| --- Log opened Fri Dec 27 00:00:30 2019 |
| 00:11 | | Pinkhair [user1@Nightstar-g7hdo5.dyn.optonline.net] has joined #code |
| 00:14 | | gnolam [quassel@Nightstar-csto10.cust.bahnhof.se] has quit [[NS] Quit: Z?] |
| 00:14 | | Pink [user1@Nightstar-g7hdo5.dyn.optonline.net] has quit [Ping timeout: 121 seconds] |
| 00:16 | | Pink [user1@Nightstar-g7hdo5.dyn.optonline.net] has joined #code |
| 00:18 | | Pinkhair [user1@Nightstar-g7hdo5.dyn.optonline.net] has quit [Ping timeout: 121 seconds] |
| 01:07 | | Degi [Degi@Nightstar-ua966j.dyn.telefonica.de] has quit [Operation timed out] |
| 01:09 | | Degi [Degi@Nightstar-5ofvqn.dyn.telefonica.de] has joined #code |
| 02:07 | | Kindamoody is now known as Kindamoody[zZz] |
| 02:16 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 02:16 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 05:05 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 05:11 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has joined #code |
| 05:11 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
| 05:55 | <&jeroud> | sshine, macdjord: Why not connect to whois.nic.<tld> and whois.iana.org in parallel? |
| 05:56 | <&jeroud> | That way you can return the result from the former if it arrives, or an appropriate error (bad tld or network error or whatever) if it doesn't. |
| 05:58 | <&jeroud> | And you can probably keep a list of known tlds (ideally in a resource file rather than hardcoded) that gets updated periodically if necessary. |
| 06:02 | <@Alek> | didn't they take out some TLDs recently though? |
| 06:05 | <&jeroud> | Sure, but only for things that are very seldom used. |
| 06:06 | <&jeroud> | So getting an awkward error for that is forgivable. |
| 06:07 | <&jeroud> | But you might want to check iana before returning any kind of "failure to reach tld whois" error anyway. |
| 06:49 | <@macdjord|slep> | Alek: It would be better to store each known TLD along with its last-confirmed-to-exist time, and expire any that get too old. |
| 06:53 | | celticminstrel [celticminst@Nightstar-0p84vo.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 08:53 | <&McMartin> | Part 2 of, apparently, 8. https://bumbershootsoft.wordpress.com/2019/12/27/raw-win32-an-application-shell/ |
| 09:02 | <&McMartin> | (After looking through my commit history, the "obvious" breakdown is: syscalls, ABI, toolchain basics; application shell; custom widget painting; player input; menus and keyboard shortcuts therefor; reacting properly to resizing the window; icons and declarative resources; custom dialog boxes) |
| 09:54 | <&McMartin> | The S is for Security https://www.youtube.com/watch?v=ozIKwGt38LQ |
| 09:55 | <&McMartin> | Also MEMS devices are super neat |
| 10:43 | | Kindamoody[zZz] is now known as Kindamoody |
| 11:08 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has joined #code |
| 11:19 | | gnolam [quassel@Nightstar-0oc.n80.224.185.IP] has joined #code |
| 11:19 | | mode/#code [+o gnolam] by ChanServ |
| 13:20 | | gnolam [quassel@Nightstar-0oc.n80.224.185.IP] has quit [[NS] Quit: Gone] |
| 13:24 | <@sshine> | woot, I got a 3% raise this year! |
| 13:27 | | * Emmy still wishes for more... ^.^; |
| 13:28 | <&ToxicFrog> | Meanwhile, I continue to do alarming things with shell escape sequences: http://toxicfrog.github.io/osc-wrapper/ |
| 13:30 | <&ToxicFrog> | (this is the setup that allows me to run weechat on my server, press a hotkey, and have it open a URL in the browser on the laptop) |
| 13:56 | < Emmy> | seems a bit roundabout 0.o |
| 14:14 | <&ToxicFrog> | Emmy: if you have an idea for a less roundabout approach, I'm all ears :) |
| 14:14 | | gnolam [quassel@Nightstar-22puij.mobileonline.telia.com] has joined #code |
| 14:14 | | mode/#code [+o gnolam] by ChanServ |
| 14:24 | < Emmy> | Well, not running the chat on a remote server would be my option, but then again, i'm a lazy ass :P |
| 14:32 | | gnolam [quassel@Nightstar-22puij.mobileonline.telia.com] has quit [[NS] Quit: Fuck SJ with an entire train] |
| 14:35 | <&ToxicFrog> | Not running the chat remotely defeats the entire purpose, so |
| 14:47 | | celticminstrel [celticminst@Nightstar-0p84vo.dsl.bell.ca] has joined #code |
