mason | lamermaster: There are bears about, and they've attacked our chickens, so they've been demoted from free range chickens to free range while we're outside with them. Not as good, but safer. | 02:39 |
---|---|---|
golinux | Your presence would stop a bear? (Was going to move to OT but you're not there . . . ) | 02:41 |
mason | Yep. I chased a big one off. He probably didn't realize I was probably more scared than he was, but there's willpower for you. :P | 02:41 |
mason | (They're black bears around here, not the grizzlies other parts of the country have. A grizzly wouldn't be impressed.) | 02:43 |
mason | I'll go back into -offtopic, but I don't much want to linger there if it's going be the same right-wing stuff that made #debianfork unpalatable. | 02:44 |
GNUmoon2 | Hi all, I hope you can help me. I'm trying out devuan (using net install iso) on a system76 laptop that is connected to the wired lan. I've confirmed it has internet access by going to the shell and wgetting a file. | 04:28 |
GNUmoon2 | However, when I try to detect disks, the installer complains that it needs non-free firmware files to operate. | 04:28 |
gnarface | GNUmoon2: does it actually complain that you need them or is it just asking if you want to include any? | 04:29 |
GNUmoon2 | namely iwlwifi-9000*.ucode files. I select No to loading the missing firmware, but I cannot proceed anyfurther. | 04:29 |
gnarface | well, what's interesting about that is it is wifi firmware | 04:30 |
gnarface | so you shouldn't need it to install if you have a wired lan connection already | 04:31 |
gnarface | also, i thought the netinstall already included that firmware anyway... weird it would be complaining | 04:31 |
GNUmoon2 | gnarface: Hi, it is asking to load them, I select No, no other complaints and it goes back to the install menu. | 04:31 |
gnarface | oh, well maybe that's fine then. select the next item in the menu and pretend like nothing happened. see if it works. | 04:32 |
gnarface | you just won't have wifi | 04:32 |
GNUmoon2 | Oh the strange thing is that it asks twice. I select No to loading the firmware, then the same menu appears again. | 04:32 |
gnarface | you can always add it again later if you change your mind | 04:32 |
gnarface | it might ask again for bluetooth | 04:33 |
GNUmoon2 | That's the weird thing, I'm selecting "Partition disks" and it asks for the wifi firmware... | 04:33 |
gnarface | theres's some combo wifi/bluetooth devices that show up as two devices, might explain it | 04:33 |
gnarface | yea, that is weird i agree | 04:33 |
GNUmoon2 | oh, i see. | 04:33 |
gnarface | yea there's several possible explanations for all this behavior so far that you've described, and the most likely one is just that it is normal | 04:34 |
gnarface | maybe it detects the wifi device, then the harddrives, then the bluetooth | 04:34 |
gnarface | or something like that | 04:34 |
gnarface | but it shouldn't matter | 04:35 |
GNUmoon2 | oh oka. | 04:35 |
GNUmoon2 | ha, now I can see the partitions! I did select no to the wifi firmware twice, but I got kicked back to the main menu without the partition info previously. | 04:36 |
GNUmoon2 | Thanks gnarface! | 04:37 |
gnarface | no problem GNUmoon2. that is expected behavior. | 04:37 |
gnarface | whenever you get kicked back out to the main menu again, often that's all there is to do with that menu selection and you can just go to the next item manually | 04:38 |
GNUmoon2 | This is my first time with Davuan/Debian so not sure what to expect :) | 04:38 |
gnarface | you can also safely re-do many of the steps, or even do some of them out of order, or skip others. if you know what you're doing, the dumbness of the tool can grant you creative flexibility | 04:38 |
gnarface | for the most part it behaves identically to the debian installer with minor changes | 04:39 |
GNUmoon2 | nice, thanks | 04:39 |
GNUmoon2 | Another question. I'm going to do full disk encryption, (https://devuan.org/os/documentation/install-guides/beowulf/full-disk-encryption). Is there any Devuan guide on using a USB with the key-file for decryption instead of the passphase? | 04:42 |
