Re: What's Meta v1?
On Sunday, February 13, 2022 2:24:59 PM EST Dan Cross wrote: > It is the version listed in the file `/var/db/pkg/Avalon.meta`. Oh my > dragonfly dev machine (which I basically never have time to poke at :-( > ...) the contents are: > > {"version":2,"packing_format":"txz","manifests":"packagesite.yaml","filesite > ":"filesite.yaml","manifests_archive":"packagesite","filesite_archive":"file > site"} > > On the other hand, my Ham radio DragonFly machine, which is running > 6.2-RELEASE, has this: > > {"version":1,"packing_format":"txz","digest_format":"sha256_base32","digests > ":"digests","digests_archive":"digests","manifests":"packagesite.yaml","file > site":"filesite.yaml","manifests_archive":"packagesite","filesite_archive":" > filesite"} > > I'm not quite sure what creates or updates that file, nor what the > implication of simply updating the "version": in the latter file from 1 to > 2 would be. My Avalon.meta file says version 1. What should I do? Pierre -- Lanthanidia deliciosa: What the kiwifruit would be if it weren't so radioactive.
What's Meta v1?
I just ran "pkg ins hs-stack" and got "WARNING: Meta v1 support will be removed in the next version". What's this mean? Pierre -- gau do li'i co'e kei do
Re: How do I reinstall kde?
On Tuesday, December 28, 2021 6:15:15 AM EST Antonio Huete Jimenez wrote: > The "kde" meta package was missing because there were dependencies that > were not building. Zrj has fixed them and now the binary packages have > been built and uploaded to the servers. Thanks! Pierre -- ve ka'a ro klaji la .romas. se jmaji
Re: How do I reinstall kde?
On Friday, December 24, 2021 7:31:12 PM EST Gerald Henriksen wrote: > ilmbase is for the OpenEXR image format (with originated at ILM aka > Industrial Light & Magic), which I believe KDE (and possibly other > things) pulls in. Okay, but what's the name of the KDE desktop package? Has it been renamed, or did it fail to build? Pierre -- I believe in Yellow when I'm in Sweden and in Black when I'm in Wales.
How do I reinstall kde?
I just upgraded kernel and world and packages. It told me it's removing a bunch of packages. One of them is Postfix; two people told me on IRC that it's because of OpenSSL and I should reinstall Postfix, so I did. Several others are py37-foo, which was replaced with py38-foo. Two others are openldap-client and ilmbase; I don't know what uses them. That leaves kde5, plasma5-plasma, and plasma5-plasma-desktop. Which package should I install to get a KDE desktop? Pierre -- sei do'anai mi'a djuno puze'e noroi nalselganse srera
Re: Problems with sites using Let's Encrypt certificates
On Wednesday, October 13, 2021 8:40:11 PM EDT James Cook wrote: > - If you upgrade to DragonflyBSD 6.0.1, the problem will go away. See > > https://www.dragonflydigest.com/2021/10/13/26267.html I'm running 6.1.0.3. Should I upgrade to the latest master? Pierre -- li ze te'a ci vu'u ci bi'e te'a mu du li ci su'i ze te'a mu bi'e vu'u ci
Miredo doesn't work right when IP address changes
My ISP's IPv6 connection stopped working several days ago (I still get an address, but trying to ping from it results in "ping6: sendmsg: No route to host"). Miredo works fine on Linux, but not on DragonFly. Every so often my connection breaks (and will break tomorrow when the thunderstorm comes). The teredo device on Linux looks like this in ifconfig: teredo: flags=4305 mtu 1280 inet6 2001:0:c38c:c38c:3cc5:519f:bcd3:2eb7 prefixlen 32 scopeid 0x0 inet6 fe80::470a:7247:656d:33e3 prefixlen 64 scopeid 0x20 inet6 fe80:::: prefixlen 64 scopeid 0x20 unspec 00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00 txqueuelen 500 (UNSPEC) RX packets 18499 bytes 7083906 (7.0 MB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 21127 bytes 2022794 (2.0 MB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 And in BSD: teredo: flags=8043 metric 0 mtu 1500 inet6 fe80::222:4dff:fe84:bb1d%teredo prefixlen 64 scopeid 0x5 inet6 fe80::::%teredo prefixlen 64 scopeid 0x5 inet6 2001:0:c38c:c38c:2002:602f:bcd3:2f25 prefixlen 128 inet6 2001:0:c38c:c38c:3c32:5e2f:bcd3:2f18 prefixlen 128 inet6 2001:0:c38c:c38c:20e2:502f:bcd3:2eb7 prefixlen 128 groups: tun Opened by PID 18239 Restarting miredo results in no address: teredo: flags=8002 metric 0 mtu 1500 groups: tun To get a single address, I have to do this: # /usr/local/etc/rc.d/miredo stop Stopping miredo. # ifconfig teredo destroy # /usr/local/etc/rc.d/miredo start Starting miredo. I'm running v6.1.0.3.g70c69-DEVELOPMENT and miredo-1.2.6_6. Is the problem in miredo or in the kernel? -- .i toljundi do .ibabo mi'afra tu'a do .ibabo damba do .ibabo do jinga .icu'u la ma'atman.
Electrum upgrade
I just ran Electrum to check how much Bitcoin I have in a wallet I haven't checked in a few years, after running pkg which told me the most recent version is installed. It says there's a version 4.1.5 available; the version in the repo is 4.0.9. Pierre -- lo ponse be lo mruli ku po'o cu ga'ezga roda lo ka dinko
Re: pkg: mkdir(/tmp): Invalid argument
On Tuesday, May 4, 2021 12:25:58 PM EDT Antonio Huete Jiménez wrote: > Yes, packages for 5.9 (old master) are 'dragonfly:5.10' because 5.10 > was a tentative version number for the next release. But as it turns > out, we branched 6.0 so now the packages for current master (6.1) are > 'dragonfly:6.2' and the 5.10 ones are removed. So should I install packages for 6.0 or 6.2? Pierre -- La sal en el mar es más que en la sangre. Le sel dans la mer est plus que dans le sang.
Re: pkg: mkdir(/tmp): Invalid argument
On Wednesday, April 14, 2021 8:08:57 PM EDT Antonio Huete Jiménez wrote: > Would it be possible for you to upgrade to latest master and retried? I just got around to rebooting. Here's what happened when I tried to install: # pkg ins dmalloc Updating Avalon repository catalogue... pkg: http://mirror-master.dragonflybsd.org/dports/dragonfly:5.10:x86:64/LATEST/ meta.txz: Not Found repository Avalon has no meta file, using default settings pkg: http://mirror-master.dragonflybsd.org/dports/dragonfly:5.10:x86:64/LATEST/ packagesite.txz: Not Found Unable to update repository Avalon Error updating repositories! Tuxillo sent a message on the 24th that 5.10 has been removed, but I'm running v5.9.0.1502, compiled on the 15th. Pierre -- sei do'anai mi'a djuno puze'e noroi nalselganse srera
Re: pkg: mkdir(/tmp): Invalid argument
On Monday, April 12, 2021 7:27:19 PM EDT Antonio Huete Jiménez wrote: > What's your dragonfly version? 5.9.0.895.gfae39-DEVELOPMENT -- sei do'anai mi'a djuno puze'e noroi nalselganse srera
pkg: mkdir(/tmp): Invalid argument
Resending. I've fixed the bug, but the problem with pkg remains. I tried to install dmalloc to debug a memory allocation bug. The interaction went like this: # pkg sea dmalloc pkg: mkdir(/tmp): Invalid argument pkg: mkdir(/tmp): Invalid argument dmalloc-5.5.2_1Portable debug memory allocation library # pkg ins dmalloc Updating Avalon repository catalogue... pkg: mkdir(/tmp): Invalid argument pkg: mkdir(/tmp): Invalid argument pkg: mkdir(/tmp): Invalid argument Avalon repository is up to date. All repositories are up to date. New version of pkg detected; it needs to be installed first. The following 1 package(s) will be affected (of 0 checked): Installed packages to be UPGRADED: pkg: 1.15.10 -> 1.16.1 [Avalon] Number of packages to be upgraded: 1 3 MiB to be downloaded. Proceed with this action? [y/N]: y [1/1] Fetching pkg-1.16.1.txz: 100%3 MiB 95.4kB/s00:32 Checking integrity... done (0 conflicting) [1/1] Upgrading pkg from 1.15.10 to 1.16.1... [1/1] Extracting pkg-1.16.1: 0% pkg: Fail to create /usr:Invalid argument [1/1] Extracting pkg-1.16.1: 100% # pkg ins dmalloc Updating Avalon repository catalogue... pkg: mkdir(/tmp): Invalid argument pkg: mkdir(/tmp): Invalid argument pkg: mkdir(/tmp): Invalid argument Avalon repository is up to date. All repositories are up to date. New version of pkg detected; it needs to be installed first. Checking integrity... done (0 conflicting) The following 1 package(s) will be affected (of 0 checked): Installed packages to be UPGRADED: pkg: 1.15.10 -> 1.16.1 [Avalon] Number of packages to be upgraded: 1 Proceed with this action? [y/N]: y [1/1] Upgrading pkg from 1.15.10 to 1.16.1... [1/1] Extracting pkg-1.16.1: 0% pkg: Fail to create /usr:Invalid argument [1/1] Extracting pkg-1.16.1: 100% I then created a file in /tmp, with no problem. Running pkg-static doesn't help. How do I fix this? Pierre -- li fi'u vu'u fi'u fi'u du li pa
BSD equivalent of sysinfo?
I'm writing a program that reads a point cloud and constructs an octree. From a 6.5 MB LAS file, it constructs a 52 MB octree file, reading and writing every block several times in the process. This takes a very long time, so I'd like to modify the store to always allocate a buffer when there's some RAM remaining (maybe 1/4 of the RAM available when the program started). On Linux, there's a function sysinfo() which does this, where info is a struct sysinfo, but this function does not exist on DragonFly BSD. How can I find out the available memory? The reason the octree is so much bigger than the LAS file is twofold. First, the blocks of the octree are nowhere near full; second, there are several variants of the LAS format, and the class LasPoint has to accommodate all of them. Pierre -- I believe in Yellow when I'm in Sweden and in Black when I'm in Wales.
Re: pkg broke after upgrade to v5.9.0.609
On Wednesday, July 22, 2020 2:17:55 PM EDT Ron Georgia wrote: > All, > > I upgraded from 5.9-DEVELOPMENT DragonFly v5.9.0.583 to 5.9-DEVELOPMENT > DragonFly v5.9.0.609. When I try to use pkg I get this error: > Shared object "libprivate_ssl.so.47" not found, required by "pkg" > > If I ln -s lib/libprivate_ssl.so.48 to /lib/libprivate_ssl.so.47 > (dangerous, I know) I get another library file missing. Any suggestions? Does pkg-static work? Pierre -- I believe in Yellow when I'm in Sweden and in Black when I'm in Wales.
Re: Upgrade from 5.5 failed
Here are the contents of /usr/local: bin llvm50 brlcad llvm60 env llvm80 etc man gcc-aux openjdk8 include openssl infopoudriere kde4sbin lib share libdata tests libexec var llvm35 www llvm36 x86_64-portbld-dragonfly5.6 llvm40 I have scripts in /usr/local/bin/ and /usr/local/sbin/ and configuration (such as Postfix) in /usr/local/etc/. What should I move elsewhere or delete before reinstalling packages? Are any of these directories obsolete? Pierre -- La sal en el mar es más que en la sangre. Le sel dans la mer est plus que dans le sang.
Re: Upgrade from 5.5 failed
On Sunday, May 17, 2020 3:45:01 PM EDT Jonathan Engwall wrote: > 5.8 obviously. As to your problem according to you, your problem is bin > files since kernel 4.00. So get rid of the old files, (that is what you say > you want to do), reload your kernel modules, start your python env so you > can run qt5, and figure out what to do with your software. > I am trying to be helpful. Cmake gives you your error. I read it, file not > found. Which file? I don't you because you haven't shared that information. > Last night I learned a new command. You can test it: > stat /usr/include/X11/Xlib.h > It should return a nice looking table, though mine machine did not. If it > does you will know you have the xorg build environment correctly installed. On DragonFly, X11 is not part of world, and packages are installed in /usr/ local/. So you can stat /usr/local/include/X11/Xlib.h . > download and then compile one from it. Then having three, delete the > oldest. What can it possibly do but lock files? Nothing good at all, that > is what. You don't understand the problem. The problem started with a failed disk, then I lost the package database, but not the installed files. Also, neither my program nor Qt5 is in Python. They're in C++. Stick around and keep learning. Pierre -- li ze te'a ci vu'u ci bi'e te'a mu du li ci su'i ze te'a mu bi'e vu'u ci
Re: Upgrade from 5.5 failed
On Saturday, May 16, 2020 9:51:38 AM EDT Jonathan Engwall wrote: > Do you have a dragonfly5.9 install disk? No. Please explain, in your own words: *why having two kernels in /boot is wrong; *what my problem is. Pierre -- The Black Garden on the Mountain is not on the Black Mountain.
