00:07:06 -!- imode has joined.
00:16:52 -!- imode has quit (Ping timeout: 256 seconds).
00:27:33 -!- atslash has joined.
00:34:14 -!- Naergon has joined.
00:36:31 -!- imode has joined.
00:49:00 -!- imode has quit (Ping timeout: 245 seconds).
01:02:46 -!- oerjan has joined.
01:24:07 * boily semaphores bonsøœirjan
01:27:34 -!- boily has quit (Quit: CONTAINMENT CHICKEN).
02:07:25 <zzo38> I managed to only partially copy a ZIP file from a floppy disk onto my computer. How can I use that partial ZIP file?
02:18:27 <esowiki> [[Along and Across]] https://esolangs.org/w/index.php?diff=55340&oldid=55339 * Oerjan * (+6) /* Semantics */ Looks like an error
02:39:21 -!- MDude has quit (Ping timeout: 264 seconds).
02:43:26 -!- MDude has joined.
02:44:48 <zzo38> I found out that bsdtar will read it.
02:45:40 -!- Naergon has quit (Ping timeout: 245 seconds).
02:52:39 -!- Guest33130 has quit (Quit: WeeChat 1.9.1).
02:54:10 -!- GeekDude has joined.
02:55:58 -!- xkapastel has joined.
03:04:31 -!- variable has quit (Quit: /dev/null is full).
03:05:18 <oerjan> <wob_jonas> I'll answer that question then <-- and thus i finally learn what your family name is... at least i _think_ i would have remembered that if i'd seen it before.
03:07:38 <shachaf> b_jonas has certainly linked to his website here before
03:09:10 <oerjan> maybe it's _anti_-memorable, then.
03:23:40 -!- friendlyGoat has joined.
03:23:47 -!- friendlyGoat has quit (Remote host closed the connection).
04:10:00 -!- doesthiswork has quit (Quit: Leaving.).
04:14:16 -!- tromp has quit (Remote host closed the connection).
04:26:32 -!- Galaxtone has joined.
04:26:38 <Galaxtone> Test changes shows Schmuu! within the code block
04:26:46 <Galaxtone> but save changes shows it outside the code block
04:37:17 -!- doesthiswork has joined.
05:10:45 -!- imode has joined.
05:18:40 <shachaf> oerjan: So recursion depth is bounded and the other looping mechanism is also bounded. :-(
05:18:53 <shachaf> But I wrote http://slbkbs.org/kak/bf.kak
05:25:26 <oerjan> . o O ( shachaf knows Jelly? )
05:25:50 <shachaf> I just remembered there was some esolang that had a character set for 256 byes.
05:26:08 <shachaf> Probably more than one. I looked on PPCG until I found an example.
05:26:48 <shachaf> Oh, this is that editor I was talking about.
05:27:29 <oerjan> yeah it's very common for golfing languages made for PPCG
05:30:51 <oerjan> so it's just forever that isn't truly forever?
05:31:53 <shachaf> Yes, this one just does 1000000 steps.
05:35:56 <shachaf> Oh man, if you add ^L to each step you can see the program as it runs. This is much better.
05:39:53 <oerjan> . o O ( now use it to play Lost Kingdom )
05:52:22 -!- MDead has joined.
05:53:35 -!- MDude has quit (Ping timeout: 245 seconds).
05:53:37 -!- MDead has changed nick to MDude.
06:11:14 -!- oerjan has quit (Quit: Nite).
06:27:11 -!- variable has joined.
06:28:00 -!- variable has quit (Client Quit).
06:35:16 -!- arseniiv has joined.
06:53:06 -!- tromp has joined.
07:21:47 -!- doesthiswork has quit (Quit: Leaving.).
07:38:47 -!- wob_jonas has joined.
07:44:50 -!- imode has quit (Ping timeout: 245 seconds).
07:45:30 -!- xkapastel has quit (Quit: Connection closed for inactivity).
07:49:44 -!- erkin has joined.
07:52:08 -!- friendlyGoat has joined.
07:52:11 -!- friendlyGoat has quit (Remote host closed the connection).
08:00:38 <wob_jonas> So when I tell Excel to save a worksheet as unicode text, it actually encodes it in utf-16. That was confusing. I expected utf-8 and didn't understand why my program wasn't working.
08:06:41 <int-e> standards-complaint
08:10:40 <shachaf> oerjan: Oh, you can set up an action to run when something happens and then nothing else happens for some amount of time.