| 14:47 | | mode/#code [+o celticminstrel] by ChanServ |
| 14:48 | < Emmy> | yea, figured. |
| 14:48 | < Emmy> | could try a bouncer, though? |
| 15:09 | <@sshine> | jeroud, yes, I could resolve/connect whois.nic.<tld> and fail somewhat gracefully not knowing whether it was a network error or a not-yet-existing TLD. |
| 15:10 | <@sshine> | jeroud, and yes, I could connect to whois.iana.org once and cache the result. but since this is a Haskell library, I don't want to rely on file caches, because this increases the complexity of the library. it's not very common, I think, for libraries to depend on implicit file caches. I could use a memory cache, which would be kind of useless if the library e.g. were used in a command-line program that exits |
| 15:11 | <@sshine> | between each query. |
| 15:12 | <@sshine> | jeroud, combining cached IANA lookup + connect to whois.nic.<tld> in parallel might be an option, but without the cache it'd be very inefficient. |
| 15:12 | <@sshine> | I'm building a web-service around this library where caching is going to be much more convenient. |
| 15:13 | <@sshine> | so my choice for the library would depend on some kind of optimal mixture between convenience, annoyance, and what people expect. |
| 15:14 | <@sshine> | okay, so here's what I'm thinking about going with: |
| 15:15 | <@sshine> | 1) if it's a recognized TLD, just look up the WHOIS server in the database. 2) if it's supposedly a new TLD, cache https://data.iana.org/TLD/tlds-alpha-by-domain.txt in memory, validate if the TLD is real, and do a 'whois.nic.<tld>' lookup. |
| 15:16 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 15:16 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 15:17 | <@sshine> | that means I'm doing at most two lookups. an average WHOIS lookup is ~1.5KB and that TLD list is ~12K, so hum. maybe make that extra safety lookup optional. |
| 15:18 | <@sshine> | that tlds-alpha-by-domain.txt is what's potentially updated daily, so I can't cache it in the executable. |
| 15:18 | <@sshine> | annoyingly they update the file daily even if there were no additions to the list. |
| 15:23 | <&ToxicFrog> | Emmy: I used to do that, in fact. The advantage of running the whole client remotely is that there's only place to install and configure. |
| 15:23 | <&ToxicFrog> | Anywhere I have an ssh client, I have my entire setup. |
| 15:24 | <&ToxicFrog> | *only one place |
| 15:24 | <&ToxicFrog> | Overall, it's a much nicer UX than using a bouncer. |
| 15:24 | < Emmy> | hmmmh. |
| 15:24 | | * Emmy ponders |
| 15:25 | | * Emmy sees opportunities for web-based bouncer front-ends there |
| 15:25 | <&ToxicFrog> | (and there is a fair amount of setup involved, since weechat is my client not just for IRC, but discord, slack, twitter, facebook chat, and my monitoring systems) |
| 15:25 | <&ToxicFrog> | Those already exist. |
| 15:25 | <&ToxicFrog> | IRCCloud is a proprietary one, and GlowingBear is a frontend for weechat. |
| 16:26 | | gnolam [lenin@Nightstar-e3tf4i.priv.bahnhof.se] has joined #code |
| 16:27 | | mode/#code [+o gnolam] by ChanServ |
| 19:41 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 20:12 | <&[R]> | `cat /proc/$client_pid/environ | tr '\0' '\n' ` D: |
| 20:12 | <&[R]> | You can just `tr '\0' \n' $file` |
| 20:14 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has quit [Connection closed] |
| 20:14 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 20:14 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has joined #code |
| 20:14 | | mode/#code [+ao VirusJTG VirusJTG] by ChanServ |
| 20:18 | <&[R]> | ToxicFrog: grep -azEo '^SSH_CLIENT.*' /proc/$$/environ | tr '\0 ' == | cut -d = -f 2 |
| 20:19 | <&[R]> | Some small further refinement: grep -azE '^SSH_CLIENT' /proc/$$/environ | tr '\0 ' == | cut -d = -f 2 |
| 20:26 | | Derakon[AFK] is now known as Derakon |
| 20:29 | <&ToxicFrog> | [R]: TIL about `grep -z`, thanks |
| 20:30 | <&ToxicFrog> | I am well aware that, for most foo, `cat bar | foo` can be replaced with `foo bar`. `cat bar | ...` is easier to experiment with, though, and the cost of forking an additional cat is negligible. |