Re: Upgrade from 5.5 failed
On Friday, May 15, 2020 12:48:25 PM EDT Jonathan Engwall wrote: > Something like this is possible if the new kernel did not transplant in > place of the old kernel. > In that case, you may have to remove the the extra kernel. > First you need to establish that you have the new kernel in place. Also you > need a back up /home/*.* plus anything else you would want. > I hope you have done this. > Can you check to see if you two kernels? Everyone who didn't just install DFly has two kernels. I have 5.9 in /boot/ kernel and 5.5 in /boot/kernel.old/. Also in /boot1/kernel I have 5.5 and in / boot1/kernel.old some older version. What went wrong is that the package database is in /var, and I didn't tar /var on the failed disk because the failed sector was in /var/log and I saw no point in transferring log files to the new disk. So when I upgraded packages, it did not delete the files belonging to the old packages. Pierre -- ve ka'a ro klaji la .romas. se jmaji
Building a project that uses Qt fails
I reinstalled packages git, cmake, qt5-core, qt5-gui, qt5-linguist, xorg, and fftw3, which (except fftw3 which is optional) are required for building Bezitopo. I then tried to build PerfectTIN, which is similar, and got errors. One was from Qt and looks like this: CMake Error at /usr/local/lib/cmake/Qt5Gui/Qt5Gui_QICNSPlugin.cmake:4 (_populate_Gui_plugin_properties): _populate_Gui_plugin_properties Macro invoked with incorrect arguments for macro named: _populate_Gui_plugin_properties Call Stack (most recent call first): /usr/local/lib/cmake/Qt5Gui/Qt5GuiConfig.cmake:201 (include) /usr/local/lib/cmake/Qt5Widgets/Qt5WidgetsConfig.cmake:100 (find_package) /usr/local/lib/cmake/Qt5/Qt5Config.cmake:28 (find_package) CMakeLists.txt:50 (find_package) CMake Error at /usr/local/lib/cmake/Qt5Gui/Qt5Gui_QJp2Plugin.cmake:4 (_populate_Gui_plugin_properties): _populate_Gui_plugin_properties Macro invoked with incorrect arguments for macro named: _populate_Gui_plugin_properties Call Stack (most recent call first): /usr/local/lib/cmake/Qt5Gui/Qt5GuiConfig.cmake:201 (include) /usr/local/lib/cmake/Qt5Widgets/Qt5WidgetsConfig.cmake:100 (find_package) /usr/local/lib/cmake/Qt5/Qt5Config.cmake:28 (find_package) CMakeLists.txt:50 (find_package) The other is from pthread (PerfectTIN uses threads, which Bezitopo does not yet) and is in the attached file. I can tell which binaries in /usr/local/bin are left over from previous packages whose package database was lost; they're for DragonFly 4.0.105, 5.0.500, or 5.0.601, instead of the current 5.0.900. But libraries in /usr/ local/lib do not have an ABI version (as far as I can tell with the file command). How can I find these obsolete files and get rid of them? Or is there another package I need to install? The repos are at https:// github.com/phma/perfecttin and https://github.com/phma/bezitopo if you'd like to try building them. PerfectTIN uses Boost, but can build without it. Pierre -- gau do li'i co'e kei do Determining if the pthread_create exist failed with the following output: Change Dir: /home/phma/crypt/build/perfecttin/dbg/CMakeFiles/CMakeTmp Run Build Command(s):/usr/local/bin/gmake cmTC_eaebe/fast /usr/local/bin/gmake -f CMakeFiles/cmTC_eaebe.dir/build.make CMakeFiles/cmTC_eaebe.dir/build gmake[1]: Entering directory '/crypt/phma/build/perfecttin/dbg/CMakeFiles/CMakeTmp' Building C object CMakeFiles/cmTC_eaebe.dir/CheckSymbolExists.c.o /usr/bin/cc-o CMakeFiles/cmTC_eaebe.dir/CheckSymbolExists.c.o -c /home/phma/crypt/build/perfecttin/dbg/CMakeFiles/CMakeTmp/CheckSymbolExists.c Linking C executable cmTC_eaebe /usr/local/bin/cmake -E cmake_link_script CMakeFiles/cmTC_eaebe.dir/link.txt --verbose=1 /usr/bin/cc CMakeFiles/cmTC_eaebe.dir/CheckSymbolExists.c.o -o cmTC_eaebe CMakeFiles/cmTC_eaebe.dir/CheckSymbolExists.c.o:CheckSymbolExists.c:function main: error: undefined reference to 'pthread_create' collect2: error: ld returned 1 exit status gmake[1]: *** [CMakeFiles/cmTC_eaebe.dir/build.make:87: cmTC_eaebe] Error 1 gmake[1]: Leaving directory '/crypt/phma/build/perfecttin/dbg/CMakeFiles/CMakeTmp' gmake: *** [Makefile:121: cmTC_eaebe/fast] Error 2 File /home/phma/crypt/build/perfecttin/dbg/CMakeFiles/CMakeTmp/CheckSymbolExists.c: /* */ #include int main(int argc, char** argv) { (void)argv; #ifndef pthread_create return ((int*)(_create))[argc]; #else (void)argc; return 0; #endif } Determining if the function pthread_create exists in the pthreads failed with the following output: Change Dir: /home/phma/crypt/build/perfecttin/dbg/CMakeFiles/CMakeTmp Run Build Command(s):/usr/local/bin/gmake cmTC_3e574/fast /usr/local/bin/gmake -f CMakeFiles/cmTC_3e574.dir/build.make CMakeFiles/cmTC_3e574.dir/build gmake[1]: Entering directory '/crypt/phma/build/perfecttin/dbg/CMakeFiles/CMakeTmp' Building C object CMakeFiles/cmTC_3e574.dir/CheckFunctionExists.c.o /usr/bin/cc -DCHECK_FUNCTION_EXISTS=pthread_create -o CMakeFiles/cmTC_3e574.dir/CheckFunctionExists.c.o -c /usr/local/share/cmake/Modules/CheckFunctionExists.c Linking C executable cmTC_3e574 /usr/local/bin/cmake -E cmake_link_script CMakeFiles/cmTC_3e574.dir/link.txt --verbose=1 /usr/bin/cc -DCHECK_FUNCTION_EXISTS=pthread_createCMakeFiles/cmTC_3e574.dir/CheckFunctionExists.c.o -o cmTC_3e574 -lpthreads /usr/libexec/binutils227/elf/ld.gold: error: cannot find -lpthreads CMakeFiles/cmTC_3e574.dir/CheckFunctionExists.c.o:CheckFunctionExists.c:function main: error: undefined reference to 'pthread_create' collect2: error: ld returned 1 exit status gmake[1]: *** [CMakeFiles/cmTC_3e574.dir/build.make:87: cmTC_3e574] Error 1 gmake[1]: Leaving directory '/crypt/phma/build/perfecttin/dbg/CMakeFiles/CMakeTmp' gmake: *** [Makefile:121: cmTC_3e574/fast] Error 2 Determining if files windows.h exist failed with the following output: Change Dir: /home/phma/crypt/build/perfecttin/dbg/CMakeFiles/CMakeTmp Run Build Command(s):/usr/local/bin/gmake cmTC_c4f14/fast
Re: Upgrade from 5.5 failed
On Wednesday, May 13, 2020 1:15:37 AM EDT Matthew Dillon wrote: > If you installed the world and kernel properly then a 'uname -a' should > show the dragonfly version. If it looks right, you can try forcing a pkg > update with: > > pkg update -f > pkg upgrade -f This still says "Your packages are up to date." "pkg -N" results in "pkg: no packages registered". It appears that the package database has been lost. Is there a way to recover it? /var was on the failed disk. I didn't tar it, because a file in /var/log had become unwritable. Pierre -- Por H o por B, los campos magnéticos se difieren dentro de un imán.
Re: Upgrade from 5.5 failed
On Tuesday, May 12, 2020 12:31:20 PM EDT Matthew Dillon wrote: > The most recent package sync was May 2nd, which includes chromium version > 81 for both -release and -master. I don't have chromium installed on that box. What else was upgraded recently? Is there something I have to change in /etc to get the latest ABI? Pierre -- Jews use a lunisolar calendar; Muslims use a solely lunar calendar.
Re: Upgrade from 5.5 failed
On Tuesday, May 12, 2020 2:38:22 AM EDT Pierre Abbat wrote: > /var (which is hammer2) and /usr are mounted, but running "nano /etc/fstab" > results in "Error opening terminal: unknown.". Figured it out: TERM=cons25 However, when I did "pkg update" and "pkg upgrade", it said "Your packages are up to date.". It's been months since I've upgraded packages, and I just upgraded the kernel; shouldn't it show that there are packages to upgrade? Pierre -- La sal en el mar es más que en la sangre. Le sel dans la mer est plus que dans le sang.
Upgrade from 5.5 failed
I pulled the latest, made and installed kernel and world, made upgrade, and just rebooted. I get the following: done. uhid0: on usbus1 dumpon: crash dumps to /dev/serno/131061401278.s1b (32. 0x20009) Starting file system checks: /dev/serno/131061401278.s1a: FILESYSTEM CLEAN: SKIPPING CHECKS /dev/serno/131061401278.s1a: clean, 87018 free (1762 frags, 10657 blocks, 0.5% fragmentation) /dev/serno/S599NE0MA759522.s1a: FILESYSTEM CLEAN: SKIPPING CHECKS /dev/serno/S599NE0MA759522.s1a: clean, 221026 free (386 frags,27580 blocks, 0.1% fragmentation) Shared object "libdmsg.so.3" not found, required by "hammer2" mount_hammer2: unable to connect to cluster controller mount_hammer2: cluster_connect(/dev/serno/S599NE0MA759522.s1e@DATA) failed hammer2_mount: '/dev/serno/S599NE0MA759522.s1e@DATA' hammer2_mount: dev="/dev/serno/S599NE0MA759522.s1e" label="DATA" rdonly=0 hammer2: using volume header #1 hammer2 mount "/dev/serno/S599NE0MA759522.s1e": no recovery needed hammer2_mount hmp=0xf80117x22000 pmp=0xf80090542e00 mount: exec mount_linprocfs not found in /sbin, /usr/sbin: No such file or directory Mounting /etc/fstab filesystems failed: startup aborted ERROR: ABORTING BOOT (sending SIGTERM to parent)! May 12 06:05:02 init: /bin/sh on /etc/rc terminated abnormally, going to single user mode Enter full path of shell or RETURN for /bin/sh: Hitting enter gets me a root prompt. Should I cpdup /boot1 back to /boot and try again? What about the old world? Or do I just need to comment out linprocfs in /etc/fstab? /var (which is hammer2) and /usr are mounted, but running "nano /etc/fstab" results in "Error opening terminal: unknown.". Pierre -- La sal en el mar es más que en la sangre. Le sel dans la mer est plus que dans le sang.
Re: Defining a locale
On Friday, April 24, 2020 8:24:57 AM EDT Pierre-Alain TORET wrote: > Hello Pierre, > > running a find for fr_FR.UTF-8, it seems that the locale files are stored in > /usr/share/locale. Those are compiled files. Where are the source files? Pierre -- Lanthanidia deliciosa: What the kiwifruit would be if it weren't so radioactive.
