| --- Log opened Thu Aug 22 00:00:56 2019 |
| 00:28 | | McMartin [mcmartin@Nightstar-ipm463.ca.comcast.net] has quit [[NS] Quit: Kernel upgrade] |
| 00:40 | | McMartin [mcmartin@Nightstar-ipm463.ca.comcast.net] has joined #code |
| 00:40 | | mode/#code [+ao McMartin McMartin] by ChanServ |
| 00:49 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 01:45 | | himi [sjjf@Nightstar-1drtbs.anu.edu.au] has joined #code |
| 01:45 | | mode/#code [+o himi] by ChanServ |
| 02:17 | | celmin|away is now known as celticminstrel |
| 02:29 | <&ToxicFrog> | I think this may be a new record for number of closing parens: (withBackgroundColor (Colour "#000000")))))))) |
| 02:37 | <@Reiv> | Impressive! |
| 03:15 | <&McMartin> | Clojure? |
| 04:46 | | Derakon is now known as Derakon[AFK] |
| 06:00 | | Vorntastic [uid293981@Nightstar-2dc.p8m.184.192.IP] has joined #code |
| 06:00 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
| 06:18 | | celticminstrel [celticminst@Nightstar-6an2qt.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 08:21 | | himi [sjjf@Nightstar-1drtbs.anu.edu.au] has quit [Ping timeout: 121 seconds] |
| 08:31 | < Yossarian> | Vorntastic: https://hackaday.com/2019/08/20/dirty-tricks-for-6502-programming/#more-372673 |
| 08:39 | <&McMartin> | My absolute favorite version of the "what the instruction is depends on where you jump" trick is this one. https://bumbershootsoft.wordpress.com/2014/05/04/cycle-exact-delays-on-the-6502/ |
| 08:40 | <&McMartin> | https://bumbershootsoft.wordpress.com/2015/12/13/raster-grip/ is a technique I managed to devise on my one, though there is a broader and very well established C64 technique it is part of. |
| 08:40 | <&McMartin> | *on my own |
| 08:45 | <&McMartin> | https://en.wikipedia.org/wiki/MOS_Technology_6502#Bugs_and_quirks is also a valuable list. |
| 08:45 | <&McMartin> | And once you have your feet under you CodeBase64 is an excellent cite for C64 stuff specifically. |
| 08:47 | <~Vorntastic> | At this point I can read most of the assembler in here at speed, the only question is semantics |
| 08:48 | | * McMartin nods |
| 08:48 | <&McMartin> | 6502 is really clean that way |
| 08:49 | <~Vorntastic> | Like I haven't opened the opcode reference in several days |
| 08:49 | <&McMartin> | It's also got the unusual property that you can meaningfully assemble it with, like sed. |
| 08:49 | <&McMartin> | The first byte of every instruction is the opcode and addressing more as an arbitrary lump, and then the argument follows |
| 08:49 | <&McMartin> | Everything that matters is at byte boundaries |
| 08:49 | <&McMartin> | This is not as true for Z80, even less true for x86, and the opposite of true for 68000 and ARM |
| 08:50 | <&McMartin> | It's not technically true for MIPS but it's close; all instructions are one word long and looking at the most significant few bits will tell you which bitfield layout (of like four total) will correctly decode the instruction |
| 08:53 | <&McMartin> | ARM is similar until you get to THUMB but even then the encodings are kind of bonkers |
| 08:55 | <&McMartin> | arguments in noncontiguous bitfields, or chunks of arguments are not in significance order, etc |
| 08:56 | <~Vorntastic> | That's wild |
| 09:04 | <&McMartin> | Remember, bitfield decoding is free in hardware, because you just hook up the wires differently from output to next input~ |
| 09:04 | <&McMartin> | (Today in statements that are maybe not completely false) |
| 09:05 | | * McMartin finishes auditing the month's UQM hacking to note which bits he needs to tear back out now that he realizes what he should have been doing all along, heads to bed |
| 09:06 | <&McMartin> | Though re: Z80 it being almost true: all numeric arguments are split at byte boundaries like in the 6502 |