08:10:54 <shachaf> Which would technically make it TC, I suppose.
08:19:33 -!- erkin has quit (Quit: Ouch! Got SIGIRL, dying...).
08:19:56 <wob_jonas> oerjan: I actually have a split profile on Stack Exchange, with my real name showing up on MathOverflow and cstheory SE and math SE, but b_jonas on most other sites. Only my real name is the default, I can't easily change that, and I haven't bothered setting it to b_jonas on all SE sites.
08:20:46 <wob_jonas> shachaf: yes, multiple esolangs with such character sets, Jelly is one of them
08:22:26 <wob_jonas> mind you, a purist esolang would just assign meaning to bytes without caring about what character to represent them; or evenand possibly even have some pretty human-writable multi-character representation that's them compiled to compressed code for the language, with arithmetic compression so it isn't even structured in bytes
08:22:46 <wob_jonas> I know about jelly because ais523 programs in it
08:23:16 <shachaf> Of course those characters exist to display bytes that have meanings independent of the characters.
08:23:16 <wob_jonas> but I don't know much more about it than that it has that strange character encoding and some really strange rules with lots of heuristics and exceptions
08:23:48 <wob_jonas> shachaf: sure, but why would you have to display them as single characters?
08:24:04 <wob_jonas> you could display them as very short sequences of ascii characters instead.
08:24:18 <shachaf> Well, in my case, it's beause I wanted a compact way to display unprintable bytes.
08:24:52 <wob_jonas> shachaf: cp437 is _almost_ good for that, except that 0x20, 0x00, and 0xFF look exactly the same
08:25:45 <shachaf> I also wanted newline not to break the line in this case.
08:26:51 <wob_jonas> "<shachaf> oerjan: Oh, you can set up an action to run when something happens and then nothing else happens for some amount of time." => is that still about that editor with try-catch as the only control construct?
08:27:22 <wob_jonas> shachaf: cp437 has a printable character for newline. some sort of music note or sun or whatever. there's one for \r and one for \n .
08:28:13 <wob_jonas> Ah no, inverted hollow circle for \n and musical note for \r
08:33:10 <wob_jonas> My bitmap font actually has the cp437 values for most of the ascii control characters, and the cp437 values drawn in thin lines for the high control characters, only they're hard to show because most terminals and font engines won't display those characters, because the control characters aren't supposed to be drawn as a glyph from the font.
08:33:37 <wob_jonas> That's a feature, as in, when you print a newline, you want them to break the line instead of show an inverted circle.
08:33:53 <wob_jonas> I think you could still display them with custom code, but that's not helpful.
08:34:16 <wob_jonas> I mean, it would be easier for the font to just copy those glyphs on some other unused characters.
08:35:44 <wob_jonas> Also, xterm in byte encoding (as opposed to utf-8 encoding; xterm wants the font encoding to match the terminal encoding, which is really silly) uses a few of the control character places in the font as line-drawing characters that you access in the terminal with ANSI control codes to shift to the alternate character set.
08:48:43 -!- sprocklem has quit (Read error: Connection reset by peer).
08:49:01 -!- sprocklem has joined.
09:02:52 <Vorpal> wob_jonas: why would you use anything except pure UTF-8 these days though in a terminal?
09:02:59 <Vorpal> UTF-16 perhaps on windows
09:03:37 <wob_jonas> Vorpal: these days I do use utf-8 in the terminal, but like five or ten years ago that wasn't so obvious
09:04:02 <wob_jonas> I was slow to adapt utf-8 because people kept changing character encodings and saying that you should use the next one because it's the best thing every five years
09:04:59 <wob_jonas> it started with iso-646-* national variants before my time, then cp437/cwi, then cp850 on dos, then cp1252 on windows and iso-8859-2 on linux. after the third character set, you don't want to change to the next hot big thing everyone's talking about, just in case it won't stick
09:05:33 <wob_jonas> so for a while I ran them mixed, had both iso-8859-2 and utf-8 terminals mixed, then eventually I started using only utf-8
09:05:56 <wob_jonas> for terminals, that is. I still have some files encoded differently.
09:05:59 <Vorpal> I think i went utf-8 like 15 years ago or so
09:06:28 <Vorpal> well, I was on mac os 9 back then, so probably had some MacRoman too
09:07:52 <Vorpal> huh, google chrome finds "a" when I search for "å" in a page
09:08:24 <Vorpal> I wonder if it is related to LC_COLLATE
09:34:10 <fizzie> In Finland there was a brief moment of ISO-8859-15 between ISO-8859-1 and UTF-8.