yanmaani | GNUmoon2: what some people do is they have the header on the USB stick | 04:46 |
yanmaani | although keep in mind flash drives are usually not great in durability | 04:46 |
yanmaani | YMMV | 04:46 |
GNUmoon2 | I see. | 04:47 |
GNUmoon2 | So using the command "cryptsetup luksHeaderBackup /dev/sda --header-backup-file /mnt/myusb"? | 04:49 |
GNUmoon2 | or I'm on the wrong track? | 04:49 |
* gnarface doesn't know | 04:50 | |
yanmaani | GNUmoon2: To put it more bluntly, I would not do that. It can break and then your entire device is fucked | 04:51 |
GNUmoon2 | yanmaani: yeah, that's the risk. security vs convenience/reliability | 04:53 |
GNUmoon2 | I think I'll have a go at the standard disk encryption approach and see how the whole Devuan experience goes :) | 04:54 |
mason | Ah, he left. | 05:09 |
mason | Something to recommend to people is to set a key they can type in, and other for the USB key to automatically decrypt. | 05:10 |
mason | LUKS1, you get four key slots. LUKS2 is... 20 or something? I don't know. | 05:10 |
mason | But then if your USB stick with key data dies, you're not hosed and you can just make another one. All you lose is automatic unlocking. | 05:11 |
mason | Same if you ever simply lose it. | 05:11 |
lamermaster | ~~~~~~~~~~~~~;5~;5~;5~;5~;5~;5~;5~;5~;5~;5~;5~;5~;5~;5~;5~;5~;5~;5~;5~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | 09:17 |
lamermaster | # | 09:17 |
Hum | I try to configure to have many resources on my local computer like kiwix and dwww. See <https://friendsofdevuan.org/doku.php/user:hum/self-reliance> for details. Do you know a search-engine that needs few resources and is easy to configure? yacy is too slow when it get to 1 mio Links, at least on my hdd. I have no SSD. I don't think of a desktop search, I like to have access using http(s). Any hints? | 12:20 |
luser978 | htdig? | 12:30 |
luser978 | Hum | 12:32 |
Hum | luser978: Thx I'll have look. | 12:37 |
Danny | hi all | 13:22 |
Danny | i'm having some trouble writing the iso image to a usb | 13:22 |
Danny | can anyone help? | 13:22 |
gnarface | Danny: using dd? | 13:23 |
Danny | i tried dd and mintstick | 13:24 |
Danny | it completes the task and all but as i boot it just doesn't detect the stick anymore | 13:24 |
gnarface | don't specify a block size with dd, and run "sync" afterwards | 13:24 |
gnarface | and make sure you unzipped the image first | 13:26 |
gnarface | those are the 3 leading causes of failure | 13:26 |
Hurgotron | gnarface: interesting. I never had issues by specifying blocksize. And ihere are any, they should be at the end of the image, while the boot block is at the start. | 13:27 |
Hurgotron | *if there | 13:27 |
Danny | this is the *desktop.iso | 13:27 |
Danny | i tried the *desktop-live.iso too | 13:28 |
gnarface | Hurgotron: seems to be a hardware specific failure for extra large block sizes (> 1MB) | 13:28 |
Danny | that one went further | 13:28 |
Hurgotron | gnarface: OK, I'll keep that in mind. 1 MB is what I usually specify to speed up things and avoid wear on flash | 13:29 |
gnarface | Hurgotron: it will actually in most cases do neither of those things | 13:29 |
gnarface | Hurgotron: it will in most cases do nothing at all | 13:29 |
Danny | i managed to select i think the first option while booting that and the desktop never appeared | 13:29 |
Hurgotron | gnarface: the speed difference is very noticeable every time I forget it :) | 13:30 |
gnarface | Hurgotron: if you're reading *from* usb or sd flash, you might benefit from a increased blocksize, maybe | 13:30 |
Hurgotron | ah well, benchmark time | 13:31 |
gnarface | Hurgotron: if you're writing *to* flash though, you won't be able to sustain the types of speeds that would benefit from it | 13:31 |
gnarface | Hurgotron: you might even slow it down | 13:31 |
Hurgotron | gnarface: https://bpa.st/FI6A | 13:45 |
Hurgotron | admittedly, the difference is less big than I thought, but that definetely will be hardware-dependent. | 13:45 |
Hum | xz -tv devuan_ascii_2.0.0_armel_raspi1.img.xz # Get the size of the image | 13:52 |
