| --- Log opened Sat Apr 29 00:00:46 2017 |
| 00:40 | | Jessikat [Jessica@Nightstar-bt5k4h.81.in-addr.arpa] has quit [Connection reset by peer] |
| 01:00 | | Derakon[AFK] is now known as Derakon |
| 05:09 | | Derakon is now known as Derakon[AFK] |
| 05:11 | | Jessikat [Jessikat@Nightstar-oer.tcr.132.82.IP] has joined #code |
| 05:14 | | Jessikat` [Jessikat@Nightstar-ip83lr.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 05:43 | | RchrdB [RchrdB@Nightstar-qe9.aug.187.81.IP] has quit [Ping timeout: 121 seconds] |
| 07:25 | | celticminstrel is now known as celmin|sleep |
| 07:40 | | macdjord is now known as macdjord|slep |
| 08:33 | < Jessikat> | Go isn't a systems language, it doesn't quite translate enough to the bare metal underneath |
| 08:34 | < Jessikat> | It's more like a lightweight scheduling/server language, which it's write capable at without being spectacular |
| 08:34 | < Jessikat> | The one area it is far ahead of the competition is ease of concurrency, but it falls behind a lot in terms of static typing |
| 08:35 | < Jessikat> | It's a bit of an odd beast |
| 08:42 | < Jessikat> | I think my definition of a systems language is 'exposes the architecture of the machine it runs on one-to-one, in explicit commands' |
| 08:42 | < Jessikat> | So C is because it offers you the machine types and the memory model via specific addressing |
| 08:43 | < Jessikat> | Javascript is not, because it doesn't even let you express arrays as contiguous blocks |
| 08:46 | < Jessikat> | I should write down the LISP library that corresponds to C syntax |
| 08:46 | < Jessikat> | That'd be amusing |
| 09:05 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 09:05 | | mode/#code [+o macdjord] by ChanServ |
| 09:07 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 09:07 | | mode/#code [+o mac] by ChanServ |
| 09:09 | | macdjord|slep [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 09:10 | | macdjord|slep [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 09:10 | | mode/#code [+o macdjord|slep] by ChanServ |
| 09:10 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 09:11 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 09:44 | | Jessikat [Jessikat@Nightstar-oer.tcr.132.82.IP] has quit [The TLS connection was non-properly terminated.] |
| 09:44 | | Jessikat [Jessikat@Nightstar-oer.tcr.132.82.IP] has joined #code |
| 09:50 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Operation timed out] |
| 09:51 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 09:51 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 09:53 | | Jessikat` [Jessica@Nightstar-bt5k4h.81.in-addr.arpa] has joined #code |
| 10:07 | | RchrdB [RchrdB@Nightstar-qe9.aug.187.81.IP] has joined #code |
| 10:29 | | Kindamoody[zZz] is now known as Kindamoody |
| 11:39 | | Jessikat` [Jessica@Nightstar-bt5k4h.81.in-addr.arpa] has quit [[NS] Quit: Leaving] |
| 11:41 | | Jessikat` [Jessikat@Nightstar-6lff9t.dab.02.net] has joined #code |
| 11:43 | | Jessikat [Jessikat@Nightstar-oer.tcr.132.82.IP] has quit [Ping timeout: 121 seconds] |
| 11:59 | | RchrdB [RchrdB@Nightstar-qe9.aug.187.81.IP] has quit [Connection closed] |
| 12:14 | | Kindamoody is now known as Kindamoody|out |
| 12:53 | | Jessikat [Jessikat@Nightstar-6lff9t.dab.02.net] has joined #code |
| 12:55 | | Jessikat` [Jessikat@Nightstar-6lff9t.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 15:56 | | celmin|sleep is now known as celticminstrel |
| 17:03 | | Stalker [Z@Nightstar-r9lk5l.cust.comxnet.dk] has joined #code |
| 17:04 | | Stalker [Z@Nightstar-r9lk5l.cust.comxnet.dk] has quit [[NS] Quit: Into the hole again, we hurried along our way, into a once-glorious garden now seeped in dark decay.] |
| 17:08 | | Jessikat` [Jessikat@Nightstar-6lff9t.dab.02.net] has joined #code |
| 17:08 | | Jessikat [Jessikat@Nightstar-6lff9t.dab.02.net] has quit [The TLS connection was non-properly terminated.] |
| 17:13 | | Stalker [Z@Nightstar-r9lk5l.cust.comxnet.dk] has joined #code |
| 17:17 | | Stalker [Z@Nightstar-r9lk5l.cust.comxnet.dk] has left #code [] |
| 17:45 | | Jessikat [Jessica@Nightstar-bt5k4h.81.in-addr.arpa] has joined #code |
| 18:11 | | Jessikat`` [Jessikat@Nightstar-jv5j45.dab.02.net] has joined #code |
| 18:14 | | Jessikat` [Jessikat@Nightstar-6lff9t.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 18:15 | | Jessikat`` [Jessikat@Nightstar-jv5j45.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 19:00 | | Kindamoody|out is now known as Kindamoody |
