"Hello, world!"rv                 >ov                   \ov                     >ov                       \ov                         >ov                           \ov                             >ov    
> 1740497227 892087 PRIVMSG #esolangs :14[[07User:BrainFuckGirl/Sandbox14]]4 10 02https://esolangs.org/w/index.php?diff=152708&oldid=152707 5* 03MihaiEso 5* (-533) 10Blanked the page
> 1740497319 956697 PRIVMSG #esolangs :14[[07Esolang:Sandbox14]]4 10 02https://esolangs.org/w/index.php?diff=152709&oldid=150939 5* 03MihaiEso 5* (+537) 10
> 1740498295 45627 PRIVMSG #esolangs :14[[07User:Hotcrystal0/My esolang ideas14]]4 10 02https://esolangs.org/w/index.php?diff=152710&oldid=152021 5* 03Hotcrystal0 5* (+43) 10
> 1740498370 23206 PRIVMSG #esolangs :14[[07User:Hotcrystal014]]4 10 02https://esolangs.org/w/index.php?diff=152711&oldid=152699 5* 03Hotcrystal0 5* (+58) 10
> 1740498490 522528 PRIVMSG #esolangs :14[[07Esolang:Sandbox14]]4 10 02https://esolangs.org/w/index.php?diff=152712&oldid=152709 5* 03Hotcrystal0 5* (+4) 10fixing name
> 1740498699 226806 PRIVMSG #esolangs :14[[07User:Hotcrystal0/Unnamed esolang 114]]4 N10 02https://esolangs.org/w/index.php?oldid=152713 5* 03Hotcrystal0 5* (+131) 10Created page with "[name here] is an [[esoteric programming language]] made by [[User:Hotcrystal0]]. It is based around the characteristics of people."
< 1740502001 951435 :craigo!~craigo@user/craigo QUIT :Ping timeout: 248 seconds
< 1740508738 620770 :CanisCorvus!~CanisCorv@shef-17-b2-v4wan-169232-cust98.vm3.cable.virginm.net JOIN #esolangs * :[https://web.libera.chat] CanisCorvus
> 1740508749 84355 PRIVMSG #esolangs :14[[07User talk:BrainFuckGirl14]]4 10 02https://esolangs.org/w/index.php?diff=152714&oldid=152704 5* 03BrainFuckGirl 5* (+343) 10/* Question */
< 1740509599 103640 :CanisCorvus!~CanisCorv@shef-17-b2-v4wan-169232-cust98.vm3.cable.virginm.net PRIVMSG #esolangs :Daft question. I'm looking for a "streaming" BF interpreter designed to work with extremely large programs. At a stretch I can handle 7GiB files with standard tools and a long start-up time but it would be nice to know if there's something that already exists that can stream programs that may be 30-50GiB or larger without pre-loading the whole
< 1740509599 605316 :CanisCorvus!~CanisCorv@shef-17-b2-v4wan-169232-cust98.vm3.cable.virginm.net PRIVMSG #esolangs :thing into memory before execution starts.
< 1740509871 259005 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :CanisCorvus: My recently-written interpreter with tagless-final encoding may be able to do that, incidentally, but it wasn't a design goal. As a starting point, are these programs inherently large or would they typically optimize to much smaller programs?
< 1740510014 533052 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :Also, are these programs generated by concatenating fragments of smaller programs, or are they irreducible into simpler operations?
< 1740510144 741153 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :FWIW anything over 4GiB of input text is not handled by the standard UNIX compiler lore; at 7GiB, you're definitely already working with unreasonably-large amounts of machine code, and it might be worth re-examining the original codegen strategy that led to this.
< 1740510207 353304 :CanisCorvus!~CanisCorv@shef-17-b2-v4wan-169232-cust98.vm3.cable.virginm.net PRIVMSG #esolangs :The programs are inherently large, optimized, and are a deliberate misuse of bf.
< 1740510258 290168 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :If you're willing to share the generator techniques, then I'm willing to experiment. You don't have to explain what you're computing but I'm quite interested.
< 1740510474 232484 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :A dissection of my interpreter is available here: https://pypy.org/posts/2024/11/guest-post-final-encoding-in-rpython.html Paraphrasing the summary: a final encoding builds input programs on the stack, emitting optimized actions to the heap; execution proceeds by interpreting the heap structure and occasionally JIT'ing.
< 1740510518 926305 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :If your program has pathologically deep loops, then the parser will fail; otherwise, the main modification I'd have to make is to stream the input program from a file on disk.