Defining a locale
As en_DK has switched to day/month/year and en_SE has disappeared, I'm writing a locale file which is en_US with year-month-day and metric units. In Linux, the locale source files go in /usr/share/i18n/locales/. Where do they go in DragonFly? Pierre -- gau do li'i co'e kei do
Re: OT: third party relay attack
On Friday, 10 January 2020 15.22.37 EST Matthew Dillon wrote: > I last looked at it a few years ago but there were numerous DNS based > services that you could use to test IP addresses and domains. But they > never worked well... they tended to block a lot of legitimate mail along > with the spam, and tended to always be out of date. I use several blacklists, some of which automatically delist an address after spam stops. I also set the reject code to 421 or 451, which tells the sender to retry until it gives up, instead of 5xx, which would tell it to give up immediately. That way, if a server is temporarily blocked, the spam run ends, and the server is delisted, legitimate mail gets through. I've seen it happen. Here's my current list of blacklists: reject_rbl_client dnsbl-1.uceprotect.net reject_rbl_client dnsbl-3.uceprotect.net reject_rbl_client psbl.surriel.com reject_rbl_client recent.spam.dnsbl.sorbs.net reject_rbl_client ix.dnsbl.manitu.net reject_rbl_client cbl.abuseat.org reject_rbl_client ubl.unsubscore.com reject_rbl_client spamsources.fabel.dk reject_rbl_client dev.null.dk reject_rbl_client dnsrbl.org reject_rbl_client truncate.gbudb.net reject_rbl_client rbl.interserver.net reject_rbl_client bl.nosolicitado.org reject_rbl_client hostkarma.junkemailfilter.com=127.0.0.2 I also greylist. That way, if a spammer sends me an email, my server tells him to try again. Fairly often, by the time he tries again, his IP address has been added to a blacklist. The Turkish spammer buys a netblock and a bunch of domains (all of the form ., e.g. enemyfree.co) and sends spam from each IP address and domain in turn. When he starts spamming from a new netblock, a few spams get through, until the entire netblock is included in a blacklist. Considering his tactics, I think a cluster of a few spams let through every few weeks is an acceptable error rate. As to the relay attack, it's someone trying to use my server as third party relay, not trying to send me spam through someone else's relay. Pierre -- When a barnacle settles down, its brain disintegrates. Já não percebe nada, já não percebe nada.
Re: OT: third party relay attack
On Friday, 10 January 2020 12.50.52 EST Matthew Dillon wrote: > There isn't a whole lot that can be done short of white-listing only > allowed originators and recipients. Most anti-spam services filter out > critical non-spam emails along with the spam. What do you mean by anti-spam services? What are some examples? What criteria do they use to block spam? Pierre -- Lanthanidia deliciosa: What the kiwifruit would be if it weren't so radioactive.
OT: third party relay attack
My mailserver is being attacked by what looks like a botnet since December 16 at 6:07 (11:07 UTC). Many hosts all over the world are sending mail purporting to be from many domains all over the world to a few domains in Russia. Most of the IP addresses are blocked by uceprotect.net; a few are blocked by other blocklists. A few are not blocked, but are rejected with "Relay access denied". The messages come at a rate of several per second. There are 133 emails stuck in leaf's mail queue, but they do not appear to be related to this attack. Pierre -- lo ponse be lo mruli ku po'o cu ga'ezga roda lo ka dinko
/var is read-only
I just noticed that the usual periodic email is not coming. I checked /var/ log/maillog and found that it's getting "Read-only filesystem" errors from maildrop several times a second. The computer has two disks. /boot, swap, and / are on one; /boot1, /var, and / crypt are on the other. /boot1 and /crypt are read-write (I created a file on each, then deleted it). Only /var is read-only (and /var/crash and /var/tmp are null-mounts from /, which is on the other disk). It is somehow writing huge reams of "Read-only filesystem" errors to a read-only filesystem. Attached is the snippet of /var/log/messages when it put /var/ into read-only mode. What should I do? Pierre -- gau do li'i co'e kei do Dec 27 03:01:21 zyxomma kernel: ahci0.0: TFES slot 16 ci_saved = fffd96c4 Dec 27 03:01:33 zyxomma kernel: ahci0.0: read NCQ error page slot=16 Dec 27 03:01:33 zyxomma kernel: ahci0.0: DONE log page target 0 err_slot=16 Dec 27 03:01:33 zyxomma kernel: ahci0.0: disk_rw: error fiscmd=0x60 @off=0x000cf150, 16384 Dec 27 03:01:33 zyxomma kernel: ahci0.0: Restart fffc96c4 Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): READ(10). CDB: 28 0 6 78 a8 0 0 0 20 0 Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): CAM Status: SCSI Status Error Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): SCSI Status: Check Condition Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): MEDIUM ERROR asc:0,0 Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): No additional sense information Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): Retrying Command (per Sense Data) Dec 27 03:01:33 zyxomma kernel: ahci0.0: TFES slot 27 ci_saved = 0ffdfec5 Dec 27 03:01:33 zyxomma kernel: ahci0.0: read NCQ error page slot=0 Dec 27 03:01:33 zyxomma kernel: ahci0.0: DONE log page target 0 err_slot=0 Dec 27 03:01:33 zyxomma kernel: ahci0.0: disk_rw: error fiscmd=0x60 @off=0x000cf150, 16384 Dec 27 03:01:33 zyxomma kernel: ahci0.0: Restart 0ffdfec4 Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): READ(10). CDB: 28 0 6 78 a8 0 0 0 20 0 Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): CAM Status: SCSI Status Error Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): SCSI Status: Check Condition Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): MEDIUM ERROR asc:0,0 Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): No additional sense information Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): Retrying Command (per Sense Data) Dec 27 03:01:33 zyxomma kernel: ahci0.0: TFES slot 0 ci_saved = 8139 Dec 27 03:01:33 zyxomma kernel: ahci0.0: read NCQ error page slot=28 Dec 27 03:01:33 zyxomma kernel: ahci0.0: DONE log page target 0 err_slot=28 Dec 27 03:01:33 zyxomma kernel: ahci0.0: disk_rw: error fiscmd=0x60 @off=0x000cf150, 16384 Dec 27 03:01:33 zyxomma kernel: ahci0.0: Restart efff8139 Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): READ(10). CDB: 28 0 6 78 a8 0 0 0 20 0 Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): CAM Status: SCSI Status Error Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): SCSI Status: Check Condition Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): MEDIUM ERROR asc:0,0 Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): No additional sense information Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): Retrying Command (per Sense Data) Dec 27 03:01:33 zyxomma kernel: ahci0.0: TFES slot 10 ci_saved = efff8ec4 Dec 27 03:01:33 zyxomma kernel: ahci0.0: read NCQ error page slot=2 Dec 27 03:01:33 zyxomma kernel: ahci0.0: DONE log page target 0 err_slot=2 Dec 27 03:01:33 zyxomma kernel: ahci0.0: disk_rw: error fiscmd=0x60 @off=0x000cf150, 16384 Dec 27 03:01:33 zyxomma kernel: ahci0.0: Restart efff8ec0 Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): READ(10). CDB: 28 0 6 78 a8 0 0 0 20 0 Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): CAM Status: SCSI Status Error Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): SCSI Status: Check Condition Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): MEDIUM ERROR asc:0,0 Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): No additional sense information Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): Retrying Command (per Sense Data) Dec 27 03:01:33 zyxomma kernel: ahci0.0: TFES slot 26 ci_saved = 07ff9ff4 Dec 27 03:01:33 zyxomma kernel: ahci0.0: read NCQ error page slot=12 Dec 27 03:01:33 zyxomma kernel: ahci0.0: DONE log page target 0 err_slot=12 Dec 27 03:01:33 zyxomma kernel: ahci0.0: disk_rw: error fiscmd=0x60 @off=0x000cf150, 16384 Dec 27 03:01:33 zyxomma kernel: ahci0.0: Restart 07ff8ff4 Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): READ(10). CDB: 28 0 6 78 a8 0 0 0 20 0 Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): CAM Status: SCSI Status Error Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): SCSI Status: Check Condition Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): MEDIUM ERROR asc:0,0 Dec 27 03:01:33 zyxomma kernel: (da0:ahci0:0:0:0): No additional sense information Dec 27 03:01:33
Re: Can't connect to IRC
On Friday, 13 December 2019 14.08.32 EST Max Herrgard wrote: > I use irc.du.se and it always works :) I added that and a few others I found on efnet.org. I still can't get on. Pierre -- lo ponse be lo mruli ku po'o cu ga'ezga roda lo ka dinko
Can't connect to IRC
This happens pretty often. I have no trouble with Freenode except that sometimes I get disconnected for no apparent reason, but I can't connect to EFnet for hours on end. The messages look like this: [04:38] [Info] Trying to connect to irc.paraphysics.net (port 6667) in 10 seconds. [04:39] [Info] Looking for server irc.paraphysics.net (port 6667)... [04:39] [Info] Server found, connecting... [04:39] [Info] Negotiating capabilities with server... [04:39] [Error] Connection to server irc.paraphysics.net (port 6667) lost: The remote host closed the connection. [04:39] [Info] Trying to connect to irc.choopa.net (port 6667) in 10 seconds. [04:39] [Info] Looking for server irc.choopa.net (port 6667)... [04:39] [Info] Server found, connecting... [04:39] [Info] Negotiating capabilities with server... [04:39] [Error] Connection to server irc.choopa.net (port 6667) lost: The remote host closed the connection. [04:39] [Info] Trying to connect to irc.teksavvy.ca (port 6667) in 10 seconds. [04:39] [Info] Looking for server irc.teksavvy.ca (port 6667)... [04:39] [Info] Server found, connecting... [04:39] [Info] Negotiating capabilities with server... [04:39] [Error] Connection to server irc.teksavvy.ca (port 6667) lost: The remote host closed the connection. Any idea how to fix it? Pierre -- When a barnacle settles down, its brain disintegrates. Já não percebe nada, já não percebe nada.
Re: One ssh hangs on exit, the other doesn't
On Thursday, 5 December 2019 03.48.18 EST Harald Brinkhof wrote: > Maybe some background processes keep running and prevent the X program from > sitting down? > > http://www.snailbook.com/faq/background-jobs.auto.html The examples at the top of that page don't make sense. If I start an xterm in an ssh session, and log out of the ssh session with the xterm still running, I expect ssh to keep running until I close the xterm. When I ssh into leopard, my local mail server, and forward the mail ports, the ssh session hangs on exit, because kmail still has the mail ports open. When I shelled into zyxomma, I exited each program before logging out of ssh. I don't understand why one ssh session closed and the other hung. Pierre -- loi mintu se ckaji danlu cu jmaji
One ssh hangs on exit, the other doesn't
I just added some text to the DocBook file of PerfectTIN, which is compiled by pandoc to PDF. Because the pandoc on my Linux box is older than that on my DragonFly box, I have to build it on DragonFly. So I typed "ssh -X zyxomma" in two konsoles, then entered the source tree in one and ran kwrite, and entered the build tree in the other and ran okular. After over an hour of typing and making, I closed kwrite and okular and hit control-D in both ssh sessions. The one that was running kwrite closed and returned me to the Linux prompt; the one that was running okular hung, and I had to kill it. Why? Pierre -- Por H o por B, los campos magnéticos se difieren dentro de un imán.
Re: Arena and threads
On Wednesday, 30 October 2019 15.26.01 EDT Matthew Dillon wrote: > Generally speaking the allocation algorithm depends on the language, the OS > doesn't have a lot to do with it. Anything linked against libc (that is, > any C program), however, will use the allocator in libc which we do > control. That allocator allocates per-thread zones. > > That said, the memory allocator itself, no matter what the language, is > still just allocating memory that is shared across all the threads. No > copying is involved. All the threads share the same address space. Any > locking is up to the program (or the language) itself. It's C++. My concern is whether there's a way to return the memory to the OS if it was allocated in a thread (say the program has been idle for a minute). In Linux, only the main thread can do that. The main thread allocates with brk; the other threads allocate with mmap. Pierre -- gau do li'i co'e kei do
Arena and threads
I'm writing a multithreaded program which produces a TIN from a point cloud. There's one UI thread, which spawns several worker threads (by default as many as the processor has), each of which works on the same data structure, consisting of points, edges, and triangles, with each triangle having a vector of dots. (For historical reasons, the corners of the triangles are called points, so I call the points in the point cloud dots.) During the conversion, points, edges, and triangles are only added, while dots are moved from one triangle to another, calling shrink_to_fit after the move. Qt also starts some threads, but I'm not concerned with them. To see how much RAM the program takes when run on a large point cloud, I ran the program, opened a previously saved TIN, cleared it, loaded a point cloud, converted it to a TIN, and did similar things while watching the task's memory allocation. I found it kept memory allocated after clearing the TIN, and loading and converting a point cloud after opening and clearing a TIN resulted in more memory allocated than loading and converting a point cloud in a fresh run of the program. I suspected a big memory leak, but found none. The reason turns out to be that each thread has its own arena. When you open or load a file, worker thread 0 does all the work, so the data structure is built in thread 0's arena. When you convert a point cloud to TIN, the point cloud, which was loaded into thread 0's arena, is moved into six triangles, still in thread 0's arena, then all threads create triangles, each in its own arena, but thread 0's allocation remains greater than the others. On my Linux laptop, the eight arenas are 64 MiB apart. There is no way to let go of freed RAM, except in the main thread's arena. I tried to check this on DragonFly, but instead of showing me each element of a map, the debugger just showed me the root node of the red-black tree. How does DragonFly handle allocation by multithreaded programs? Pierre -- li fi'u vu'u fi'u fi'u du li pa
Re: DragonFlyBSD Project Update - colo upgrade, future trends
On Thursday, 1 August 2019 11.51.32 EDT Steffen Nurpmeso wrote: > I say nonetheless, i do not have twitter nor facebook, nor will > ever have unless chip implant mandatory for not being expelled. > I see no problems with mail except patch review, but which does > neither hurt me, and then also can be worked around by placing the > patch inline. Setting up mailman is terrible, though, and it > seems Python 2 series becomes obsoleted, so that there a painful > step to Python 3 and the "new" mailman had to be achieved. Maybe > sr.ht (iirc) is a solution. After trying to set up Mailman and getting lost, I chose Sympa for the Bezitopo mailing list. Pierre -- When a barnacle settles down, its brain disintegrates. Já não percebe nada, já não percebe nada.
