Time Nick Message 00:01 Emerald2 In this particular IRC channel? People come here to get help with stuff mostly. 00:01 rubenwardy the IRC channel is private, you need to be invited 00:02 nri oh well what are irc channels anyway? 00:02 nri is this private? 00:02 Emerald2 A channel is like a chatroom. 00:02 Emerald2 ruben, you mean Redcat's IRC channel? 00:02 rubenwardy yes 00:02 GreenDimond By "invited" he means told what the channel name is 00:03 nri oh ok ruben 00:04 nri well got to go 00:04 GreenDimond Hey rubenwardy, that awards PR is still collecting dust :3 00:05 rubenwardy lol 01:53 Xiong Docs say run up to four quarries, one on each edge of the square pit. This is unnecessary. 01:55 Xiong I have stacked two quarries, one smack atop the other; this works. Much less cabling of course and no additional output piping at all. Conveniently enough, the lower quarry ejects just fine. 01:56 Xiong Unclear if lower Q ejects into upper Q's buffer or both eject into pipe... and does it matter? 02:00 Xiong Obviously the working volumes differ by one floor. But the upper Q does its top and immediately finds the next floor bare air, so advances to work on the next floor down together with lower Q. I don't see why this can't continue for a stack of 4 or more. It might be more reasonable to run banks side by side... 9 looking very manageable. Of course the same slight non-overlap. Not sure that's a real issue. 02:05 Xiong One snag is that my two Q setup ran into is sync. Both started together, both soon wind up working on the same node... so no increase in speed. I threw a block at the ceiling and restarted the upper Q. Soon both were cutting the same floor from opposite directions for a total rate of 2 cubic meters mined per second. Don't know if the random block helped or if reset alone does enough. 02:06 * Xiong pizza 05:38 Sveta https://forum.minetest.net/viewtopic.php?id=7354 the file is missing? 06:27 Xiong zacts, the story goes that Erdos gave a lecture, starting by chalking up an equation. He smiled at the students and said, "Well that's obvious. So..." and ground to a halt. Turned around, looked at the board, looked some more. Grabbed his notes and ran from the hall. An hour later he walked in, tie loose, covered in sweat, beaming from ear to ear. "I was right! It *is* obvious." 06:27 Xiong Oops. Funky client. 10:03 powerolive ... 10:03 powerolive ... 10:06 powerolive ... 10:18 Sveta powerolive: help you with anything? 13:26 Xiong I need to amend my previous report on stacked quarries. Apparently the *upper* Q has a little difficulty clearing its buffer. Oddly enough, the lower Q "purges cache" every so often; but all upper Q slots filled for me (once so far). Hitting the button to "disable" triggered the cycle as usual and the upper Q did eject properly; and resumed when reenabled, still correctly out of sync. I haven't actually allowed the upper buffer to 13:35 Xiong Okay now I'm stumped. Just now the lower Q autopurged; I stood there and watched the entities float in the tube. (They didn't get stuck in upper Q.) Then upper autopurged and again, it all went out. 13:38 Xiong It's unclear to me what governs autopurge. It's plain that disabling a Q always triggers a purge; and a unpowered Q just sits there with its pants full. I've noted many autopurges with a buffer much less than full. 13:43 Xiong Sorry I'm annoyed at the misuse of language. The quarry contains a 12-slot buffer which clears on some conditions. The word "cache" implies some random request that is fetched. Buffers are cleared, not purged. Since this jargon spans both hardware and software worlds, I don't see how it was got wrong. 13:43 Xiong I'm going to go root around in the quarry code. 14:30 Xiong https://gist.github.com/Xiong/e87ba38c9eb625d53c703645e52051b7 14:33 Xiong Well that says it all. I don't like to be harsh in criticism but there's a conflict between nice and factual. I'm terribly sensitive to other people's pain and I've spent much time repackaging my opinions. I don't see how to make this sound good. 14:39 Xiong I have implemented buffers in microelectronic hardware, several programming languages, and (if the concept may be extended) in cases as prosaic as ripening bananas. I have never planned to clear a buffer randomly. It's often wise to clear before full; overruns can be bad. There's often a value in clearing periodically, perhaps without even checking if there's anything in the buffer. It's often wise to look for a signal from downst 14:41 Xiong There's enough pseudo-randomness in any sort of complex device. 14:44 Xiong Now, no matter how long I test, I will never be certain whether a quarry's buffer overrun is due to bad luck or my insanity in stacking them. 15:25 JDCodeIt Hi All, I am seeing quite a few of these messages in the log and no update on servers.minetest.net: servers.minetest.net/announce not found (Timeout was reached) (response code 0) 15:39 calcul0n JDCodeIt, it happens, i guess it's just a connexion problem 15:39 calcul0n iirc you can change the curl timeout in advanced settings if it's a problem 16:19 JDCodeIt Looks like the default curl timeout is 5 seconds, so I will try increasing it 17:27 JDCodeIt Question is curl timeout one of the parameters that can be changed while the server is running - and have any effect? 21:27 MinetestBot 02[git] 04pauloue -> 03minetest/minetest: Use true pitch/yaw/roll rotations without loss of precision by pgimen… 13d5456da https://git.io/fhHIr (152019-02-07T21:26:06Z) 23:00 powerolive hi