< 1740510561 135575 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :Oh, I guess "deep" is contextual here; BF Joust thinks 3 is deep. My interpreter's parser thinks that like 2000 is deep, and I only hit that with a fuzzer.
< 1740510837 682569 :CanisCorvus!~CanisCorv@shef-17-b2-v4wan-169232-cust98.vm3.cable.virginm.net PRIVMSG #esolangs :I don't have anything available online for the project yet so I'll have to get on that. Essentially it's an array initialisation followed by millions/billions of moves and outs. Though if I follow through on the dafter part of the plan there would be several blocks with some glue logic.
< 1740510941 384554 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :Okay. Do you do any I/O? There are several interpreters that should be able to stream non-interactive programs and evaluate them at compile-time, although I don't know how well they would perform.
< 1740510977 266404 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :...That alone is actually pretty good motivation for me, since my interpreter is pretty fast. The GNU Lightning intepreter linked on-wiki would also be a good candidate for modification.
< 1740511197 12984 :CanisCorvus!~CanisCorv@shef-17-b2-v4wan-169232-cust98.vm3.cable.virginm.net PRIVMSG #esolangs :Vast amounts of output. The standalone version is designed so that once the initial array initialisation is done, the rest of the program could be interpreted in chunks.
< 1740511301 352901 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :Alright, yeah, that sounds like something I'd be interested in. Ping me when you're ready to share. No rush; I have a few 
< 1740511325 171907 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :*few more weeks of work for customers. (I really need to fix my SSH.)
< 1740511372 232328 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :I'm not gonna ping anybody directly, but I think some folks highlight on "Brainfuck" in the logs, which neither of us said yet.
< 1740511385 443726 :APic!apic@apic.name PRIVMSG #esolangs :Good Night!
< 1740511397 596651 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :u'i Good night everybody~
< 1740511950 268198 :CanisCorvus!~CanisCorv@shef-17-b2-v4wan-169232-cust98.vm3.cable.virginm.net PRIVMSG #esolangs :Will do. Thanks!
< 1740512227 909272 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :No worries.
< 1740512304 641844 :zzo38!~zzo38@host-24-207-52-143.public.eastlink.ca PRIVMSG #esolangs :If a brainfuck program includes any commands that are outside of any loop, then they do not need to be kept in memory after they have been executed, since they are only executed once.
< 1740512422 665783 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :Yep. As such, an interpreter could partially evaluate its input program up to the first request for interactive I/O. But, just like with the streaming input feature, it's a design requirement.
< 1740514390 306017 :b_jonas!~x@88.87.242.184 PRIVMSG #esolangs :wow, multi-gigabyte brainfuck programs. this sounds esoteric.
< 1740514544 395136 :b_jonas!~x@88.87.242.184 PRIVMSG #esolangs :so the homepage of Vodafone has this weird login form where I can log in as a client of their mobile phone service providing. and I have found earlier that many times when I tried to log in, the form doesn't work, but this was inconsistent. specifically this is one of these overengineered client-side script forms that has the login button disabled by default, and it wasn't obvious when it worked and 
< 1740514550 681365 :b_jonas!~x@88.87.242.184 PRIVMSG #esolangs :when it didn't. anyway, yesterday I figured out one of the multiple factors about this, and it's so stupid that I have to share.
< 1740514748 932086 :b_jonas!~x@88.87.242.184 PRIVMSG #esolangs :apparently the form is trying to track whether I filled the login ID field with something syntactically valid, and whether I entered something nonempty into the password field. it doesn't enable the submit button until both of those are true. apparently when you type something into a field then it recalculates its idea about whether the data in that field is valid for just that field, but it does not 
< 1740514754 940164 :b_jonas!~x@88.87.242.184 PRIVMSG #esolangs :recalculate its idea when I paste the X selection into a field with middle mouse click. 