| 19:50 | | Derakon[AFK] is now known as Derakon |
| 19:58 | | Turaiel[Offline] is now known as Turaiel |
| 20:35 | | RchrdB [RchrdB@Nightstar-qe9.aug.187.81.IP] has joined #code |
| 22:36 | | Turaiel is now known as Turaiel[Offline] |
| 23:04 | | * McMartin solves his network issues in RISC OS in the dumbest way possible |
| 23:04 | <&McMartin> | ssh works, but scp does not |
| 23:05 | <&McMartin> | Solution: base64-encode the files I want to transfer, copy-paste them through SSH |
| 23:05 | <&McMartin> | Sadly, RISC OS is not unix-like enough to do this the "sensible" way in SSH because that would imply that pipes work the way Unix wants, and at least on this Pi 3, they don't. |
| 23:06 | <&McMartin> | (The "sensible" way being piping the result of tar on one end through an ssh connection told to use tar xf - as its shell on the other) |
| 23:20 | <&ToxicFrog> | Does >redirection work? If so you could ssh > foo.tar and then untar it |
| 23:30 | <&McMartin> | It's the other direction that's been the problem. |
| 23:30 | <&McMartin> | I'm trying to pull stuff from the RISC OS machine into the Unix one. |
| 23:30 | <&McMartin> | What I *really* need to do is just get a build of scp that actually works on the Pi 3. |
| 23:30 | <&McMartin> | And it's extremely vexing that scp and sftp neither work when ssh and ssh-keygen both do. |
| 23:40 | < himi-cat> | It works nicely for me! |
| 23:40 | < himi-cat> | . . . using raspbian |
| 23:41 | <&McMartin> | Yes, there is zero chance that this is a hardware issue. |
| 23:41 | <&ToxicFrog> | McMartin: do they intrinsically not work, or do they just default to disabled in sshd_config? |
| 23:41 | <&McMartin> | They destroy the entire system, necessitating yanking the plug. |
| 23:41 | <&ToxicFrog> | Or does RISCOS not have the necessary binaries on either end? |
| 23:41 | <&ToxicFrog> | o.O |
| 23:42 | < himi-cat> | Wow - that's an unusual response |
| 23:42 | <&McMartin> | So, I *suspect* this is because it's a Pi 3 and I'm relying on a slightly dodgy automatic binary patcher to get around the fact that the ABI in SVC mode changed incompatibly between ARMv7 and ARMv8. |
| 23:42 | <&McMartin> | Its filesystem is even less like POSIX than DOS's is, and its notion of memory protection isn't far removed from Windows 95 |
| 23:43 | < himi-cat> | Is it an ARMv8 specific RISC OS build? |
| 23:43 | <&McMartin> | So the chip going wild can smash basically everything. |
| 23:43 | <&McMartin> | Uh, no, the whole problem is that it's an ARMv7 RISC OS build which will hang the process unpatched because it uses the no-longer-existent SWP instruction |
| 23:43 | < himi-cat> | Hm |
| 23:44 | < himi-cat> | Okay, that makes sense |
| 23:44 | <&McMartin> | However, these are only used inside libc, more or less, which is why the dodgy autopatcher more or less works for many things. |
| 23:44 | <&McMartin> | Including, as it happens, ssh and ssh-keygen. |
| 23:44 | < himi-cat> | Curious that scp ans sftp both fail if ssh doesn't |
| 23:44 | <&McMartin> | Quite so. |
| 23:44 | < himi-cat> | Have you tried a plain ftp, or possibly even netcat? |
| 23:44 | <&McMartin> | sftp fails with the message "socketpair: protocol not supported" which may be a config issue |
| 23:45 | <&McMartin> | See, you keep naming Unix tools, when this is for all practical purposes an Acorn Archimedes |
| 23:45 | < himi-cat> | You don't even have /netcat/ available? |
| 23:45 | <&McMartin> | I could probably try to build it from source! |
| 23:45 | <&McMartin> | But, I mean |
| 23:46 | <&McMartin> | I don't have *cat* cat available. |
| 23:46 | <&McMartin> | Because It Is Not A UNIX System. |
| 23:46 | <&McMartin> | "cat" in fact lists the contents of your current directory |
| 23:47 | <&McMartin> | Note also that "." is the directory separator and the ports of Unix tools are alarmingly inconsistent about autoswapping it back and forth between / |
| 23:47 | < himi-cat> | I figured in their efforts to make a system that at least /pretends/ to be useful they'd have ported some tools like that |
| 23:47 | < himi-cat> | (the people making RISC OS distributions for Raspberri Pis, that is) |
| 23:48 | <&McMartin> | So, leaving that aside, because that's an entirely different issue |
| 23:48 | <&McMartin> | (I'll get to the hilarity that is attempting to use gcc in a bit...) |