Hum | xzcat devuan_ascii_2.0.0_armel_raspi1.img.xz | pv -s 1891M > /dev/sdX # replace X with the sdcard | 13:52 |
Hum | Just another way to write an image to flash | 13:52 |
Danny | so i try to start devuan-live from the bootable usb and it just stops after an error i think | 13:55 |
Danny | it says IPv6: ADDRCONF (NETDEV_UP): eth0: link is not ready | 13:55 |
Danny | then it goes to a console | 13:56 |
Danny | what can i do about this? | 13:56 |
Danny | cool | 14:03 |
lamermaster | Ok, I found a possible workaround to have no xset, here a little fork (xsetfork.c): https://termbin.com/pv0o | 14:04 |
lamermaster | cool, glad to see that you use devuan on raspberry. you will have to modify the size of partition, because devuan image is too small. | 14:06 |
lamermaster | I use sda1 sda2 with the kernel on sda1 as fat, and sda2 is a new partition with about 30gb to hold devuan for raspberry (on sanedisk usb). | 14:07 |
lamermaster | not working dpms | 14:26 |
buZz | Hum: i like dd status=progress , instead of pv | 14:29 |
lamermaster | on old unix, progress is not there. | 14:31 |
buZz | lamermaster: i dno how old cause debian stable has it | 14:32 |
lamermaster | hi... does linux-image-i386 exist in the packages?` | 15:00 |
fsmithred | lamermaster, there's linux-image-686 and linux-image-686-pae | 15:06 |
lamermaster | I install by hand | 15:19 |
lamermaster | is there a vmlinuz 686 from an http? | 15:19 |
lamermaster | I just need the file and the modules | 15:19 |
lamermaster | ok, I took vmlinuz and linux modules from the following deb: wget -c http://http.us.debian.org/debian/pool/main/l/linux/linux-image-4.9.0-13-686-pae_4.9.228-1_i386.deb | 15:34 |
lamermaster | and I installed the i386 on /target (destination) with : debootstrap --no-check-gpg --foreign --arch i386 --include=wpasupplicant ascii . http://pkgmaster.devuan.org/merged | 15:35 |
Hum | buZz: Yes, as you write it, I use status=progress more often then pv. In past I also used ddrescue until I found status=progress | 15:42 |
fsmithred | lamermaster, sorry, I was afk. Maybe you want hd-media files? https://pkgmaster.devuan.org/devuan/dists/beowulf/main/installer-i386/current/images/ | 16:08 |
systemdlete2 | Scary experience just now. My beowulf VM froze up (somewhat); Was able to get a console (which was also slow). Logged in and killed firefox-esr. When I returned to the desktop, not only was firefox gone, but so were 2 shell windows I had had open. | 18:34 |
systemdlete2 | I've been experiencing freeze-ups with beowulf frequently, at least in a VM. My testbox running beowulf (as host, not VM) doesn't seem to have these issues, but I am not entirely sure if I am comparing them fairly. | 18:35 |
systemdlete2 | Both are updated and upgraded. | 18:36 |
systemdlete2 | Now, I had been trying to close firefox, thinking that was the only app running on the desktop (besides the two shells, which use far less resources). It would not close, and by then, the clock (mine shows seconds) was frozen, which was when I resorted to the console. | 18:37 |
ErRandir | Firefox probably ate all your memory and you ended in swapping hell. To confirm checkout the load with uptime. Next time run it with a ulimit -v limit or turn off swap. | 18:46 |
systemdlete2 | That's what I'm thinking also. I was trying to compare speed test sites, and a couple of them, despite claiming they don't use js, are full of... | 18:50 |
systemdlete2 | (that brown stuff) | 18:51 |
systemdlete2 | Swapping hell might help to explain why the shell windows disappeared I guess. | 18:51 |
systemdlete2 | They probably got starved for memory also. | 18:52 |
systemdlete2 | This firefox heck is repeatable. Very repeatable. | 18:52 |
systemdlete2 | Thanks ErRandir. I'll look at it. Though I'm not sure what to do about it other than confirm, once again, that ff is a pig | 18:52 |
systemdlete2 | (some websites more than others also) | 18:53 |
HumanG33k | hi from a rescue boot and chroot there is a wa yto get last try to startup logs ? | 20:12 |