Re: DragonFlyBSD Project Update - colo upgrade, future trends
On Monday, 29 July 2019 13.10.27 EDT Constantine A. Murenin wrote: > Greylisting through PF spamd is also an option. Crater and leaf run Postfix, as do my mail servers. I use Postgrey. Pierre -- The gostak pelled at the fostin lutt for darfs for her martle plave. The darfs had smibbed, the lutt was thale, and the pilter had nothing snave.
Re: DragonFlyBSD Project Update - colo upgrade, future trends
On Monday, 29 July 2019 11.27.09 EDT Matthew Dillon wrote: > I feel that we do not have a choice here. Privately-run mail systems, in > general, are almost dead due to the spam load. I have to forward my own > personal domain email through GMail just to be able to continue using it > and my GMail spam mailbox consistently contains more than 3000 spams in it > (30-day expiration, so ... 100+ spams per day). And that doesn't count the > ones Google auto-deletes immediately or the ones my smtp server discards. > I've tried everything possible to keep my personal domain and dragonfly's > domain email usable but its an impossible task. On my mailserver (bezitopo.org), I use greylisting combined with IP address blacklists. Very little spam gets in. A few times a year I get several spams a day, because a spammer has bought a new netblock, but even when that happens, much of the spam is blocked because, between the first try from an IP address (which is blocked by greylisting) and the second, the IP address landed in a blacklist. Most of the spam I get comes through leaf. I can't block it, because my mail server sees leaf's IP address, not the spammer's. I filter some of it to a spam folder, but that's not as effective as blocking IP addresses. Moving to a web forum is not going to make spam go away. They get spam too. Pierre -- Don't buy a French car in Holland. It may be a citroen.
Re: How to get the thread number and every thread's ID of a running process?
On Tuesday, 25 June 2019 21.11.07 EDT Nan Xiao wrote: > Hi all, > > Greeting from me! > > I am a new user of DragonflyBSD, and want to know if there is some > method to get the thread number and every thread's ID of a running > process? I try to explore "/proc/$pid" directory, but can't find > related information. I also can't find out-of-box API to fulfill these > functions. > > Could anybody give some hints? Thanks very much in advance! If the process is on top, type "H" while running top. Type "ps aux -H". The output is wider than 80 characters, so hit the maximize button on your xterm first. Pierre -- Jews use a lunisolar calendar; Muslims use a solely lunar calendar.
Re: Computers for cross-platform development
On Friday, 24 May 2019 12.30.21 EDT Gerald Henriksen wrote: > The NUCs are reasonably standard Intel stuff, but I suspect it will > come down to asking on the various forums for those BSDs with a > specific model, or buying one machine and try installing them all and > see what happens. Sounds good. I think I'm going for NUC. > It will also depend on whether you need all the hardware to work or > not. What do you mean by all the hardware? What might not work? On Friday, 24 May 2019 12.39.01 EDT Justin Sherrill wrote: > Get a single large machine and virtualize each one of the > environments. It won't be as fast, but that may not matter if you are > only building, say, weekly. Two of my computers (one holds all my email and the other all my local Git repos) are over six years old. Neither has suffered a hard drive failure, but I just replaced a failed drive on my laptop. I need at least two new computers to take over if they fail. I also need a computer (which could be one of them) to drive a plotter; the new computers and the plotter will be in the other room. On Friday, 24 May 2019 22.00.19 EDT Zachary Crownover wrote: > I double the sentiment for virtualization if your intent is a continuous > integration system for testing and artifact building. As Justin mentioned, > the frequency of the runs adjusts the value of VM vs physical. I don’t know > if Travis CI supports every OS on your list, but it might and you could set > up a testing framework with it. If it’s open source work it could also be > free through them and you wouldn’t have to worry about computer costs at > all. Travis does only Linux and macOS, its Linux is Ubuntu Trusty and Xenial (I'm running Bionic, and I think Trusty is too old for the version of CMake I require), and AFAICT it doesn't have big-endian. Pierre -- sei do'anai mi'a djuno puze'e noroi nalselganse srera
Computers for cross-platform development
I'm planning to get several small computers so that I can test my software on several OSes. One will compile binaries for Windows; the others will run OpenBSD, NetBSD, probably FreeBSD, and maybe DragonFly (I already have a DragonFly box, but it's slow compared to my laptop). I'm looking at these: https://www.newegg.com/Barebone-Mini-Computers/Category/ID-3 How can I make sure that all the hardware works on all the BSDs? The reason I say "probably FreeBSD" is that another computer I'm going to get is a Power9 box from Raptor for big-endian testing. I know of two OSes that are big-endian on Power9: Adélie Linux and FreeBSD. However, while it is Tier 1 on Adélie, it is Tier 2 on FreeBSD. So I may or may not set it up as dual- boot. Pierre -- I believe in Yellow when I'm in Sweden and in Black when I'm in Wales.
Re: What programming languages and operating systems that will to be used after Jesus to return?
On Saturday, 13 April 2019 16.20.10 EDT Quantum Robin wrote: > Hello! > > Apocalipse 13: 15-18 > > What programming languages and operating systems that will to be > used after Jesus to return? > > In other words, what are the programming languages and operating systems > that will to be used in the Antichrist's rule? Jesus's return is shortly after Antichrist's rule, and is followed (I believe) by the Millennium, which means there's also the question of what computer software will be used during the Millennium. There are beasts mentioned in Revelation, and (fishing) nets, open (doors), and free (people) mentioned in various parts of the Bible. But there are no dragonflies. Dragons are mentioned in Revelation, and flies are mentioned among the plagues leading up to the Exodus, but the word "μυια" (fly) does not occur in the Greek NT. Python is mentioned in Acts 16:16, but it is not a programming language or even a snake, but a spirit used by oracles. Snow is mentioned, but no Snobol AFAIK. Julia is mentioned in Romans 16, but again she is not a programming language. There is not enough information in the Bible to answer the question. But I would be very interested in what kinds of computers we will use in the Olam haBa, which will be in a new universe (Rev. 21:1) with completely different physical laws (Rev. 21:16). Pierre -- Lanthanidia deliciosa: What the kiwifruit would be if it weren't so radioactive.
Re: /compat/linux/proc
On Friday, 8 March 2019 09.38.38 EST Aleksej Lebedev wrote: > Well, it seems to work. Sounds good. > In fact, I think you can. I only have dragonfly on my server at Hetzner, > so I installed geogebra there and run it using ssh X11-forwarding. You > can too. I can't run it here for two reasons: * I can't reach zyxomma directly from this hotel. I have to go through marozi, which does not have X. * Zyxomma is on a satellite connection, which adds about 600 ms of latency. When I'm home, mooncat and zyxomma are on the same LAN, and I can easily run X programs over ssh. Pierre -- Por H o por B, los campos magnéticos se difieren dentro de un imán.
Re: /compat/linux/proc
On Friday, 8 March 2019 07.39.43 EST Aleksej Lebedev wrote: > I don't know anything about GeoGebra, but there is geogebra-5.0.513.0 in > the official DragonFly repo. > You can try it with > > sudo pkg install geogebra > > Hope this helps. And when you run it, what happens? I already installed it. I can't try *running* it until I get home. The "Message from geogebra" was the result of running "pkg install geogebra". Pierre -- La sal en el mar es más que en la sangre. Le sel dans la mer est plus que dans le sang.
Re: /compat/linux/proc
On Friday, 8 March 2019 04.28.55 EST Antonio Huete Jiménez wrote: > Hi Pierre, > > DragonFly has only ever supported 32-bit Linux emulation and that was > removed a while ago. Is there a way to get GeoGebra to work? I can't test it until I get home, which will be late Sunday. Pierre -- The Black Garden on the Mountain is not on the Black Mountain.
/compat/linux/proc
I am at a math conference and got a book "Exploring Advanced Euclidean Geometry with GeoGebra". Since the version on my laptop (4.0.34 for Bionic) does not work (known bug), I installed it on zyxomma, and got this message: Message from geogebra-5.0.449.0: This software wants linprocfs mounted on /compat/linux/proc. Add the following line to /etc/fstab: linprocfs /compat/linux/proc linprocfs rw 0 0 I added the line and then did the following: # ls /compat ndis # mkdir /compat/linux # mkdir /compat/linux/proc # mount /compat/linux/proc mount: exec mount_linprocfs not found in /sbin, /usr/sbin: No such file or directory Where do I get mount_linprocfs? I have two files by that name, but one's a 32- bit version from my previous DragonFly box, darner, and the other is for DragonFly 4.0.100. Pierre -- li fi'u vu'u fi'u fi'u du li pa
Can't shell into leaf on IPv6
$ ssh -6 leaf.dragonflybsd.org ssh_exchange_identification: read: Connection reset by peer $ ssh -4 leaf.dragonflybsd.org Last login: Sat Nov 3 18:31:40 2018 from 2001:5b0:211c:3 This happens on both DragonFly and Linux. Any idea why? Pierre -- Por H o por B, los campos magnéticos se difieren dentro de un imán.
Re: desktop problem
On Wednesday, 20 February 2019 01.44.26 EST Seann Aswell wrote: > Of course if anyone has actually attempted to get KDE5 to work on > DragonFly, they will likely chime in... It's been some months since I actually used KDE on DragonFly, but it is still installed and I should be able to get it up. The last reason I logged into DragonFly (as opposed to ssh-ing into it) was to ssh into my Linux laptop to kill a program that was grabbing all the memory it could and making it impossible to even move the mouse. I got KDE running by installing a login manager (namely sddm). sddm is installed in /usr/local/, so its startup script (as are others) is in /usr/ local/etc/rc.d/. I just tried to enable it with rcrun, but it crashed. I'll be away for a few days and won't be able to fix it till I get back. I have kwrite running on DFly over ssh. The KDE version is 5.47.0 (5.44.0 on Linux). I'm not sure what package you have to install to get the KDE desktop, but it may be one of the plasma packages. Pierre -- Don't buy a French car in Holland. It may be a citroen.
Re: DragonFly on Power9?
On Sunday, 13 January 2019 20.56.08 EST Gerald Henriksen wrote: > Why do you want to do big-endian testing? One of the Bezitopo programs reads geoid files in various formats and converts them to a common format so that the main Bezitopo program doesn't have to read many different formats. One of the input formats is the NGS format: https:// www.ngs.noaa.gov/GEOID/GEOID12B/GEOID12B_data.shtml . "The binary formats of the Unix format data files follow the IEEE Std. 754 binary format (big- endian), as they are developed in a Unix environment." To make sure that convertgeoid can read both endians, I got Alaska, Guam, and Samoa in big- endian and Hawaii, Puerto Rico, and Lower 48 in little-endian. But I haven't run the program on a big-endian machine. > Big-endian is essentially dead, with Fedora (which for years has > offered both BE and LE Power) dropping big-endian support due to both > a lack of resources and the fact that new Power based projects are LE. > > Red Hat EL 8 is LE only on Power. > > And IBM itself started the process to switch Power to being a LE only > platform due to client/vendor desires 4 years ago: > > https://www.ibm.com/developerworks/community/blogs/fe313521-2e95-46f2-817d-4 > 4a4f27eba32/entry/just_the_faqs_about_little_endian?lang=en That page states "No, there will be no effect on AIX or IBM i application environments as a result of this change." AIX on Power is big-endian. AIX is a Unix and should therefore run Bezitopo (and Quadlods and any other software I write on Linux or BSD) without change. Pierre -- Lanthanidia deliciosa: What the kiwifruit would be if it weren't so radioactive.