< 1740514896 944399 :ais523!~ais523@user/ais523 JOIN #esolangs ais523 :(this is obviously not my real name)
< 1740515320 460717 :Sgeo_!~Sgeo@user/sgeo JOIN #esolangs Sgeo :realname
< 1740515512 545849 :Sgeo!~Sgeo@user/sgeo QUIT :Ping timeout: 252 seconds
> 1740516459 445074 PRIVMSG #esolangs :14[[07Talk:SLet14]]4 10 02https://esolangs.org/w/index.php?diff=152715&oldid=152593 5* 03I am islptng 5* (+619) 10Removed redirect to [[Talk:SLet (Old 3)]]
< 1740517071 371008 :ais523!~ais523@user/ais523 PRIVMSG #esolangs :CanisCorvus: I think the only real issue with making a streaming BF interpreter would be matching the loop brackets – it'd be easy to write one which scans the program for the other end of the loop every time it encounters a loop, but also very slow
> 1740517083 446576 PRIVMSG #esolangs :14[[07Snakel14]]4 10 02https://esolangs.org/w/index.php?diff=152716&oldid=152556 5* 03Ractangle 5* (+6) 10/* Not specifying a type or specify a non-existent type to a variable (as long as it's not defined) */
> 1740517113 298677 PRIVMSG #esolangs :14[[07Snakel14]]4 10 02https://esolangs.org/w/index.php?diff=152717&oldid=152716 5* 03Ractangle 5* (+0) 10/* Not specifying a type or specify a non-existent type to a variable (as long as it's not defined) */
< 1740517155 156390 :ais523!~ais523@user/ais523 PRIVMSG #esolangs :if you can deal with some extra constraints on how the loop structure works (e.g. "the number of [ and ] instructions is small enough that all their locations can be fit into memory", or "[ and ] instructions can be annotated with comments showing the size of the loop") then you could work around that problem
> 1740517201 318945 PRIVMSG #esolangs :14[[07Comment14]]4 10 02https://esolangs.org/w/index.php?diff=152718&oldid=151505 5* 03Ractangle 5* (+3) 10/* MarkupL */
< 1740517244 227911 :ais523!~ais523@user/ais523 PRIVMSG #esolangs :another thing worth wondering is "can you fit the entire program into address space, even if it doesn't fit into memory?" – if you can, then you can make use of the OS page caching mechanisms more or less for free, which would give a nice performance boost without adding extra complexity
> 1740517314 757847 PRIVMSG #esolangs :14[[07Comment14]]4 10 02https://esolangs.org/w/index.php?diff=152719&oldid=152718 5* 03Ractangle 5* (+11) 10/* Gammaline languages */
> 1740517340 166723 PRIVMSG #esolangs :14[[07Comment14]]4 10 02https://esolangs.org/w/index.php?diff=152720&oldid=152719 5* 03Ractangle 5* (-10) 10/* Snakel */
< 1740517935 986386 :CanisCorvus!~CanisCorv@shef-17-b2-v4wan-169232-cust98.vm3.cable.virginm.net PRIVMSG #esolangs :The interpreter would ideally need a step to create a loop lookup file. One of the goals is that it should work on machines that don't have enough memory to hold the whole file.
> 1740518120 31309 PRIVMSG #esolangs :14[[07HQ9Sharp14]]4 N10 02https://esolangs.org/w/index.php?oldid=152721 5* 03I am islptng 5* (+1180) 10Created page with "{{wrongtitle|title=HQ9#}}  HQ9# is an esolang created by islptng. It is designed to be more usable than [[HQ9+]]. ==Commands== 
 H prints Hello World Hw prints Hello world HW prints HELLO WORLD hw prints hello world H, prints Hello, World H! prints Hello World
> 1740518308 360362 PRIVMSG #esolangs :14[[07HQ9Sharp14]]4 10 02https://esolangs.org/w/index.php?diff=152722&oldid=152721 5* 03I am islptng 5* (+82) 10/* Commands */
< 1740518372 970845 :tromp!~textual@2a02:a210:cba:8500:6ddc:c1a9:bc13:1391 QUIT :Quit: My iMac has gone to sleep. ZZZzzz…
< 1740518734 120527 :ais523!~ais523@user/ais523 PRIVMSG #esolangs :CanisCorvus: so modern computers have separate physical memory and address space – the address space is generally much larger
< 1740518736 534611 :tromp!~textual@2a02:a210:cba:8500:b949:287e:6bbd:873b JOIN #esolangs * :Textual User
< 1740518802 625546 :ais523!~ais523@user/ais523 PRIVMSG #esolangs :and I'd expect it to be bigger than the typical hard drive or SSD, although I'm not too tuned into just how big hard drives are getting nowadays
< 1740518854 458176 :ais523!~ais523@user/ais523 PRIVMSG #esolangs :having a file loaded into address space doesn't necessarily load the contents into physical memory, but you can mention a specific byte of the file and the processor and OS will cooperate to ensure it's loaded while you're accessing it (and will unload it again to save memory once it's no longer in use, if necessary)
< 1740518885 233915 :ais523!~ais523@user/ais523 PRIVMSG #esolangs :so it's an efficient way to handle very large files
< 1740519213 750304 :fizzie!irc@selene.zem.fi PRIVMSG #esolangs :I think spinning rust is 12, 16, maybe up to 20 terabytes these days.