09:38:25 <fizzie> I've booted to Windows (10) for the first time since November 2016, and it's having considerable trouble getting all the updates installed.
09:39:03 <Vorpal> i ran into an issue the other week, where I wouldn't get any user accounts listed on the login screen
09:39:10 <Vorpal> unless I booted into safe mode, where it worked
09:39:29 <Vorpal> the fix was to go to safe mode, open msconfig and change to "normal startup"
09:39:57 <fizzie> I'm on my fourth reboot now, and it somehow managed to start two different update mechanisms (Windows Update, and some sort of more dist-upgrade-ish thing) at the same time.
09:40:55 <Vorpal> yeah this will probably take a while, always does with windows
09:46:00 <fizzie> It's also a little bit awkward, because every time it reboots I have to unplug the monitor's USB cable, otherwise the keyboard won't work in GRUB prompt and I can't select Windows.
09:47:17 <wob_jonas> fizzie: luckily nobody adapted iso-8859-16 here, because the Euro was never introduced in Hungary, there was no such unicode character split as with the romanian s with comma, none of the missing letters used in Hungarian, and anyway, people used the new version of cp1250 instead.
09:48:11 <wob_jonas> Mind you, when MS added new characters including the euro sign and the s with comma to cp1250 and cp1252, they effectively created a new character set and possible confusion because they didn't give it a new name
09:48:32 <wob_jonas> that's sort of cheating, so I guess there is an extra step I didn't mention too
09:49:58 <wob_jonas> and of course cp1250 differs more from 8859-2 than cp1252 does from 8859-1, with some characters moved, but those didn't impact Hungarian texts
09:52:25 <wob_jonas> plus you can count cwi as a separate character set from cp437, and I think there were some legacy character sets used on old pre-IBM-PC computer sets that put a few accented letters to the ascii printable range but doesn't match the iso-646 assignment
09:54:11 <fizzie> Back in the university newsgroups there was a running joke to say "onneksi ääkköset eivät ole enää ongelma" ("fortunately <the accented characters in the Finnish alphabet> are no longer an issue"), every time they in fact were an issue, which they had been with a steady frequency for decades.
09:55:33 <fizzie> (fi:aakkoset = en:alphabet, so people use "ääkköset" to refer to the non-English parts.)
10:33:41 -!- variable has joined.
10:40:45 -!- KawaiiBunny has joined.
10:41:25 -!- KawaiiBunny has left.
11:10:58 -!- tromp has quit (Remote host closed the connection).
11:15:46 <HackEso> ympyrkeämme lojuvallanne imaamistani saansaanjohtamaustani uurtumaavanaan purjoita radallemäksyttä viertämilleni kohtaisemmissaaverkeamme jeenpaikaamin
11:16:07 <fizzie> Some bots[who?][weasel words] seem to think it's unnecessary.
11:17:13 <shachaf> Huh, I never thought about the fact that sorting networks don't need decidable equality between elements.
11:17:22 -!- tromp has joined.
11:26:00 <shachaf> So you can sort real numbers for example.
11:26:43 <int-e> Yeah, I didn't get the point that the outputs may be represented differently from the inputs.
11:28:55 <int-e> the program computing min(a,b) can be different from the program computing a and the program computing b.
11:29:02 <int-e> (in fact, it has to be different in some cases)
11:30:35 <shachaf> Oh, you mean you might have a different converging sequence or something
11:31:42 <shachaf> But you can also sort lists of streams, where that isn't true
11:31:46 <int-e> My confusion was that I thought that the output of a sorting network is a permutation of the inputs, and I didn't see how that could be subverted.
11:31:53 <shachaf> (And you still don't have decidable equality)
11:32:14 <int-e> shachaf: no, the same is true for streams
11:33:22 <int-e> I'm taking an intensional perspective here. If this were Haskell I would be looking at the thunks.
11:33:32 <shachaf> I guess that depends on how you represent them
11:33:38 <int-e> Which a pure program cannot observe.
11:34:00 <shachaf> This was the confusion I had too, which is why this was surprising
11:35:10 <shachaf> So programs that use decidable equality are getting more information in a sense than sorting networks are
11:35:54 <shachaf> Can they do the sort in fewer comparisons?