lamermaster | interesting, this kernel of arm runs faster than devuan. https://www.brobwind.com/wp-content/uploads/2019/01/2019_01_25_rpi3_a586f2b.bin.gz perfs are quite good, devuan is not faster. | 20:25 |
rwp | systemdlete2, Also check if the OOM Killer (out of memory killer) triggered and killed anything important. Check the system logs for kernel space logs of oom killer actions. | 20:33 |
rwp | systemdlete2, I would also use 'vmstat' to look at the si/so fields and see if they are higher than 2 and if so then swap thrash is happening. | 20:35 |
rwp | Remember than when the kernel OOM Killer strikes the application getting killed has no ability to react. It's just a "kill -9 $pid" equivalent action. The process is removed. | 20:37 |
rwp | I like to watch the 'htop' live bar graph of memory usage when testing memory intensive applications like web browsers. Useful information. | 20:37 |
rwp | HumanG33k, Yes. Logs will be in /var/log/ directory. If you have 'bootlogd' installed then it will save the boot logs to /var/log/boot file that you can browse. Otherwise just the standard logs. | 20:38 |
HumanG33k | thx rwp i just install bootlogd i hope i can have a more verbose output | 20:39 |
lamermaster | this is p�s.c to do it without hop and even ps ;) https://termbin.com/gwkb (ps.c) | 20:42 |
rwp | Good luck HumanG33k! Since you are rescue booting I assume you are fighting with a problem there. | 20:42 |
HumanG33k | yes i m ^^ | 20:43 |
HumanG33k | humm i have WARNING: Device /dev/ram8 not initialized in udev database even after waiting 10000000 microseconds. | 20:48 |
HumanG33k | when i try to update-grub2 any clue ? | 20:49 |
rwp | Sorry. Not a clue about /dev/ram8 not initialized. | 20:49 |
HumanG33k | it s not only 8 but look like it s a loop | 20:50 |
HumanG33k | ^^ | 20:50 |
rwp | I miss the previous simplicity of grub v1 where I don't remember having any of the problems with the current grub v2. | 20:51 |
rwp | grub v2 does more dynamic automatic probing, and often gets things wrong. :-( | 20:51 |
rwp | You might try saving the output to a pastebin and then posting a link and seeing if a second eye from anyone here sees something? | 20:52 |
HumanG33k | for now i just try to debug a little more | 20:53 |
HumanG33k | maybe i miss a service startup on the chroot | 20:53 |
rwp | Hmm... Initializing /dev/ramX in a chroot? That seems wrong. | 20:55 |
rwp | In the chroot are you mounting the live /dev so that the devices from the hosting system are available? I generally "mount -o bind /dev /srv/chroot/ceres/dev" or similar. | 20:56 |
rwp | That's just an example for a concrete discussion and of course needs to be different in other environments. | 20:56 |
rwp | The /dev is almost always required in a chroot. /proc sometimes depending upon the applications being run. Other directories such as /sys maybe depending upon how invasive the program is. | 20:57 |
HumanG33k | mount --rbind /dev $TARGETDIR/dev | 20:57 |
HumanG33k | mount --make-rslave $TARGETDIR/dev | 20:57 |
rwp | I don't know anything about those options. Probably appeared after I already learned older ways of doing things. Sorry. :-( | 20:58 |
HumanG33k | no trouble | 20:58 |
HumanG33k | i will retry pretty sure the good old option work better as always this days | 20:59 |
rwp | Reading the docs... --rbind will not be needed for /dev since there are no recursive bind mounds below it. | 20:59 |
rwp | There is /dev/shm which is a tmpfs but that has never been needed by me. | 21:00 |
rwp | There is /dev/pts which is used by dpkg but I just ignore dpkg whining about it. It complains but does not really need it. | 21:00 |
lamermaster | Is XFS better than EXT4? | 21:01 |
rwp | Better? That depends upon the workload. There are no absolute rules. Including this one. :-) But I am using xfs most places. | 21:01 |
HumanG33k | me to, and agree for rules. | 21:02 |
rwp | xfs is *different* from ext4. Most of those differences I prefer. Some of those differences are negatives but I avoid those and enjoy the positives. | 21:02 |
lamermaster | what is negative with XFS, for instance? I am new to xfs. | 21:02 |