| 23:48 | <&McMartin> | ... it simply isn't the case that netcat is a "basic utility" without which you cannot operate, and if you think this is the case you've been living in a pure-UNIX land for entirely too long |
| 23:49 | <&McMartin> | There are ports of Bash and GNU Coreutils, but I'd have to manually patch out the SWPs for each file to have any hope at all they'd work, and ideally one would use the workflows that are native to that system, instead of insisting that it look like Unix and complain whenever it doesn't. |
| 23:49 | <&McMartin> | If I want Unix I have something like five unix systems in the same room as it. |
| 23:50 | < himi-cat> | So what's the RISC OS equivalent tool for getting data to and from a network connected system? |
| 23:50 | <&McMartin> | So, GCC in fact did get a port and there is enough of a POSIX compatibility layer to make things mostly work from source if you're lucky and aren't relying on things like, say, pipes or forks |
| 23:50 | < himi-cat> | FTP isn't exactly *nix only |
| 23:51 | <&McMartin> | I'm not willing to run ftp traffic over the network~ |
| 23:51 | < himi-cat> | netcat is just a network pipe - if there's something functionally similar you could probably use that at the RISC OS end |
| 23:51 | < himi-cat> | Wait, this isn't on a private network? |
| 23:51 | <&McMartin> | Ideally not entirely. |
| 23:51 | <&McMartin> | That's why I'm sticking to SSH ports - the core SSH solution is actually a terminal emulator called Nettle |
| 23:52 | <&McMartin> | Which can speak SSH2 and also handles xterm codes right (which the core task window does not) |
| 23:52 | < himi-cat> | About as Unix a thing as you could suggest! |
| 23:52 | <&McMartin> | Not really? No more than PuTTY is on Windows. |
| 23:53 | < himi-cat> | Yeah, a port of a *nix terminal and an ssh client |
| 23:53 | <&McMartin> | Right, so, the thing about netcat is that it's a minimal tool that you use to let your program system's pipe operations move stuff between files |
| 23:53 | <&McMartin> | This implies, like, a pipe system that works like Unix does. |
| 23:54 | < himi-cat> | Regardless, that's not a particularly helpful bit of joshing |
| 23:54 | <&McMartin> | RISC OS as near as I can tell has file redirection but can't actually do pipes. |
| 23:54 | <&McMartin> | If it *did* ssh alone would be enough. |
| 23:54 | < himi-cat> | File redirection for inpu? |
| 23:54 | < himi-cat> | input, rather? |
| 23:55 | < himi-cat> | Which direction are you trying to get your data |
| 23:55 | < himi-cat> | ? |
| 23:55 | <&McMartin> | I'm trying to copy from the RISC OS machine to the wider world |
| 23:55 | <&McMartin> | The other direction I can basically use a web browser. |
| 23:56 | < himi-cat> | Can you run a web /server/ on the RISC OS side? |
| 23:56 | <&McMartin> | ... maybe if I had more than three days of experience with it |
| 23:56 | <&McMartin> | So, I hadn't mentioned the whole thing about building software on it |
| 23:56 | <&McMartin> | See, as noted, . is the directory separator |
| 23:56 | <&McMartin> | normally unix ports of tools swap . and / on all file names |
| 23:57 | <&McMartin> | gcc, however, has decided that instead, it will insist on loading c.foo and h.foo when you call or #include stuff |
| 23:57 | <&McMartin> | This necessitates a directory structure that whatever you're trying to build probably doesn't have. |
| 23:57 | < himi-cat> | Go gcc devs |
| 23:58 | <&McMartin> | There are cross tools. I've gotten them installed and working in a VM (they don't build on versions of gcc > 4.x, hooray!) and they pass their tests but I haven't figured out how to actually executue builds of the packages the package maintainers use. |
| 23:58 | <&McMartin> | Oh no, that's the RISC OS community doing that on purpose because that's how their C devkits worked in 1993 |
| 23:59 | <&McMartin> | And things like that are why I'm treating the RISC OS system as roughly equivalent to the ZX81 and the Commodore 128. |
| 23:59 | < himi-cat> | Their C compilers did weird shit to work around a different directory separator? |
| 23:59 | <&McMartin> | The part where things get exciting is that unlike those two it is *supposed* to actually be up-to-speed enough to be able to speak to the modern world |
| 23:59 | <&McMartin> | No, because it wasn't weird shit |
| --- Log closed Sun Apr 30 00:00:06 2017 |