11:36:13 <shachaf> I think I've asked a similar question before, about <=>
11:41:11 <shachaf> Also what happens when you use a sorting network on a lattice?
11:41:44 <shachaf> Where the swap operation computes inf and sup
11:45:41 <shachaf> Hmm, also neat: https://cstheory.stackexchange.com/questions/25485/does-the-0-1-principle-apply-to-merge-networks
11:55:02 <int-e> shachaf: Hmm, sorting networks are meaningful for distributive lattices. For arbitrary lattices, different sorting network may give different results: https://paste.debian.net/1026562/
11:58:04 <int-e> Will the output always be sorted even in an arbitrary lattice? I don't know.
11:58:55 <int-e> (assuming one uses a sorting network that is valid for total orders)
12:15:20 -!- Guest27988 has changed nick to ocharles.
12:34:07 <wob_jonas> shachaf: I'd like sorting functions (in libraries) and sorted search trees that guarantee to work for non-transitive comparison operators too, in the sense that they don't segfault or otheriwse run into undefined behavior, don't lose or duplicate items, and
12:35:11 <wob_jonas> if you can divide the input set to a lower and a higher part such that the comparison function compares each item of the lower part as less than each item of the upper part, then the sorter function will sort each of the items of the lower part before each item of the upper part, and the search tree will find an item in the lower part if you search
12:37:15 <wob_jonas> You can imagine that as if you made a new comparison function that is actually transitive and antisymmetric, and whenever it says an item is less than the other, then the original (non-transitive comparison function) also say the first item is less than the second; and then the sort and search tree must behave as if you sorted by the new comparison
12:37:15 <wob_jonas> function, but they can sort equal elements in any way.
12:38:06 <wob_jonas> This would be useful so that if you want to run untrusted code in a sandbox, and that code wants to sort with a custom comparison function, you still get meaningful results. Also useful for trusted but buggy comparison functions.
12:39:01 <wob_jonas> Most search trees and sort functions actually satisfy that, but a few don't, notably very old versions of perl had a sort function that can segfault for an inconsistent comparison function. (Current perl still has some problems with sorting, but not to the segfault levels.)
12:39:45 <wob_jonas> So for some libraries, including the C++ standard library and the rust standard library, it's quite possible that all the library authors had to do was to write this as a guarantee in the docs for future compatibility, without changing a line of the actual code.
12:40:07 <wob_jonas> (I'll have to look at the details of some of the libraries to check if that's true.)
12:41:12 <Vorpal> wob_jonas: for C++ it could vary between implementations
12:41:25 <Vorpal> I seem to remember at least one std::sort that is a modified quicksort
12:41:38 <Vorpal> and one that was insertion sort based
12:42:36 <Vorpal> I'm* pretty sure your comparison function is supposed to produce a consistent ordering according to most language specifications and documentation
12:42:37 <wob_jonas> And when sqlite had the wrong comparison function, it did lose an item after all (but it still didn't segfault, which is nice).
12:42:47 <wob_jonas> Vorpal: yes, and that's the problem, or so I say.
12:43:02 <wob_jonas> The language specs shouldn't make that requirement.
12:43:19 <Vorpal> I think you are doing it wrong if int strcmp(char* a, char* b) { return random(); }
12:44:05 <Vorpal> I don't see the use for supporting something like that
12:45:49 <Vorpal> wob_jonas: yes it is probably useful to ensure it doesn't crash (for the sandbox case), but apart from that? Eh
12:47:52 <Vorpal> wob_jonas: also what do you think of languages like Python that provides a key function as opposed to a comparison function to the sort method?
12:48:16 <Vorpal> though I guess you could still overide __cmp__ or something like that on your type instead
12:48:26 <wob_jonas> Vorpal: the key function is a nice convenience, but sometimes you find it's not enough
12:49:28 <wob_jonas> Unless of course you can define custom comparison functions on your own types and use them as a key, but in that case you have the original problem of non-transitive user-defined comparison functions.
12:50:07 <Vorpal> hm it does have __ge__, __gt__, __lt__ and so on
12:50:19 <Vorpal> __cmp__ was apparently a python 2 thing
12:50:46 <Vorpal> wob_jonas: which apart from "don't crash from inside sandbox" I maintain isn't a problem
12:51:05 <Vorpal> could you elucidate why it is useful to support it?