rwp | Most things about xfs are positive good things. It effectively has a built in fsck so that one never needs to do an external fsck. Mostly. Therefore crash recovery is always faster than with ext4. | 21:03 |
rwp | I had to say something good before the bad. | 21:03 |
rwp | ext4 handles incrementally increasing the file system size better. Such as with lvm and using lvextend to incrementally increase the size of the file system. | 21:04 |
rwp | With xfs the allocation strategy defeats adding small incremental amounts. In xfs the allocation plan is to divide the entire space allocated into quads. Then allocate among the quads. | 21:05 |
rwp | If one then adds a very small incremental chunk, that is also divided into quads. And then allocation from there when needed. Which is not so efficient. | 21:05 |
lamermaster | what is more stable for a fileserver running NFS (server) ? xfs or ext4? | 21:05 |
rwp | Therefore adding small incremental amounts to xfs is not good for file system block allocation efficiency. I avoid it. | 21:05 |
rwp | Perfectly good to double the size of an xfs file system however. Doubling is almost always almost an excellent performance plan. | 21:06 |
rwp | I have no problems over NFS for either ext4 or xfs. So have nothing from there to pick one or the other. | 21:06 |
rwp | Do you use file ACLs? For whatever reason in some Linux kernels I have turned up some file ACL issues with both xfs and ext4 that I haven't been able to reduce to root cause yet. | 21:07 |
rwp | Therefore I suggest that if you are using those over NFS that you specifically test that functionality to ensure that it is not a problem for you. | 21:08 |
rwp | As I said it has always worked okay but then I tripped over some specific version problems where it works fine on one kernel but not in another. And I haven't been able to conclude on the actual problem yet. | 21:08 |
lamermaster | I have a lot of small files, but sometimes about 5-6 GB files. | 21:09 |
HumanG33k | if you have ressources make capacity planning/test and run for finding the best for you | 21:09 |
rwp | NFS has many advantages. Such as being about the most well used network file system. But network file systems are higher latency. | 21:09 |
lamermaster | With sync, it is slow, but more reliable. async too risky. | 21:10 |
rwp | Latency over NFS has been something we have been suffering through forever. And often modifying our work flows to avoid it. Working in /tmp and so forth. | 21:10 |
rwp | We have always used async. | 21:10 |
rwp | Most of my work experience is in technical environments. VLSI chip design labs. Places like that. | 21:11 |
lamermaster | I lot data with async,... large files mostly. | 21:11 |
rwp | Not working in a financial institution doing money transfers or anything. | 21:11 |
rwp | When people worry about reliability of sync versus async one must also be aware of the environment in which it is being used. | 21:12 |
rwp | When we are doing things like circuit simulations we do hundreds of thousands of them. | 21:13 |
rwp | If on a particular day the power to the building fails, the backup power UPS units drain, and the systems are forced to shutdown, we will lose those running jobs regardless of the setting of sync. | 21:13 |
rwp | So we would rather have the much higher performance of async to get more jobs done while the power is on. | 21:14 |
lamermaster | On bsd, I see also : ffs, mfs, udf. there are plenty of choices. | 21:18 |
rennj | openzfs be done with fsck | 21:19 |
rennj | and perhaps in the future work across winblows,crapple,linux,bsd's,opensolaris even..1 fs for all | 21:20 |
lamermaster | zfs is too heavy for ARMs boards. | 21:20 |
rennj | send/recv import | 21:20 |
rennj | couple TB in 1GB of ram so ive read.. | 21:20 |
rennj | arc zil and such yeah hybrid storage | 21:21 |
mason | rennj: Hey, what are you doing in here? o/ | 21:37 |
mason | lamermaster: zfs is okay on ARM. If you've got a gig of RAM you can make it work. | 21:38 |
lamermaster | thank you | 21:44 |
lamermaster | lot of choices | 21:44 |
rennj | embedded nand and stuff jffs2 or it that nor...i forget | 21:51 |
