011 at 16:03, Matthias Rampke wrote: >From what I gather SMP Kernels with IO APIC enabled don't work. I just settled >on using UP since I only have one core anyway. You may try setting >hw.apic_io_enable=0 in /boot/loader.conf Sorry, that wasn't worded well. I meant to say that SMP + IO APIC doesn't seem to work with x40/x60-ish ThinkPads, so this is very probably a ThinkPad specific issue, not a DragonFly issue. -matthiasr
Re: Some problems installing dragonFly BSD. Can anybody help?
On Freitag, 6. Mai 2011 at 15:30, Ivan Uemlianin wrote: Dear All > > I am installing DragonFly BSD onto a Thinkpad X60. I recently installed on a X40, so let's see. There is also a page on the Website about the T42[1] and the X61s[2], especially most of the latter should apply to your case as well. > Symptoms: > > Major: > > 1. After the post-installation configuration step, I chose reboot. > While rebooting, the system seemed to crash and I ended up inside a > debugger. However, after a successful shutdown and reboot, the system > seemed to run OK. >From what I gather SMP Kernels with IO APIC enabled don't work. I just settled >on using UP since I only have one core anyway. You may try setting >hw.apic_io_enable=0 in /boot/loader.conf > > 2. During configuration dhcp seemed to find the network (at least, a > dialogue appeared, giving me some info ending with "Status: Active"), > but after I'd rebooted into X and launched Firefox, Firefox could not > find external urls. So, perhaps dhcp didn't work and I should configure > the network manually. Is this on the WLAN? Which card? I'm having massive trouble getting ath(4) to work with most WLAN routers, e.g. I don't receive the DHCPOFFERs from my home WLAN. My phone's WiFi hotspot (Motorola Milestone / CyanogenMod 6) works perfectly, though. I'm sorry I haven't gotten around to reporting this properly yet. > > 3. The window manager is TWM. On exiting TWM, the screen froze and > didn't return me to the console. I had to go to another console and > kill X manually. When it freezes, what does /var/log/Xorg.0.log say? > > Minor: > > 4. I don't think I set the correct keyboard map (and/or screen map?). > Most of the keyboard works as it should, but shift-3 does not produce > the expected "£" --- it doesn't produce anything, so maybe the keyboard > map is OK but the screen map is wrong? Which layout are you expecting to use, and what are the respective lines in /etc/rc.conf? > > 5. I was expecting FVWM as the window manager: I thought the docs said > that was the default. Presumably I can change the window manager. > Should it have been FVWM? Is the fact that I got TWM instead a symptom > that some config was wrong? Probably ... (I'm not using the GUI image, so rather wild guessing here). You could a) check that FVWM2 is installed (just run fvwm2 from within twm) b) add "exec fvwm2" as the last line of your $HOME/.xsession (and check that it is the *only* exec line, of course) > > All in all, it doesn't look very healthy. Can anybody indicate what > might have gone wrong? Or, if I re-install, what signals I should be > looking out for? It's not so bad :) Regards, matthiasr [1] http://www.dragonflybsd.org/docs/user/ThinkpadT42/ [2] http://www.dragonflybsd.org/docs/user/ThinkpadX61s/
Re: postgresql904 did not install
On Tue, 3 May 2011, 78dd085bd...@gmail.com wrote: dfly# bmake install tl;dr: It appears you have PostgreSQL 9.0.3 installed. Use "bmake update". The longer explanation: pkgsrc has separate make targets for upgrading an already installed package. "replace" simply replaces the package, not minding the consequences. This will break things if a shared library version changes, because dependent packages won't find their .so's any more. "update" also rebuilds and updates any dependant. For a full update of all outdated packages (e.g. using pkg_chk), using "bmake update" will work, but take longer than necessary because some packages may be rebuilt more than once. pkg_rolling-replace will take the whole dependency tree and sort the list of packages such that they can be "bmake replace"'d one after the other safely. It is a good idea to always use pkg_rolling-replace for updating packages. Regards, matthiasr.
Re: virtualbox & "greetings"
I think I messed up the recipients on this, so here it is again. Forwarded message: > From: Matthias Rampke > To: Justin Sherrill > Date: Sonntag, 3. April 2011 11:47:41 > Subject: Re: virtualbox & "greetings" > > On Sonntag, 3. April 2011 at 03:32, Justin Sherrill wrote: > On Sat, Apr 2, 2011 at 10:52 AM, * William > wrote: > > > What ever comes up with the ISO image; is missing some basic shell set-up > > > because it FAULT-s duiring start-up. > Could you describe this in more detail? When does it fault? Which version of > VirtualBox and DragonFly did you try? > > > Virtualbox seems to have trouble with DragonFly on every other Virtualbox > > release. i.e. I've heard reports that the Virtualbox 4 release has trouble, > > but recent Virtualbox 3 releases did not. Your mileage may vary. > > VirtualBox 4.0.4 (mostly) works for me (master with x86_64) - only the ahci > initialisation hangs quite often (about 50% of the boots). Just reset the VM > or break into the debugger (ctrl-alt-esc) and type reset until it works. > > Regards, M. >
Re: Dual use Filesystem
On Montag, 21. März 2011 at 00:31, Justin C. Sherrill wrote: > ext3 or UFS might be at least readable for each side. However, the > lowest-common-denominator of DOS (i.e. FAT) is possibly the most portable. >From my experience, ext2fs seems to be the most widespread filesystem among >the unix-ish systems. beware though, that many of the *BSD drivers can not >handle inode sizes larger than 128 byte (the default on linux is now 256 >bytes) and most of the extensions (including those enabled by default with >mke2fs) are not supported either. As far as I know, only linux supports journaling (i.e. ext3). The NetBSD ext2fs driver has recently gained support for large inodes, it may be worth backporting that. Last time I tried the ext2 tools (from pkgsrc) did not work on DragonFly because they could not determine the size of the devices. It may be worth porting them too (because without journaling the filesystem won't mount R/W after an unclean shutdown and will need to be fsck'ed). I haven't tried UFS with DFly and Linux yet, it may be worth a shot too. msdosfs works too, of course, but it is very limited. Regards, Matthias
Re: Dragonfly network changes
On Fri, Feb 18, 2011 at 09:29, Chris Turner wrote: > On 02/18/11 00:53, Francois Tigeot wrote: > >> Do they offer IPv6 ? > > man gif(4) I think whether an ISP offers native IPv6 a very valid question and everybody should ask it to theirs once in a while … in the meantime I'm very happy with the (free) service of sixxs.net – their tunneling is fast (haven't noticed any bandwitdth shortage; adds about 20ms latency for me) and works even through NATed connections with dynamic IPs. I haven't tried their software on DFly yet, but it's in pkgsrc (net/aiccu) and is generally very portable, so it should work (remember to load tun). After your tunnel is up and stable for a few weeks you can even get a /48 subnet assigned. My home network now has 65535 times the size of all the v4 Internet ;) Have fun :) Matthias
Re: How to tell where/how space is used?
On Mon, Feb 14, 2011 at 06:39, Pierre Abbat wrote: > 2 GB Hammer partition This is something you're not supposed to do(tm): the HAMMER(5) man page recommends a minimum size of 50GB [1]. > It *started out* 15% full. It's still 15% full; there's > not much on it. That's probably just HAMMER metadata. [1] http://leaf.dragonflybsd.org/cgi/web-man?command=hammer§ion=5 "Large File Systems & Multi Volume"