< 1740519265 448426 :fizzie!irc@selene.zem.fi PRIVMSG #esolangs :(I looked into bumping my bulk storage drives from 4T to either 8T or 12T, but didn't do it yet.)
< 1740519281 485085 :b_jonas!~x@88.87.242.184 PRIVMSG #esolangs :and these days you can have file systems larger than a hard disk in at least three different levels (hardware raid, software raid, or ZFS file system level support for multi-volume file systems
< 1740519411 994148 :JAA!~JAA@user/meow/JAA PRIVMSG #esolangs :24 TB drives are readily available, and 26 TB ones are just about accessible to mere mortals. The bigger ones are large-scale customer exclusives.
< 1740519497 8283 :JAA!~JAA@user/meow/JAA PRIVMSG #esolangs :The sweet spot for price per storage is around 18-20 TB currently, at least where I am.
< 1740519535 474545 :JAA!~JAA@user/meow/JAA PRIVMSG #esolangs :24 TB are available from all manufacturers, I think. 26 TB not yet.
< 1740519692 18649 :zzo38!~zzo38@host-24-207-52-143.public.eastlink.ca PRIVMSG #esolangs :I have 500GB hard drive and only use 6%
< 1740519793 918099 :tromp!~textual@2a02:a210:cba:8500:b949:287e:6bbd:873b QUIT :Quit: My iMac has gone to sleep. ZZZzzz…
< 1740519881 9709 :tromp!~textual@2a02:a210:cba:8500:b949:287e:6bbd:873b JOIN #esolangs * :Textual User
< 1740519941 407352 :ais523!~ais523@user/ais523 PRIVMSG #esolangs :I'm surprising that the step up from "generally available" to "hard to get hold of" is such a small amount – if you needed more storage, wouldn't you just get two 24TB drives at that point?
< 1740520002 919955 :b_jonas!~x@88.87.242.184 PRIVMSG #esolangs :I bought 4 TB hard disks in 2020-09. wow, that was four years ago already! it feels like it was so recent. I think 16 TB ones were already available back then, and indeed now they are selling 26 GB ones for mortals. 
< 1740520011 595620 :fizzie!irc@selene.zem.fi PRIVMSG #esolangs :https://diskprices.com/?locale=uk is where I usually go look up the numbers.
< 1740520037 155201 :fizzie!irc@selene.zem.fi PRIVMSG #esolangs :On that page, cheapest price/TB numbers are in the 12, 14, 16 TB range, mostly.
< 1740520179 510228 :fizzie!irc@selene.zem.fi PRIVMSG #esolangs :(Also I never remember to toggle off the "Used" box.)
< 1740520281 447267 :b_jonas!~x@88.87.242.184 PRIVMSG #esolangs :I believe the prices are nontrivial to interpret because the different qualities of commercial hard disks come in different sizes, as in, the higher quality non-SMR drives are more likely to come in larger capacities and full size
< 1740520304 781354 :b_jonas!~x@88.87.242.184 PRIVMSG #esolangs :so the capacity isn't the only varaible 
< 1740520492 409220 :b_jonas!~x@88.87.242.184 PRIVMSG #esolangs :also it's still weird how large capacity SSDs have grown to
< 1740521047 890222 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :CanisCorvus: A good BF parser should be able to match all brackets at parse time, and so it shouldn't be a problem for the backend to only handle fully-formed loops. The main concern is that a single loop might be too large.
< 1740521145 960921 :b_jonas!~x@88.87.242.184 PRIVMSG #esolangs :hmm… brackets are probably well-nested so even for a large file you should be able to preprocess them into some helper table so that you can always find the matching one quickly by scanning only part of the source code, even if there are lots of brackets
< 1740521189 877210 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :Maybe we should clarify assumptions. Is the assumption that the program text fits on disk but not in RAM, or is the assumption that the program text is streamed into the interpreter?
< 1740521262 366512 :tromp!~textual@2a02:a210:cba:8500:b949:287e:6bbd:873b QUIT :Quit: My iMac has gone to sleep. ZZZzzz…
< 1740521372 188547 :tromp!~textual@2a02:a210:cba:8500:b949:287e:6bbd:873b JOIN #esolangs * :Textual User
< 1740521564 499289 :korvo!~korvo@2604:a880:4:1d0::4d6:d000 PRIVMSG #esolangs :CanisCorvus: ^^? Does this make sense? It's a big deal e.g. for video files, where seeking backwards is different for files on disk vs a network stream.