DragonFly on Power9?
I recently found out about the Talos and upcoming Blackbird computers from Raptor (https://www.raptorcs.com/) and am thinking of getting one for big- endian testing. Talos comes with Debian, which is little-endian, but Adélie Linux and FreeBSD are big-endian and can run on it. Any thoughts on porting DragonFly to Talos? Pierre -- loi mintu se ckaji danlu cu jmaji
Re: Coupla questions on DragonFly as a desktop OS
On Sunday, 28 October 2018 04.04.39 EDT John Long wrote: > Hi, > > I have some miscellaneous questions on the way to finding a new desktop > OS. Thanks to anybody who would help with this: > > Acrobat reader: I have a lot of manuals in .pdf format and I haven't > found anything I can tolerate other than acroread. I'm not saying there > isn't anything good, I just haven't seen it. I don't need forms support > or change tracking or any fancy stuff. I want 2-up display and I need > internal links to work. I also have a bunch of .djvu but the djview > reader is mostly ok. I use okular, which is available in dports. It's not in Ravenports yet, because KDE isn't, but it is in dports. Pierre -- lo ponse be lo mruli ku po'o cu ga'ezga roda lo ka dinko
Re: Where can I buy a Ryzen box?
On Tuesday, 11 September 2018 00.46.25 EDT Gerald Henriksen wrote: > OpenCL would depend on if that has been ported to Dragonfly or not, > but it at least in theory can work with the AMD open source drivers, > Intel drivers, and on the CPU. Both DPorts and Ravenports have both opencl and libclc packages. > The Ryzen with APU (Raven Ridge codename) is likely just too new > having only been released in February and even as late as May there > were still a lot of stability issues on Linux (when using the APU). So > without Dragonfly support for the APU you would need a separate, > supported video card in order to get anything other than perhaps text. > > Given the lack of Raven Ridge support you may want to consider pricing > out getting a plain Ryzen CPU + separate GPU, may either save you > money or allow to get a better GPU for the same money unless you > specifically want the ability to drop back to the APU in the future if > the drivers get ported to Dragonfly. Is the APU part of the CPU or a separate chip? Does the Threadripper come with an APU? Pierre -- ve ka'a ro klaji la .romas. se jmaji
Re: Where can I buy a Ryzen box?
On Monday, 10 September 2018 11.39.56 EDT Matthew Dillon wrote: > If you want to use the GPU for computation you may have to run Linux. In > fact, you probably would have to. Could I do it with the CUDA compiler in Julia (which of course requires getting Julia to work on DragonFly)? Or could I do it with OpenCL? Michael Neumann told me off-list that DFly doesn't support the built-in GPU of the Ryzen. Does that mean than I have to get a separate GPU even if I can't program it? > I usually just buy the parts on Amazon or NewEgg, and throw the computer > together, but if you don't have experience doing that sort of thing its > easy to get the wrong bits. Are the wrong bits bits that don't work with DragonFly, or bits that don't work with each other? If the latter, the guys at the computer shop should be able to advise me. Pierre -- Lanthanidia deliciosa: What the kiwifruit would be if it weren't so radioactive.
Where can I buy a Ryzen box?
I'm thinking of getting some new computers in a few months: * A System 76 Leopard, with lots of cores and GPU, but not Ryzen * A Ryzen box, with GPU for computation, running DragonFly * A few lesser boxes with other OSes. I doubt that anyone sells Ryzen boxes preloaded with DragonFly, so I'm trying to find a Ryzen box without an operating system. But googling "ryzen no operating system" brings up computers that have Ryzen and an operating system, but no something else. Where do you suggest I look? Pierre -- Jews use a lunisolar calendar; Muslims use a solely lunar calendar.
Re: Ravenports, what's missing for you?
I see Erlang is in Ravenports. I'd like to see also Haskell (which is in dports as "ghc") and Julia. I just upgraded my laptop from Artful Aardvark to Bionic Beaver. Artful has an old version of Julia. Bionic doesn't have Julia; I had to download the binary. I tried compiling the Github repo, but apparently am missing some dependency. Pierre -- ve ka'a ro klaji la .romas. se jmaji
Postfix users, heads up
I upgraded from 5.0 (compiled 2017-10-17) to 5.2 on Friday. Sometime today, my laptop spontaneously half-rebooted, so I tried to shell into it from my DFly box and found that Postfix was giving an error every minute. It now requires that there be an smtpd_relay_restrictions or smtpd_recipient_restrictions (I had neither) with a certain clause. I copied the parameter from one of my other mailservers, deleted the check_client_access and check_recipient_access clauses (zyxomma is accessible only via IPv6, the address changes once a thunderstorm, and all the emali it gets is forwarded from another box), and was up and running again. Pierre -- Jews use a lunisolar calendar; Muslims use a solely lunar calendar.
Re: DragonFly on VirtualBox
On Friday, 2 March 2018 20.54.08 EST Aaron LI wrote: > Hi Pierre, > > You're using EFI with VirtualBox to boot the DFly guest, which I haven't > tried. > > For EFI boot, you need to make sure that file "/EFI/boot/bootx64.efi" file > exists on your EFI partition (i.e., ad0s0), which is the default boot code. > If that file doesn't exist, you can copy "/boot/boot1.efi" over. > > You can also refer to the example installation script: > /usr/share/examples/rconfig/hammer_uefi.sh > > But I'm not sure whether VirtualBox's EFI implementation works with DFly's > EFI boot loader... The file is there, but it's /EFI/BOOT/BOOTX64.EFI, not /EFI/boot/bootx64.efi. Is there a way in VirtualBox to install DragonFly without EFI? Pierre -- The Black Garden on the Mountain is not on the Black Mountain.
Re: DragonFly on VirtualBox
On Friday, 2 March 2018 00.24.48 EST Aaron LI wrote: > I've installed DragonFly BSD (4.8, 5.0, and master) in VirtualBox (on Debian > Linux) using the installer, and all work well. > > Maybe you can try: > > 1. Boot into DFly through the ISO > 2. Find the disk into which DFly is installed, e.g., "ad0" (for me) or "da0" > 3. Install the boot code: fdisk -B /dev/ad0 (change to your disk) >(According to the installer, "fdisk -B" is used instead of "boot0cfg", >which installs the /non-interactive/ boot code as described in the >EXAMPLES in boot0cfg(8). >Could somebody explain in more detail about these two boot codes??) > 4. Try reboot from the disk (you may need adjust the booting orders) > > Hope this helps solve your problem. fdisk -B results in "Missing operating system". The disk is laid out like this, according to file: ad0: boot sector, pt 1 type ee ad0s0: boot sector, 32-bit FAT, label "EFI" ad0s1: boot sector ad0s1a: FFS /boot ad0s1b: swap ad0s1d: HAMMER / Pierre -- ve ka'a ro klaji la .romas. se jmaji
DragonFly on VirtualBox
On my new laptop (which came with Linux) I made two VirtualBoxen and installed DragonFly on one and OpenBSD on the other. The OpenBSD install worked fine. The DragonFly came up with "Missing operating system" or the like. I rebooted it from the ISO file and ran boot0cfg. It looks like all files are installed, but when I rebooted it this time, it came up "F1 ??" and wouldn't boot. I think it's supposed to say "F1 DF/BSD". Did I skip a step, or am I missing a file or something? Pierre -- loi mintu se ckaji danlu cu jmaji
Spam sent to leaf
Every so often, I get a spam sent to my leaf account. I checked the last one, which is SEO spam. It was sent from 202.160.134.13. I tried to connect to port 25 and got a timeout, so apparently it is not running a mail server. This sort of spam can be easily blocked with greylisting, which I use on my mail servers. I see that Postgrey is in the repo. Is anyone else getting spam on leaf addresses? Would it be good to greylist it? Pierre -- Jews use a lunisolar calendar; Muslims use a solely lunar calendar.
Kmail version
I just got my laptop back. (The on button circuit broke down.) On caracal (the laptop), I am running: *Ubuntu Xenial Xerus *Linux 4.4.0 *Kmail 5.1.3 *KDE 4.14.16. On zyxomma, I'm running: *DragonFly BSD 5.0.0.2 *Kmail 4.14.10 *KDE 4.14.30. Xenial was a year and a half old when Artful Aardvark came out. DragonFly 5.0 was released last October. Why is Kmail newer on Xenial? Pierre -- The Black Garden on the Mountain is not on the Black Mountain.
Small form factor with two NICs and/or wireless
My laptop caracal just conked out and I've ordered a new one from System 76. (This is the same laptop that I brought to Berkeley years ago to see if Matt Dillon could load DragonFly on it. He couldn't, so it's been running Ubuntu ever since.) My other boxen leopard (Ubuntu, May 2012) and zyxomma (DFly, April 2013) are slower than caracal. Both are small-form-factor boxes from Aleutia. System 76 sells a NUC with Core i7 (same as caracal, and same as the new laptop), but NUCs appear to have only one Ethernet port, and one reason I got the model I did for zyxomma (which they no longer make) is that it has two Ethernet ports and thus can act as a gateway. Only five devices (one of which is the phone) can get IPv4 addresses. I'm thinking of setting up some more computers in the other room, so I'll need a gateway. Do you know of some small-form-factor computers with two Ethernet ports, wireless which I can set up as an access point, and comparably fast multicore processors which work well on both Linux and DragonFly? Pierre -- La sal en el mar es más que en la sangre. Le sel dans la mer est plus que dans le sang.
Could you check the build instructions?
http://bezitopo.org/developers.html I'm planning to release another version of Bezitopo in about three weeks. I wrote the build instructions while building it on a fresh Cygwin install and extrapolated the packages needed on other OSes. I've had KDE on my DragonFly box for many years, so I already had Qt when I made the GUI. Also, I don't have FreeBSD and am guessing that the packages are the same. Could someone who doesn't have Qt5 installed, or who runs FreeBSD, check the list of packages? Pierre -- I believe in Yellow when I'm in Sweden and in Black when I'm in Wales.
Re: DragonFly 5.0 released!
On Monday, October 16, 2017 11:25:13 AM EDT Matthew Dillon wrote: > Hi folks! So 5.0 is out the door, and this is probably going to be our > stable release for the next few months. Master is about to get a pile of > commits and should be considered a bit more experimental over the next few > months than it would normally be. So if you normally run master, but you > want stability, we recommend sticking with the 5.0 release for a few months. > > Now that 5.0 is out the door, master is going to start getting some really > fun performance commits for beefier systems with lots of cores. However, > the most important of these commits is also relatively dangerous as it > involves avoid global IPIs by not fully synchronizing the page tables > backing the buffer cache. Any corruption in that area of the kernel can > directly effect filesystem integrity, which is why I am posting this > warning now. I just tried "pkg search" and got this error: pkg: Repository Avalon missing. 'pkg update' required pkg: Repository Avalon load error: access repo file(/var/db/pkg/repo- Avalon.sqlite) failed: No such file or directory pkg: http://mirror-master.dragonflybsd.org/dports/dragonfly:4.10:x86:64/ LATEST/meta.txz: Not Found pkg: http://mirror-master.dragonflybsd.org/dports/dragonfly:4.10:x86:64/ LATEST/packagesite.txz: Not Found I'm running 4.9 compiled August 8. Do I have to upgrade in order to get any packages? Pierre -- I believe in Yellow when I'm in Sweden and in Black when I'm in Wales.