| 09:06 | <&McMartin> | But all register arguments are glommed into the opcodes in vaguely but not entirely consistent ways |
| 09:06 | <&McMartin> | But some of *that* is that you write LD A, X instead of TXA |
| 09:06 | <~Vorntastic> | I'm just reminded of reading about MMIX and going "he describes features in ways that suggest they're not common" |
| 09:07 | <&McMartin> | And LD A, 47 as LDA #47 |
| 09:07 | <&McMartin> | But then there are also three opcodes that mean "start over but use this other table of instructions" because they have Too Many Opcode+Hardcoded Operand Combinations |
| 09:08 | <&McMartin> | I actually never properly learned MIX nor MMIX, just kind of seat-of-the-pantsed it when I encountered code in it |
| 09:08 | <~Vorntastic> | And I thought this couldn't be true because these things are far too sensible |
| 09:09 | <~Vorntastic> | And learning about how assembler is in general was a very ... |
| 09:09 | <&McMartin> | There's "sensible" and then there's "we can build it with 4,000 transistors"~ |
| 09:10 | <~Vorntastic> | It was not pleasant |
| 09:11 | <&McMartin> | The joy of assembler is largely "this is what you can do with stone tools" |
| 09:11 | <&McMartin> | And MMIX ends up looking rather a lot like, well, a low level intermediate language |
| 09:11 | <&McMartin> | Hell, MIPS does, even with its bonkers jump delay slots. |
| 09:11 | <&McMartin> | (specifically, MIPS looks like three-address code~) |
| 09:11 | <&McMartin> | Though once you get a good feel for it, 6502 ends up looking like two-address code. |
| 09:13 | <&McMartin> | viz. https://bumbershootsoft.wordpress.com/2017/03/11/getting-a-decent-and-fast-prng-out-of-an-8-bit-chip/ and how that insight kicked me out of my tailspin |
| 09:14 | <&McMartin> | Heh |
| 09:14 | <&McMartin> | Also re: inconsistent or non-orthogonal instructions |
| 09:14 | <&McMartin> | "Then on x86, anything goes, except for 16-bit indexed memory accesses, where of the 63 possible ways you could write an instruction of the form mov [bp+si+3], exactly 17 are legal." |
| 09:24 | <~Vorntastic> | What. |
| 09:39 | <&McMartin> | The rule for an address in 16-bit x86 is: at most one of [BX, BP] plus at most one of [SI, DI] plus at most one constant, and you can't omit everything |
| 09:39 | <&McMartin> | 3*3*2-1 = 17 |
| 10:10 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has joined #code |
| 10:50 | <@abudhabi> | That feeling when you have one of those camouflage moths sitting on the edge of your laptop screen, but you just let it sit there. |
| 10:50 | <&Reiver> | "Hey little dude, I have no beef with you. Chill." ? |
| 11:12 | < Emmy> | as long as they don't randomly flap through your cone of vision |
| 11:34 | | gnolam [quassel@Nightstar-0oc.n80.224.185.IP] has joined #code |
| 11:34 | | mode/#code [+o gnolam] by ChanServ |
| 12:13 | <@abudhabi> | Reiver: Yes. |
| 12:14 | <@abudhabi> | https://i.imgur.com/5cpd7CW.jpg |
| 13:17 | | gnolam [quassel@Nightstar-0oc.n80.224.185.IP] has quit [[NS] Quit: Gone] |
| 13:33 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has quit [Connection closed] |
| 13:33 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has joined #code |
| 13:33 | | mode/#code [+ao VirusJTG VirusJTG] by ChanServ |
| 14:13 | | celticminstrel [celticminst@Nightstar-6an2qt.dsl.bell.ca] has joined #code |
| 14:14 | | mode/#code [+o celticminstrel] by ChanServ |
| 14:52 | | Kindamoody[zZz] is now known as Kindamoody |
| 15:05 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has quit [Operation timed out] |
| 15:20 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 15:20 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 15:52 | | gnolam [lenin@Nightstar-e3tf4i.priv.bahnhof.se] has joined #code |