rennj | nand vs nor..fs choices | 21:51 |
brocashelm | what's a good minimalist graphical browser on a devuan desktop running a pentium-era cpu? i checked out netsurf and qutebrowser | 21:51 |
rennj | wear leveling and all | 21:51 |
rennj | JFFS2 replacements: LogFS, UBIFS, and YAFFS | 21:53 |
Hum | brocashelm: depends on the website, imho. Some sites browse well with netsurf, other need a firefox like engine. Some sites have a nice text/lynx mode, if it fits to you, e.g. https://reiseauskunft.bahn.de/bin/query.exe/dl | 21:53 |
lamermaster | I used JFS but once it crashes, good luck, it takes ages to recover and to boot the machine. | 21:53 |
Hum | lamermaster: btrfs! ;) If you have database-like workloads, the COW is hell and you should disable it and have no advantages about ext4 ;) | 21:55 |
Hum | but otherwise I like it :) | 21:56 |
lamermaster | never heard of btrfs :( | 21:57 |
Hum | similar to zfs but in the kernel and the fs of the future for ten years or something like that. Now we are getting slowly into future, since btrfs is used in some distros by default. See wikipedia: https://en.wikipedia.org/wiki/Btrfs or for details https://btrfs.wiki.kernel.org/index.php/Main_Page | 22:00 |
rennj | redhat is certainly pushing it | 22:02 |
rennj | ubuntu has openzfs fu.. | 22:02 |
rennj | suse what are they pushing | 22:02 |
Hum | rennj: redhat is pushing btrfs back, imho, to sell their own solution | 22:03 |
Hum | did redhat change their way? | 22:03 |
Hum | https://en.wikipedia.org/wiki/Btrfs#Commercial_support | 22:05 |
Wonka | weeeell, I'm hearing about fucked up btrfs instances often enough | 22:05 |
Wonka | I wouldn't want to use it currently | 22:05 |
Hum | Hi Wonka: Yeah, me too, when I read my screen messages loudly. But I haven't problems for one or two years, even if I filled my partitions totally | 22:06 |
rennj | wonky world..haha | 22:06 |
rennj | so redhat is not pushing btrfs but oracle is | 22:06 |
rennj | meh..data loss perhaps on customers | 22:07 |
Hum | "In 2020, Btrfs was selected as the default file system for Fedora 33." | 22:07 |
rennj | https://access.redhat.com/discussions/3138231 | 22:07 |
mason | Jamie's commentary is the take-away of that discussion. | 22:11 |
rennj | well fedora is suppose to be upstream right | 22:14 |
rennj | hehe | 22:14 |
rennj | yeah | 22:15 |
rennj | the beta alpha bits of fufture redhat | 22:15 |
Hum | [1] (sorry german article) from 2017 states that Red Hat will buy permabit and will get access to cloud storage solution(s) with data deduplication and compression. [1] https://www.golem.de/news/rhel-7-4-red-hat-beendet-unterstuetzung-fuer-btrfs-1708-129266.html | 22:24 |
HumanG33k | one step at time everything becoming "open source" | 22:24 |
Hum | HumanG33k: even your genome? | 22:25 |
HumanG33k | (not so usefull except if you have $$) | 22:25 |
HumanG33k | hum not pretty sure dna base storage are patent | 22:26 |
Hum | open source doesn't need patents at first, only copyright. idk how the legal situation about that is | 22:27 |
HumanG33k | open source vs free(dom) software | 22:28 |
phogg | it's literally the same thing | 22:29 |
Hum | HumanG33k: I like the story about OpenStreetMap: A guy respects the copyright of google maps and creates a libre solution. Google needs to lower the price of Google Maps :D | 22:29 |
Hum | "free(dom) software" is a nice solution to emphasize the core meaning of libre | 22:30 |
HumanG33k | btw it s a complex question in a complex world. | 22:30 |
rennj | "Give 'em the razor; sell 'em the blades" | 22:35 |
rennj | the os is free, just buy this hardware | 22:35 |
buZz | Hum: did you see that google was busted a bunch of times, making malicious edits on OSM | 22:35 |
buZz | Hum: they switched directions on one-way streets in london, completely breaking navigation for a while | 22:36 |
Hum | buZz: No, I didn't read about it. | 22:38 |
buZz | i follow #osm for a while now, since i worked in GIS for a while | 22:38 |
buZz | https://www.wired.com/2012/01/osm-google-accusation/ | 22:39 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!