Re: Encrypt your home folder after installation
On Sunday, October 15, 2017 1:05:50 AM EDT Olle Harstedt wrote: > Hi, > > Thanks for your excellent reply. > > OK, I just accepted the default partition settings during > installation, and it seems it did not create a home folder, so I guess > I'm screwed unless I can shrink the root partition without formating > it. Possible? Otherwise, a reinstallation might be my best option > here. (Except that I physically have to move the hard-drive since dfly > won't read USB on the X220 laptop ><) As far as I can tell, this is not possible. Can you add another hard disk? > A related question: Is it possible to encrypt a HAMMER PFS? You mean filesystem encryption (as opposed to block encryption, which is what cryptsetup does)? I don't see anything about it in the man page. Pierre -- Jews use a lunisolar calendar; Muslims use a solely lunar calendar.
Re: Encrypt your home folder after installation
On Thursday, October 5, 2017 12:03:02 AM EDT Olle Harstedt wrote: > Hi dfly! > > I've installed dfly and wish to encrypt my home folder. I discussed it > breifly on #dragonflybsd, and it seems like the best option is to > encrypt the whole disk *during* installation, not after. Are there any > other options? Are you encrypting just /home, or the whole disk? It's been a long time since I've installed any OS. On zyxomma (my DragonFly box) I have a plaintext slice, which is /, /home, and most other filesystems, and a ciphertext slice, which is /crypt and /usr/obj, which I put as a PFS on /crypt because I was once compiling lots of stuff at once and /crypt is bigger. On my Linux laptop and mailserver, I do have /home encrypted, but /usr is plaintext. No matter how you set up disk encryption, /boot must be plaintext, because that's where the encryption code is stored. Pierre -- ve ka'a ro klaji la .romas. se jmaji
libgd conflicts with uk-libgd
On installing gnuplot, I got a problem. pkg check -d reported this: Checking all packages: 100% apache24 is missing a required shared library: libaprutil-1.so.0 apache24 is missing a required shared library: libapr-1.so.0 apache24 is missing a required shared library: libdb-5.3.so.0 gcc-aux is missing a required shared library: libmpfr.so.4 gcc-aux is missing a required shared library: libmpc.so.3 gcc47 is missing a required shared library: libmpfr.so.4 gcc47 is missing a required shared library: libmpc.so.3 gnuplot is missing a required shared library: libgd.so.6 libva has a missing dependency: libGL llvm36 is missing a required shared library: libedit.so.0 pulseaudio is missing a required shared library: libXtst.so.6 pulseaudio is missing a required shared library: libsndfile.so.1 pulseaudio is missing a required shared library: libltdl.so.7 xdriinfo has a missing dependency: libGL >>> Missing package dependencies were detected. >>> Found 1 issue(s) in the package database. pkg: No packages available to install matching 'libGL' have been found in the repositories >>> Summary of actions performed: libGL dependency failed to be fixed >>> There are still missing dependencies. >>> Try fixing them manually. >>> Also make sure to check 'pkg updating' for known issues. pkg updating has a notice from four years ago about libGL which says to use portmaster. Is it still valid? Should I use portmaster? When installing all the packages for gnuplot, it installed uk-libgd, but that conflicts with libgd. Which one should I have installed? Pierre -- Don't buy a French car in Holland. It may be a citroen.
Re: New experimental statistics collector in master
On Saturday, July 29, 2017 10:54:42 PM EDT Matthew Dillon wrote: > If you have gnuplot installed (pkg install gnuplot), kcollect can generate > some useful graphs based on the in-kernel data. Well, it will be boring if > the machine isn't doing anything :-). There are options to use gnuplot to > generate a plot window in X or a .jpg or .png file, and other options to > set the width and height and such. At the moment the gnuplot output uses a > subset of statically defined fields to plot but ultimately the field list > it uses will be specifiable. Typo in the man page: -s Smooth plot output (only applicable to gunplot). Do you mean "applicable only to gnuplot"? Pierre -- Don't buy a French car in Holland. It may be a citroen.
Re: Packages for 4.9
On Thursday, June 29, 2017 1:07:26 PM EDT karu.pruun wrote: > The option "-f" is at the end: > > % pkg upgrade -f After doing that a few times, so that it no longer downloads or upgrades any packages, it does this: # pkg upgrade -f Updating Avalon repository catalogue... Avalon repository is up to date. All repositories are up to date. Checking for upgrades (1024 candidates): 38% llvm36 has no direct installation candidates, change it to llvm35? [Y/n]: y Checking for upgrades (1024 candidates): 45% liborcus07 has no direct installation candidates, change it to liborcus? [Y/ n]: Checking for upgrades (1024 candidates): 83% gcc47 has no direct installation candidates, change it to gcc? [Y/n]: Checking for upgrades (1024 candidates): 100% Processing candidates (1024 candidates): 100% pkg: sqlite error while executing UPDATE packages SET name=?1 WHERE name=?2; in file pkg_jobs.c:1714: UNIQUE constraint failed: packages.name pkg: sqlite error while executing UPDATE packages SET name=?1 WHERE name=?2; in file pkg_jobs.c:1714: UNIQUE constraint failed: packages.name pkg: sqlite error while executing UPDATE packages SET name=?1 WHERE name=?2; in file pkg_jobs.c:1714: UNIQUE constraint failed: packages.name Checking integrity...Assertion failed: (strcmp(uid, p->uid) != 0), function pkg_conflicts_check_local_path, file pkg_jobs_conflicts.c, line 386. Child process pid=7381 terminated abnormally: Abort trap Pierre -- Jews use a lunisolar calendar; Muslims use a solely lunar calendar.
Re: Packages for 4.9
On Thursday, June 29, 2017 2:08:47 AM EDT Pierre Abbat wrote: > I got "illegal option -- f". I'm letting it run, but then I'll try the > command again with -f. It still says "illegal option". What is the correct option? Pierre -- When a barnacle settles down, its brain disintegrates. Já não percebe nada, já não percebe nada.
Re: Packages for 4.9
On Monday, June 26, 2017 9:33:30 AM EDT Francois Tigeot wrote: > The only unusual thing you will have to do is use pkg -f upgrade, > the -f flag is necessary because recent versions of pkg(8) failed > to properly record shared libraries in packages I got "illegal option -- f". I'm letting it run, but then I'll try the command again with -f. Pierre -- loi mintu se ckaji danlu cu jmaji
Packages for 4.9
I'm planning to upgrade my 4.5 box, which has been doing its thing for over a year, to 4.9. Are packages built for 4.9? Is there anything unusual I have to do when upgrading? Pierre -- The Black Garden on the Mountain is not on the Black Mountain.
Bezitopo on DragonFly
I'm in the process of the first release of Bezitopo. It used to compile on Raspbian Wheezy but now requires at least Jessie. So I'd like to know, and indicate on the website, what is the minimum version of DragonFly necessary to run it. I have 4.5.0 and it works fine. If you have older, would you try compiling and running it? Bezitopo website: http://bezitopo.org Code repository: https://github.com/phma/bezitopo/ Once you've Cmade it, type: $ make -j 5 && ./bezitest (the number after -j depends on your cores and hyperthreads) Pierre -- When a barnacle settles down, its brain disintegrates. Já não percebe nada, já não percebe nada.
Re: UTF-8
On Thursday, February 04, 2016 17:12:21 Vasily Postnicov wrote: > As a reminder: Don't forget to run cap_mkdb after any changes to > /etc/login.conf Thanks, that fixed it. Pierre -- lo ponse be lo mruli ku po'o cu ga'ezga roda lo ka dinko
Re: UTF-8
On 02/ 4/16 11:38 AM, Dmitrij D. Czarkoff wrote: Pierre Abbat said: Once I do that, what else do I have to do? Do I have to put something in /etc/profile? Check output of "locale". I put "en_US.UTF-8" in /etc/login.conf, logged in on a VT, ran locale, and found LANG set to "en-US.UTF-8". What gives? I manually set LANG to "en_US.UTF-8" in a konsole, opened the Bezitopo source in kate, and found that all the files are displayed in ISO-8859-1. What else do I have to set? There are degree signs and other non-ASCII characters in comments, all encoded in UTF-8. There is no mention of UTF in /etc/profile , ~/.profile , or any other configuration file in my home directory that I could think of. Pierre
Re: UTF-8
On 02/ 4/16 07:00 AM, Romick wrote: On Wed, Feb 03, 2016 at 01:39:32PM +0100, Dmitrij D. Czarkoff wrote: Pick one lang per class. Yeah, and there is no such locale as ``en-US.UTF-8'', use ``en_US.UTF-8'' (eq. just remove second lang=...) :) Does it matter where I put the line? Should I change "ru_RU.KOI8-R" to "ru_RU.UTF-8"? Or should I remove the Russian class? I don't know why it's there. Once I do that, what else do I have to do? Do I have to put something in /etc/profile? Pierre
Re: UTF-8
On 02/ 3/16 07:29 AM, Romick wrote: On Tue, Feb 02, 2016 at 01:57:22PM +, Pierre Abbat wrote: I already had this in login.conf : default:\ :passwd_format=sha512:\ :copyright=/etc/COPYRIGHT:\ :welcome=/etc/motd:\ :setenv=MAIL=/var/mail/$,BLOCKSIZE=K,FTP_PASSIVE_MODE=YES:\ :path=/sbin /bin /usr/sbin /usr/bin /usr/games /usr/local/sbin /usr/loc$ :nologin=/var/run/nologin:\ :lang=en_US.UTF-8:\ :lang=en-US.UTF-8:\ ^ remove lang=en-US.UTF-8 please (the last one, with en-US) There are now no occurrences of "lang=en-US.UTF-8" in /etc/login.conf . Should I also remove the "lang=ru_RU.KOI8-R" line? Pierre
Re: UTF-8
On 02/ 2/16 04:11 PM, Dmitrij D. Czarkoff wrote: Pierre Abbat said: Should I put something in /etc/profile, or what? Could you paste output of "locale" command? LANG="en-US.UTF-8" LC_CTYPE="C" LC_COLLATE="C" LC_TIME="C" LC_NUMERIC="C" LC_MONETARY="C" LC_MESSAGES="C" LC_ALL=""
Re: Valgrind
On 02/ 2/16 06:22 PM, Vasily Postnicov wrote: Do you mean an uninitialized pointer problem? Do you have a source code of that software that you think causes a trouble? If you think your memory is corrupted because of uninitialized pointer, I suggest to compile problematic code with clang with AddressSanitizer. Just don't forget to compile llvm with compiler-rt option. I hope this helps. Didn't try this by myself, though It's on Github. Follow the links from http://bezitopo.org/. It's not an uninitialized pointer, it's a double variable, which is a control point of a Bézier triangle. The function is returning something like 1e-300 instead of 280. I inserted a debugging statement to output "ctrlpt garbage" if it ever returned such a number. On Linux, inserting the statement made the bug go away, but it reappeared on DragonFly. I ran bezitest under Valgrind a few weeks ago and fixed some uninitialized variables and allocation mixups. Pierre
Re: Valgrind
On 02/ 2/16 08:19 PM, Vasily Postnicov wrote: Let me correct my answer a bit. You calculate 'which' as a sum of floating point comparisons with some "weight" from 1 to 8, but this is ok, as a result of comparison is 0 or 1. But, there might be a problem. The greatest possible 'which' is the sum of "weights" 1+2+3+4+8+12=30. (Assuming all conditionals return 1). Your ctrltab array has only 16 elements. So if one floating point conditional wrongly returns 1 (due to lack of precision) you will get out-of-bounds error. I think this is the most probable scenario Thanks! I put a breakpoint on the "ctrlpt garbage" line and found that which was 23. I then corrected the line crit=1/2.0*(1/dist(xy(*a),xy(*b))+1/dist(xy(*c),xy(*a))+1/dist(xy(*b),xy(*c))); to crit=1/(2.0*(1/dist(xy(*a),xy(*b))+1/dist(xy(*c),xy(*a))+1/dist(xy(*b),xy(*c; which fixed the problem. The bug showed up on triangles much smaller than a meter. Pierre
Re: UTF-8
On 02/ 1/16 07:38 PM, Thomas Nikolajsen wrote: Well. Isn't there login.conf? Yes, see manual, 'man login.conf' or online, https://www.dragonflybsd.org/cgi/web-man?command=login.conf=ANY I already had this in login.conf : default:\ :passwd_format=sha512:\ :copyright=/etc/COPYRIGHT:\ :welcome=/etc/motd:\ :setenv=MAIL=/var/mail/$,BLOCKSIZE=K,FTP_PASSIVE_MODE=YES:\ :path=/sbin /bin /usr/sbin /usr/bin /usr/games /usr/local/sbin /usr/loc$ :nologin=/var/run/nologin:\ :lang=en_US.UTF-8:\ :cputime=unlimited:\ :datasize=unlimited:\ :stacksize=unlimited:\ :memorylocked=unlimited:\ :memoryuse=unlimited:\ :filesize=unlimited:\ :coredumpsize=unlimited:\ :openfiles=unlimited:\ :maxproc=unlimited:\ :posixlocks=unlimited:\ :sbsize=unlimited:\ :vmemoryuse=unlimited:\ :lang=en-US.UTF-8:\ :priority=0:\ :ignoretime@:\ :umask=022: Should I put something in /etc/profile, or what? Pierre
UTF-8
How can I set up DragonFly so that for all users, all files, the shell, and processes come up in UTF-8 by default? I have files written in English and Russian, files with Hebrew names, and the like. Pierre
Kmail on DragonFly
Do any of you use Kmail on DragonFly? My laptop (which runs Ubuntu) is in the shop, so I'm using zyxomma. When my connection drops, it throws up the KWallet dialog box, as if the password stored in the wallet were wrong. There's nothing wrong with the password; the connection went down. Sometimes it gets stuck and displays "Retrieving folder data" when I click on a message, and never gets around to showing the message. I've also had problems with Akonadi not starting, though it starts correctly now. Pierre
McCabe complexity
I've been working on a land surveying program for over eight years. According to pmccabe, the most complicated functions are pointlist::maketin (62), which creates a triangulated irregular network, and triangle::subdivide (58/65), which subdivides a triangle into monotonic regions. The TIN uses a tangledly hierarchical data structure called "winged edge", in which points, edges, and triangles all point to each other. It took me years to make it work right. I checked HAMMER source code with pmccabe. The most complex function is hammer_vfs_mount, at 51. In hammer2, the most complex is hammer2_cluster_resolve, at 79/90. In the hammer utility, the main program is 93/144, and in hammer2 68/78. Whoever said McCabe complexity should be at most 10 never worked on such complex data structures as a winged edge or a filesystem. Pierre -- ve ka'a ro klaji la .romas. se jmaji
Re: DragonFly on DigitalOcean
On Monday, December 14, 2015 13:31:05 Nuno Antunes wrote: > Hi Pierre, > > Here’s what you have to do to install dfly 4.4 on DigitalOcean: I think you misunderstood. I wasn't asking how I could install DFly on DO, but how DO could install DFly so that they can support it. Also, what would be a good argument for them to support it? On Friday, December 11, 2015 12:42:10 Zachary Crownover wrote: > Vultr works quite well, but you have to submit a support ticket to unblock > SMTP traffic from your system, otherwise mail will just pool up on the > system and you'll see weird things like 50+ concurrent instances of dma > running. I may consider Vultr for my next server. Oddly, I have a similar problem: I can send mail from my droplet on IPv4, but not IPv6. I got around this by opening port 2525 on my primary server. Pierre -- The Black Garden on the Mountain is not on the Black Mountain.