< 1740522066 388291 :tromp!~textual@2a02:a210:cba:8500:b949:287e:6bbd:873b QUIT :Quit: My iMac has gone to sleep. ZZZzzz…
< 1740522163 527039 :JAA!~JAA@user/meow/JAA PRIVMSG #esolangs :fizzie: Yeah, prices are wildly different by location. The Seagate Exos and Toshiba Enterprise Capacity models have been cheapest here for years now. I don't mind the 5 year warranty. :-)
< 1740522260 839469 :JAA!~JAA@user/meow/JAA PRIVMSG #esolangs :ais523: Most people would probably optimise for money/TB. If you absolutely need the density (e.g. limited connectors or tiny case), you'd go for the highest available capacity.
< 1740522309 587045 :ais523!~ais523@user/ais523 PRIVMSG #esolangs :right
< 1740522331 633526 :ais523!~ais523@user/ais523 PRIVMSG #esolangs :or perhaps "cheapest disk that has the capacity I need"
< 1740522340 738389 :JAA!~JAA@user/meow/JAA PRIVMSG #esolangs :Well, yeah
< 1740522374 986398 :JAA!~JAA@user/meow/JAA PRIVMSG #esolangs :At some point, tape becomes the more efficient approach. I haven't done the maths recently, but it's somewhere between 100 and 200 TB of data, I think.
< 1740522396 832288 :JAA!~JAA@user/meow/JAA PRIVMSG #esolangs :But whether that's a viable path depends on whether you need hot access or cold storage.
< 1740522731 594037 :JAA!~JAA@user/meow/JAA PRIVMSG #esolangs :I'd like to play with host-managed SMR, but those drives never make it to plebs like me.
> 1740523776 170558 PRIVMSG #esolangs :14[[07User:Aadenboy14]]4 M10 02https://esolangs.org/w/index.php?diff=152723&oldid=152572 5* 03Aadenboy 5* (+238) 10help me out here!
< 1740524565 59531 :zzo38!~zzo38@host-24-207-52-143.public.eastlink.ca PRIVMSG #esolangs :I thought of a variant of semaphore locking, which will be instead of a single number, the value will be a sequence of integers, where the first nonzero number must be positive (but all zero is also allowed). The usual semaphore is the case where this is only a single number. There are two other features as well.
< 1740524611 540203 :zzo38!~zzo38@host-24-207-52-143.public.eastlink.ca PRIVMSG #esolangs :Firstly, you should be able to have atomic locking of multiple such sequences at once. Secondly, you may have multiple adjustments in sequence where the value has to be valid after each intermediate value as well.
< 1740524950 465009 :zzo38!~zzo38@host-24-207-52-143.public.eastlink.ca PRIVMSG #esolangs :(Allowing atomic locking of several objects at once is also sometimes helpful, I think, even if they are remote objects in some cases.)
< 1740525114 353398 :zzo38!~zzo38@host-24-207-52-143.public.eastlink.ca PRIVMSG #esolangs :Although there is a few ways to implement reading/writing lock, with this you would have initial value [1,0], reading lock [0,-1], and writing lock [-1,0].
< 1740525534 224137 :ais523!~ais523@user/ais523 PRIVMSG #esolangs :zzo38: part of the reason that this sort of primitive isn't generally provided, even though it would be useful, is that it's hard to implement in hardware
< 1740525585 354540 :ais523!~ais523@user/ais523 PRIVMSG #esolangs :although by using a compare-and-exchange loop you can generally produce arbitrary atomic operations on aligned memory up to a given number of bytes (e.g. 16 bytes on x84_64) which might be enough to fit in a sequence if you were using small integers
< 1740525625 264296 :ais523!~ais523@user/ais523 PRIVMSG #esolangs :I think standard libraries and operating systems generally prefer to keep their primitives specialised so that they can be optimised for the purpose – locking and unlocking is often a performance bottleneck
< 1740526003 708152 :zzo38!~zzo38@host-24-207-52-143.public.eastlink.ca PRIVMSG #esolangs :OK, although there are other considerations like I mentioned, and specialized locking might make it difficult to make the more general case to be atomic, I might think (unless they had considered that in the design, although that might make it more complicated). The arbitrary atomic operations that you describe might help, although then the convention of their use must be used.
> 1740526333 731680 PRIVMSG #esolangs :14[[07SLet14]]4 10 02https://esolangs.org/w/index.php?diff=152724&oldid=152665 5* 03I am islptng 5* (+302) 10/* Quine */
> 1740527761 885130 PRIVMSG #esolangs :14[[07SLet14]]4 10 02https://esolangs.org/w/index.php?diff=152725&oldid=152724 5* 03I am islptng 5* (+140) 10/* Implementation */