| 20:31 | <&[R]> | You can do: `< $file foo` instead |
| 20:31 | <&[R]> | Gives you the same benefit |
| 20:37 | <&[R]> | All in one command now: `sed -Ez '/^SSH_CLIENT/ s|.*=([^ ]*) .*|\1|p; d' /proc/$$/environ` |
| 20:39 | <&ToxicFrog> | I am aware of that, too. |
| 20:40 | <&[R]> | Oh, TIL: sed -Ez 's|SSH_CLIENT=([^ ]*) .*|\1|p; d' /proc/$$/environ |
| 22:05 | <&McMartin> | MULTI-REGISTER DRIFTING |
| 22:05 | <&McMartin> | The 65816 is made entirely of rusty nails and broken glass |
| 22:05 | <&McMartin> | I'd say "I'm amazed anyone ever actually used this for anything" but then again there are like three actually significant uses of it and quite plausibly only one counts as "major" |
| 22:06 | <&McMartin> | I have never seen a programming manual, much less an official one, that has so many instances of "the programmer must take extra care not to..." "... or disaster will surely result" |
| 22:06 | <&McMartin> | Even instruction decoding depends on run-time processor flags that are expected to change multiple times within a function |
| 22:07 | <&McMartin> | And yeah |
| 22:07 | <&McMartin> | They Know What They Did |
| 22:09 | <&McMartin> | In x86, whether you're loading a 16-bit or 8-bit value depends on the operand, more or less: MOV AL, 3 vs MOV AX, 3 vs MOV EAX, 3 for byte/word/dword |
| 22:10 | <&McMartin> | On 65816 whether lda #3 loads a byte or a word depends on the run-time state of a processor status flag regarding whether the accumulator is in byte mode or double-byte mode |
| 22:10 | <&McMartin> | And this means that whether that instruction is two or three bytes long also depends on that flag |
| 22:11 | <&McMartin> | Which means if you lose control of your flag bits somehow you can end up losing instruction sync within the code stream |
| 22:18 | <&[R]> | That's amazing |
| 22:19 | <&[R]> | Did they use that trash fire to heat their offices? |
| 22:22 | <&McMartin> | Pretty sure there's a reason this was only used for the Apple IIgs, the SNES, and accelerator boards for C64 and Apple II lines |
| 22:22 | <&McMartin> | *As an accelerator chip* it's actually completely fine, and they nailed important parts of backcompat that previous 8-bit revisions of the 6502 did not nail |
| 22:22 | <&McMartin> | It's also got some features when in "be an accelerator for an 8-bit system" mode that remind me uncomfortably of Zachtronics games, but for real |
| 22:23 | <&McMartin> | In particular, the high byte of the 16-bit accumulator becomes a secret second 8-bit accumulator that you can swap in and out |
| 22:23 | <&McMartin> | The Z80 has one of those two but you're only really supposed to use it for, like, simulations of supervisor mode |
| 22:24 | <&McMartin> | Also it's got 24-bit pointers and the high byte is used like a segment register on x86 |
| 22:25 | <&McMartin> | And while I've mocked the x86 for spending 32 bits to form a 20-bit address, I can at least acknowledge that this buys something that strictly disjoint banks like the 65816 does not have |
| 22:25 | <&McMartin> | Overall still fine except for a few headaches for someone who would want to, well, design the Apple IIgs. I should look into exactly how they made that work |
| 22:26 | <&McMartin> | There's a stripped down only-16-address-pins version of the 65816 called the 65802 that was pin-compatible with the 8-bit chips |
| 22:26 | <&McMartin> | Despite being crippled I actually have way fewer objections to that crippled version! |
| 22:35 | | macdjord|slep is now known as macdjord |
| 22:45 | <&McMartin> | Poor Life Choices Instruction Set Computer |
| 23:12 | <@Alek> | I swear I've read that the 65802 was more popular, more widely used... |
| 23:13 | <@Alek> | than the 65816, I mean |
| 23:13 | <@Alek> | but eh, it's been so long... I'm probably remembering something wrong. |
| 23:51 | <&McMartin> | I can see the 65802 being less insane tbh |
| 23:51 | <&McMartin> | In any other application I feel like the Motorola 68000 would be a better choice |
| 23:58 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has quit [Ping timeout: 121 seconds] |
| --- Log closed Sat Dec 28 00:00:32 2019 |