DragonFly on DigitalOcean
Several days ago I set up a server on DigitalOcean. They offer several Linux distros and FreeBSD, but no other BSD. How hard would it be for them to offer DragonFly? They use Bochs for virtualization. Pierre -- sei do'anai mi'a djuno puze'e noroi nalselganse srera
trying to upgrade packages, I get a segfault
I ran "pkg upgrade", and it has to upgrade everything because the ABI changed from 4.2 to 4.6. I got this: glproto-1.4.17.txz : 100% 15 KiB 15.7kB/s00:01 ghostscript9-x11-9.06_11.txz : 100% 36 KiB 37.2kB/s00:01 Checking integrity... done (11 conflicting) pkg: Cannot solve problem using SAT solver: upgrade rule: upgrade local xscreensaver-5.29_4 to remote xscreensaver-5.34 cannot install package xscreensaver, remove it from request? [Y/n]: y conflict rule: The following packages conflict with each other: jpeg- turbo-1.4.2(r), jpeg-8_6(r) conflict rule: The following packages conflict with each other: jpeg-8_6(r), jpeg-turbo-1.4.2(r) upgrade rule: upgrade local jpeg-8_6 to remote jpeg-8_6 cannot install package jpeg, remove it from request? [Y/n]: n pkg: Cannot solve problem using SAT solver: cannot remove package (null), remove it from request? [Y/n]: y upgrade rule: upgrade local xplanet-1.3.0_7 to remote xplanet-1.3.0_8 cannot install package xplanet, remove it from request? [Y/n]: upgrade rule: upgrade local kdeartwork-4.14.3 to remote kdeartwork-4.14.3 cannot install package kdeartwork, remove it from request? [Y/n]: pkg: Cannot solve problem using SAT solver: cannot remove package (null), remove it from request? [Y/n]: upgrade rule: upgrade local webp-0.4.3 to remote webp-0.4.4 cannot install package webp, remove it from request? [Y/n]: upgrade rule: upgrade local kstars-4.14.3_1 to remote kstars-4.14.3_2 cannot install package kstars, remove it from request? [Y/n]: pkg: Cannot solve problem using SAT solver: cannot remove package (null), remove it from request? [Y/n]: ^C At this point I figured it was a mistake to keep jpeg, so I tried again: # pkg upgrade Updating Avalon repository catalogue... Avalon repository is up-to-date. All repositories are up-to-date. Checking for upgrades (900 candidates): 100% Processing candidates (900 candidates): 100% Checking integrity... done (11 conflicting) pkg: Cannot solve problem using SAT solver: upgrade rule: upgrade local xscreensaver-5.29_4 to remote xscreensaver-5.34 cannot install package xscreensaver, remove it from request? [Y/n]: conflict rule: The following packages conflict with each other: jpeg- turbo-1.4.2(r), jpeg-8_6(r) conflict rule: The following packages conflict with each other: jpeg-8_6(r), jpeg-turbo-1.4.2(r) upgrade rule: upgrade local jpeg-8_6 to remote jpeg-8_6 cannot install package jpeg, remove it from request? [Y/n]: pkg: Cannot solve problem using SAT solver: cannot remove package �x, remove it from request? [Y/n]: upgrade rule: upgrade local xplanet-1.3.0_7 to remote xplanet-1.3.0_8 cannot install package xplanet, remove it from request? [Y/n]: upgrade rule: upgrade local kdeartwork-4.14.3 to remote kdeartwork-4.14.3 cannot install package kdeartwork, remove it from request? [Y/n]: pkg: Cannot solve problem using SAT solver: cannot install package , remove it from request? [Y/n]: upgrade rule: upgrade local webp-0.4.3 to remote webp-0.4.4 cannot install package webp, remove it from request? [Y/n]: upgrade rule: upgrade local kstars-4.14.3_1 to remote kstars-4.14.3_2 cannot install package kstars, remove it from request? [Y/n]: pkg: Cannot solve problem using SAT solver: Child process pid=22665 terminated abnormally: Segmentation fault How do I make it work? Pierre -- sei do'anai mi'a djuno puze'e noroi nalselganse srera
Re: "wlanmode hostap" doesn't work
On Saturday, November 21, 2015 11:04:57 Vasily Postnicov wrote: > It seems, that iwn driver does not support this mode. You can get supported > modes by following commands: > > ifconfig wlan0 create wlandev > ifconfig wlan0 list caps > > If you can see HOSTAP in the output, then your device can be used as access > point. > By myself, I recommend Atheros adapters. I think they are the most > supported and feature-rich. The computer came with the IWN adapter and an antenna and was marketed as an access point. If the deficiency is in the driver, I'd like to help fix it. # ifconfig wlan0 list caps drivercaps=2581e201htcaps=5006e Pierre -- ve ka'a ro klaji la .romas. se jmaji
undefined reference to `crc32_tab' building hammer
I try to rebuild world and get the following. How can I fix it? phma --- --- hammer --- cc -O -pipe -I/usr/src/sbin/hammer/../../sys -DALIST_NO_DEBUG -std=gnu99 - Wmissing-include-dirs -Wsystem-headers -Werror -Wall -Wformat-security -Winit- self -Wno-pointer-sign -Wextra -Wno-unused-parameter -Wstrict-prototypes - Wmissing-prototypes -Wpointer-arith -Wold-style-definition -Wreturn-type - Wcast-qual -Wwrite-strings -Wswitch -Wshadow -Wcast-align -Wunused-parameter - Wchar-subscripts -Winline -Wnested-externs -Wredundant-decls -o hammer hammer.o ondisk.o blockmap.o cache.o misc.o cycle.o cmd_show.o cmd_softprune.o cmd_history.o cmd_blockmap.o cmd_reblock.o cmd_rebalance.o cmd_synctid.o cmd_stats.o cmd_remote.o cmd_pseudofs.o cmd_snapshot.o cmd_mirror.o cmd_cleanup.o cmd_info.o cmd_version.o cmd_volume.o cmd_config.o cmd_recover.o cmd_dedup.o cmd_abort.o crc32.o -lm -lutil -lmd -lhammer --- _SUBDIR_all_jscan --- --- jfile.o --- cc -O -pipe-std=gnu99 -Wmissing-include-dirs -Wsystem-headers -Werror - Wall -Wformat-security -Winit-self -Wno-pointer-sign -Wextra -Wno-unused- parameter -Wstrict-prototypes -Wmissing-prototypes -Wpointer-arith -Wold- style-definition -Wreturn-type -Wcast-qual -Wwrite-strings -Wswitch -Wshadow - Wcast-align -Wunused-parameter -Wchar-subscripts -Winline -Wnested-externs - Wredundant-decls -c /usr/src/sbin/jscan/jfile.c --- _SUBDIR_all_hammer --- crc32.o: In function `crc32': crc32.c:(.text+0x1f): undefined reference to `crc32_tab' crc32.o: In function `crc32_ext': crc32.c:(.text+0x49): undefined reference to `crc32_tab' --- _SUBDIR_all_iscontrol --- --- fsm.o --- --- _SUBDIR_all_hammer --- *** [hammer] Error code 1 make[4]: stopped in /usr/src/sbin/hammer 1 error make[4]: stopped in /usr/src/sbin/hammer --- _SUBDIR_all_iscontrol --- cc -O -pipe -I/usr/src/sbin/iscontrol/../../sys/dev/disk/iscsi/initiator - std=gnu99 -Wmissing-include-dirs -Wsystem-headers -Werror -Wall -Wformat- security -Winit-self -Wno-pointer-sign -Wextra -Wno-unused-parameter -Wstrict- prototypes -Wmissing-prototypes -Wpointer-arith -Wold-style-definition -Wno- maybe-uninitialized -Wno-unused-but-set-variable -Wno-unused-but-set-parameter -Wno-unused-local-typedefs -Wno-unused-value -c /usr/src/sbin/iscontrol/fsm.c --- _SUBDIR_all_hammer --- *** [_SUBDIR_all_hammer] Error code 2 make[3]: stopped in /usr/src/sbin --- _SUBDIR_all_iscontrol --- A failure has been detected in another branch of the parallel make make[4]: stopped in /usr/src/sbin/iscontrol *** [_SUBDIR_all_iscontrol] Error code 2 make[3]: stopped in /usr/src/sbin --- _SUBDIR_all_jscan --- A failure has been detected in another branch of the parallel make make[4]: stopped in /usr/src/sbin/jscan *** [_SUBDIR_all_jscan] Error code 2 make[3]: stopped in /usr/src/sbin 3 errors make[3]: stopped in /usr/src/sbin *** [_SUBDIR_all_sbin] Error code 2 make[2]: stopped in /usr/src .CURDIR='/usr/src' .OBJDIR='/usr/obj/usr/src/world_x86_64/usr/src' LD_LIBRARY_PATH='' MACHINE_ARCH='x86_64' MACHINE='x86_64' MAKEFILE='' .TARGETS='all' .ERROR_TARGET='_SUBDIR_all_sbin' .MAKE.LEVEL='2' 1 error make[2]: stopped in /usr/src .CURDIR='/usr/src' .OBJDIR='/usr/obj/usr/src/world_x86_64/usr/src' LD_LIBRARY_PATH='' MACHINE_ARCH='x86_64' MACHINE='x86_64' MAKEFILE='' .TARGETS='all' .ERROR_TARGET='_SUBDIR_all_sbin' .MAKE.LEVEL='2' *** [everything] Error code 2 make[1]: stopped in /usr/src .CURDIR='/usr/src' .OBJDIR='/usr/obj/usr/src' LD_LIBRARY_PATH='' MACHINE_ARCH='x86_64' MACHINE='x86_64' MAKEFILE='' .TARGETS='buildworld' .ERROR_TARGET='everything' .MAKE.LEVEL='1' 1 error make[1]: stopped in /usr/src .CURDIR='/usr/src' .OBJDIR='/usr/obj/usr/src' LD_LIBRARY_PATH='' MACHINE_ARCH='x86_64' MACHINE='x86_64' MAKEFILE='' .TARGETS='buildworld' .ERROR_TARGET='everything' .MAKE.LEVEL='1' *** [buildworld] Error code 2 make: stopped in /usr/src 1 error make: stopped in /usr/src -- Don't buy a French car in Holland. It may be a citroen.