| 15:52 | | mode/#code [+o gnolam] by ChanServ |
| 17:03 | | Kindamoody is now known as Kindamoody|afk |
| 18:18 | | Vorntastic [uid293981@Nightstar-2dc.p8m.184.192.IP] has quit [[NS] Quit: Connection closed for inactivity] |
| 19:12 | | Derakon[AFK] is now known as Derakon |
| 19:30 | <&McMartin> | Meanwhile, in another channel entirely, out of left field |
| 19:30 | <&McMartin> | https://nurpax.github.io/posts/2019-08-18-dirty-tricks-6502-programmers-use.html |
| 19:30 | <&McMartin> | These are almost all C64-specific except for some things that have names as standard compiler optimizations |
| 19:30 | <&McMartin> | But it was last week and it seems relevant! |
| 19:36 | <~Vornicus> | Oh *that* is how they do it |
| 19:37 | <~Vornicus> | I always remembered games that would load in and run themselves and that was 1. rude and 2. black fucking magic |
| 19:46 | <~Vornicus> | (looking at one of the winners, which made it so loading his prg would land a jmp right inside the basic idle loop) |
| 20:26 | <~Vornicus> | 10: sizeof(critter_stats) |
| 21:26 | <&McMartin> | The basic idle loop technique is the one that's actually reliable |
| 21:28 | <&McMartin> | Also you can be less rude by loading into the cassette read buffer and then replicating the (consistent) data from BASIC's runtime up to the warm restart vector |
| 21:28 | <&McMartin> | and there is enough room in the cassette read buffer to write a small program to properly load and run some other file, so you can get a bootstrap loader out |
| 21:54 | | Kindamoody|afk is now known as Kindamoody |
| 22:14 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 22:14 | | macdjord|slep [macdjord@Nightstar-rslo4b.mc.videotron.ca] has quit [Connection closed] |
| 22:14 | | Kindamoody [Kindamoody@Nightstar-eubaqc.tbcn.telia.com] has quit [Connection closed] |
| 22:14 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 22:14 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 22:14 | | KiMo|autorejoin [Kindamoody@Nightstar-eubaqc.tbcn.telia.com] has joined #code |
| 22:15 | | macdjord|slep [macdjord@Nightstar-rslo4b.mc.videotron.ca] has joined #code |
| 22:15 | | mode/#code [+o macdjord|slep] by ChanServ |
| 22:15 | | KiMo|autorejoin is now known as Kindamoody |
| 22:53 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 23:06 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has quit [Ping timeout: 121 seconds] |
| 23:28 | | Kindamoody is now known as Kindamoody[zZz] |
| 23:36 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has joined #code |
| 23:36 | | mode/#code [+ao VirusJTG VirusJTG] by ChanServ |
| 23:48 | <&Derakon> | Mental note: in C# at least, float.NaN != float.NaN. |
| 23:48 | <&Derakon> | This may be common. |
| 23:48 | <&Derakon> | But it means I can't use NaN for "this variable has not yet been initialized". |
| 23:50 | <@Reiv> | That is correct and expected behavior |
| 23:50 | <@Reiv> | NaN is NULL and the whole point of NULL is 'you don't know' |
| 23:50 | <@Reiv> | 'we have no data' |
| 23:50 | <@Reiv> | "Is this no data the same as that no data?" "How the hell am I supposed to answer that question" |
| 23:50 | <&Derakon> | Yeah, I'm used to thinking in terms of Python's None which is its own special type. |
| 23:51 | <@Reiv> | Yeah, that's /explicitly/ no data |
| 23:51 | <&Derakon> | But that's more like a null pointer, which you couldn't assign to a float. |
| 23:51 | <@Reiv> | Right |
| 23:51 | <@Reiv> | There is theoretically something equivalent to isnull() in there I am certain |
| 23:51 | <@Reiv> | Guess I'm gonna be learning the precise syntax along with you, of course~ |
| --- Log closed Fri Aug 23 00:00:57 2019 |