Re: undefined reference to `crc32_tab' building hammer
On Friday, May 29, 2015 19:56:08 John Marino wrote: It should go without asking, but are you sure your tree is the latest? Specifically, does it have this commit? http://gitweb.dragonflybsd.org/dragonfly.git/commitdiff/992ed92cba35afe93e4b 02bd9f1cf4178f8ab21f That would explain it. The latest change in the git log is dated May 28 08:54:09 2015 -0700. I pulled the repo yesterday, tried to build it, got the error, and didn't try it again because of a thunderstorm, then tried again today. Pierre -- lo ponse be lo mruli ku po'o cu ga'ezga roda lo ka dinko
What is killlwps in top?
I ran gdb on a program, which segfaulted. I tried to exit gdb, but it didn't exit. I tried to kill the program being debugged, then tried to kill gdb, but neither died. The program being debugged shows as status killlwps in top. gdb is in wait state. What is killlwps, and how do I kill the program? phma -- The Black Garden on the Mountain is not on the Black Mountain.
Re: The DragonFly Mail Agent is now the default MTA
On Saturday, March 07, 2015 22:10:18 John Marino wrote: New installations of DragonFly will have the DragonFly Mail Agent (dma) configured by default. This is the first step to removing Sendmail from base. This particular change will only have a small impact on existing installations. The switch is controlled by /etc/mailer.conf, and this file is never overwritten. However, there was a small change in the default configuration of periodic.conf: I just checked where this file is; on my system it's /etc/mail/mailer.conf , not /etc/mailer.conf . Pierre -- I believe in Yellow when I'm in Sweden and in Black when I'm in Wales.
Re: HEADS UP - pkg may delete itself (solution)
On Tuesday, February 17, 2015 11:34:06 Robin Hahling wrote: Recent versions of pkg(8) (older than 1.4.12) contain a bug that could cause a suppression of pkg itself during an upgrade or a package installation. If this happened, you end up without pkg(8) on your system. Here is a simple solution to fix that: # cd /usr make pkg-bootstrap I rebuilt kernel and world, then did pkg upgrade. I did not have to reinstall pkg after pkg upgraded itself, though I did before. After upgrading, I was able to compile bezitopo and run bezitest; before, cmake reported that the compiler did not work. Pierre -- sei do'anai mi'a djuno puze'e noroi nalselganse srera
Re: OT: can someone help me with an open-source CAD program?
On Wednesday, February 18, 2015 11:16:55 Zachary Crownover wrote: With regard to C++ GUI libraries, I'm personally partial to QT. Not a problem for me, as it's available as a package on both OSes. Can you write a GUI (at least a rudimentary one) for Bezitopo? I'll have to write volume routines, cross-section routines, and other surveying stuff. Will there be a problem making it run on both Qt4 and Qt5? Pierre -- Jews use a lunisolar calendar; Muslims use a solely lunar calendar.
Re: OT: can someone help me with an open-source CAD program?
On Wednesday, February 18, 2015 16:16:49 Matthias Play wrote: Hi Pierre, for the GUI part I would recommend Ultimate++ (www.ultimatepp.org) that is a C++ framework which allows to build GUI apps quite fast. It is also a nice library if you intend to use a SQL db. The documentation is not that comprehensive, but one gets responses to questions very quickly in the forums of the project. It's available for several Linux distros, though Launchpad is down at the moment. Are packages available for BSD? I will make a representation of each curve as a 3D Bézier spline approximation. Do you know how to draw these in perspective or parallel projection? Pierre -- li ze te'a ci vu'u ci bi'e te'a mu du li ci su'i ze te'a mu bi'e vu'u ci
OT: can someone help me with an open-source CAD program?
I've been working on a surveying CAD program for several years and it's nearing the point where I'd like others to hack on it. I need help with the file format, import and export, and GUI. It runs on DragonFly (my build system is hosed, but I just ran it on leaf) and Linux. If you can help, or know someone who can, let me know. Pierre -- li fi'u vu'u fi'u fi'u du li pa
pkg 1.4.6 deleted itself - I thought this might happen!
pkg 1.4.6 had a bug in which, if told to upgrade a package, it would leave the database thinking the package is installed, but its files are actually missing. I just tried to install a package, and pkg found it had to upgrade itself. Here's what happened: # pkg ins minicom Updating Avalon repository catalogue... Avalon repository is up-to-date. All repositories are up-to-date. New version of pkg detected; it needs to be installed first. The following 1 packages will be affected (of 0 checked): Installed packages to be UPGRADED: pkg: 1.4.6 - 1.4.11 The process will require 51 KB more space. 2 MB to be downloaded. Proceed with this action? [y/N]: y Fetching pkg-1.4.11.txz: 100%2 MB 99.0k/s00:22 Checking integrity... done (0 conflicting) [1/1] Upgrading pkg from 1.4.6 to 1.4.11... [1/1] Extracting pkg-1.4.11: 20% pkg: archive_read_extract(): truncated input [1/1] Extracting pkg-1.4.11: 100% [1/1] Deleting files for pkg-1.4.11: 100% # pkg ins minicom pkg: not found So I have a downloaded pkg package in /var/cache/pkg/. What do I do with it? Pierre -- ve ka'a ro klaji la .romas. se jmaji
pkg thinks a package is installed, but its files aren't
I tried to build one of my projects and it told me that cmake requires a version of libexpat that isn't installed. [phma@zyxomma ~/crypt/build/bezitopo/dbg]$ make -j 3 ./bezitest Shared object libexpat.so.6 not found, required by cmake # pkg ins cmake Updating Avalon repository catalogue... Avalon repository is up-to-date. All repositories are up-to-date. The following 1 packages will be affected (of 0 checked): Installed packages to be UPGRADED: cmake: 3.0.2 - 3.1.0 The operation will free 1 MB. 3 MB to be downloaded. Proceed with this action? [y/N]: y Fetching cmake-3.1.0.txz: 100%3 MB 784.2k/s00:04 Checking integrity... done (0 conflicting) [1/1] Upgrading cmake from 3.0.2 to 3.1.0... [1/1] Extracting cmake-3.1.0: 0% pkg: archive_read_extract(): truncated input [1/1] Extracting cmake-3.1.0: 100% [1/1] Deleting files for cmake-3.1.0: 100% [phma@zyxomma ~/crypt/build/bezitopo/dbg]$ make -j 3 ./bezitest /usr/local/bin/cmake: not found # pkg ins cmake Updating Avalon repository catalogue... Avalon repository is up-to-date. All repositories are up-to-date. Checking integrity... done (0 conflicting) The following 1 packages will be affected (of 0 checked): Installed packages to be UPGRADED: cmake: 3.0.2 - 3.1.0 The operation will free 1 MB. Proceed with this action? [y/N]: y [1/1] Upgrading cmake from 3.0.2 to 3.1.0... pkg: unlinkat(usr/local/share/licenses/cmake-3.0.2/BSD3CLAUSE): No such file or directory pkg: unlinkat(usr/local/share/licenses/cmake-3.0.2/LICENSE): No such file or directory pkg: unlinkat(usr/local/share/licenses/cmake-3.0.2/catalog.mk): No such file or directory [1/1] Extracting cmake-3.1.0: 100% [phma@zyxomma ~/crypt/build/bezitopo/dbg]$ make -j 3 ./bezitest Shared object libexpat.so.1 not found, required by cmake phma@zyxomma ~/crypt/build/bezitopo/dbg]$ ls /usr/lib/*expat* ls: /usr/lib/*expat*: No such file or directory [phma@zyxomma ~/crypt/build/bezitopo/dbg]$ ls /usr/local/lib/*expat* ls: /usr/local/lib/*expat*: No such file or directory [phma@zyxomma ~/crypt/build/bezitopo/dbg]$ ls /lib/*expat* ls: /lib/*expat*: No such file or directory # pkg upg expat Updating Avalon repository catalogue... Avalon repository is up-to-date. All repositories are up-to-date. Checking integrity... done (0 conflicting) Your packages are up to date. # ls /usr/local/lib/*expat* /usr/local/lib/libexpat.a /usr/local/lib/libexpat.so.1.6.0 /usr/local/lib/libexpat.so /usr/local/lib/libexpat.so.6 /usr/local/lib/libexpat.so.1 Should I file a bug in pkg? Its version is 1.4.6. Pierre -- When a barnacle settles down, its brain disintegrates. Já não percebe nada, já não percebe nada.
Re: polkit does not start
On Tuesday, January 20, 2015 20:21:32 karu.pruun wrote: I am using 64 bit machines in both cases. It's a puzzling situation since the relevant ports all seem to be identical on both machines. Tested with your com.mycompany.Test1 and got the same result. Thanks for mentioning the dbus webpage, I'll try and see if I can get more info how the dbus EXTERNAL authentication mechanism is supposed to work: what the strings 30 (FreeBSD) vs 34323934393637323935 (DragonFly) stand for. 34323934393637323935 is obviously 4294967295 (2^32-1) in ASCII hex. What it means in the protocol, I don't know. Pierre -- Jews use a lunisolar calendar; Muslims use a solely lunar calendar.
trying to get xdmcp to work
Zyxomma runs DragonFly. Caracal runs Linux. I'm trying to get a Xephyr or Xnest window on caracal to run a session on zyxomma. Both have kdm. kdm.log on zyxomma ends with: Invalid MIT-MAGIC-COOKIE-1 keyInvalid MIT-MAGIC-COOKIE-1 keyInvalid MIT-MAGIC- COOKIE-1 keyInvalid MIT-MAGIC-COOKIE-1 keyInvalid MIT-MAGIC-COOKIE-1 key # tcpdump -i em0 port 177 tcpdump: verbose output suppressed, use -v or -vv for full protocol decode listening on em0, link-type EN10MB (Ethernet), capture size 65535 bytes 05:20:20.168710 IP6 caracal.53230 zyxomma.xdmcp: UDP, length 7 05:20:22.168724 IP6 caracal.53230 zyxomma.xdmcp: UDP, length 7 05:20:22.691578 IP6 zyxomma.xdmcp caracal.53230: UDP, length 61 05:20:22.692564 IP6 caracal.53230 zyxomma.xdmcp: UDP, length 457 05:20:24.692965 IP6 caracal.53230 zyxomma.xdmcp: UDP, length 457 05:20:25.084306 IP6 zyxomma.xdmcp caracal.53230: UDP, length 61 05:20:25.084573 IP6 zyxomma.xdmcp caracal.53230: UDP, length 52 05:20:25.084638 IP6 zyxomma.xdmcp caracal.53230: UDP, length 52 05:20:25.085192 IP6 caracal.53230 zyxomma.xdmcp: UDP, length 29 05:20:27.085849 IP6 caracal.53230 zyxomma.xdmcp: UDP, length 29 05:20:31.086083 IP6 caracal.53230 zyxomma.xdmcp: UDP, length 29 05:20:39.086018 IP6 caracal.53230 zyxomma.xdmcp: UDP, length 29 05:20:55.085212 IP6 caracal.53230 zyxomma.xdmcp: UDP, length 29 05:21:27.085312 IP6 caracal.53230 zyxomma.xdmcp: UDP, length 29 05:21:40.127998 IP6 zyxomma.xdmcp caracal.53230: UDP, length 107 at which point Xephyr quits. /usr/local/share/config/kdm/kdmrc: [Xdmcp] # Whether KDM should listen to incoming XDMCP requests. # Default is true Enable=true Zyxomma is accessible from the outside only by IPv6; to find it one would have to sniff packets or break into leaf (or have root on leaf). I'm not immediately concerned about leaving the port open, but once I get it working, restricting access to my boxes would be nice. Pierre -- La sal en el mar es más que en la sangre. Le sel dans la mer est plus que dans le sang.