12:51:36 <wob_jonas> Vorpal: maybe I've just seen too many buggy comparison functions. some are still out there and cause surprising bugs (though not segfaults most of the time)
12:52:14 <wob_jonas> Vorpal: but mostly I'm saying this because many of those functions could guarantee this freely, because the implementation already satisfies the stonger guarantee, just doesn't document it;
12:52:41 <Vorpal> possibly it should in fact raise some sort of exception if it detects that the comparison method is inconsistent, but I think it is bad to just quietly accept it
12:52:44 <wob_jonas> Vorpal: and some languages like haskell and rust and perl are already striving for the goal that they don't segfault when you write bad code unless you do certain unsafe things,
12:53:27 <Vorpal> because then the list is possibly not sorted as you intended and you might not notice it
12:53:48 <wob_jonas> plus in some of them you can also have values that aren't copiable, so it's likely already either impossible for the sort function to duplicate a value, or else it actually breaks the language guarantees if it does duplicate a value;
12:54:32 <wob_jonas> plus doing a wrong sort where an item is duplicated or lost can already cause security bugs or segfaults in user code that uses the result.
12:54:38 <Vorpal> how will uncopyable (or unmovable) values ever work with a sort function?
12:55:04 <wob_jonas> "possibly it should in fact raise some sort of exception if it detects that the comparison method is inconsistent, but I think it is bad to just quietly accept it " => perhaps, but you can't really strive for that, because in general it's impossible to check that the comparison function is consistent
12:55:04 <Vorpal> wob_jonas: sorting incorrectly can also lead to security bugs
12:55:19 <wob_jonas> Vorpal: they're still movable or swappable, just not copiable
12:55:45 <Vorpal> so it's like ConcurrentModificationException in java, it reports that on a best effort
12:56:23 <wob_jonas> Vorpal: even even trying to detect a bad sort function would likely make performance worse, or at least you'd have to change the code of the sort function
12:56:38 <wob_jonas> but just making the stronger guarantee is actually free in half of the functions, so you should just do it there
12:57:06 <wob_jonas> that would also make it clear by contrast where it's not guaranteed and where you can get segfaults
12:57:28 <wob_jonas> hmm, I should check if perl's numeric sort is actually buggy in the way I think it is, and report a proper bug if it is
12:57:44 <wob_jonas> because I think it does use an inconsistent function and that can cause security problems
12:57:54 <Vorpal> a wild idea though: Have a type or annotation system strong enough to be able to express contracts like "this is a sort function", then enforce running quickcheck like tests on functions that end up being used that way
12:58:02 <Vorpal> as part of the build process
12:58:03 <wob_jonas> so I should test that for myself, and report if it's bad, or convince myself that it's correct if it is correct
12:58:26 <wob_jonas> Vorpal: quickcheck functions aren't enough when sort functions can be wrong in very sneaky ways
12:58:54 <wob_jonas> there's a good reason why that sort bug I found in sqlite wasn't found for so many years, despite that it's been present
12:59:08 <Vorpal> wob_jonas: anyway, an incorrect ordering could still lead to bad bugs, even if elements aren't duplicated or lost
12:59:58 <wob_jonas> and some of those bugs are bugs that programmers should learn to avoid, which is a different thing we should spread to them
13:00:31 -!- doesthiswork has joined.
13:03:39 <wob_jonas> And sure, for those few functions where it wouldn't be easy to guarantee the behavior for wrong sort functions, yes, keep them and document that. There are really those algorithms.
13:41:33 -!- xkapastel has joined.
13:57:47 -!- qob_jonas has joined.
13:59:38 -!- qob_jonas has quit (Client Quit).
14:05:45 -!- SopaXorzTaker has joined.
14:11:38 <wob_jonas> Vorpal: ah yes. now I'm getting the GPDR spam, from websites I haven't used for ages and where it's not even clear why they even try to collect data:
14:14:22 <wob_jonas> in particular from (or in the name of) ARM, who ask to register on their site before you can download the technical reference manuals of their product, a completely ridiculous requirement
14:56:46 <fizzie> The ARM thing is terrible.
14:57:21 <fizzie> NVIDIA had something similar around CUDA SDKs. (Just got their GDPR spam the other day.)
15:09:59 -!- user24 has joined.
15:17:37 <Vorpal> gitlab was interesting, got a mail I had to accept their TOS again before I could continue using the service, blocking me at login.
15:17:46 <Vorpal> not explaining what they changed
15:18:12 <Vorpal> there should be a law that you must publish a diff
16:08:40 <wob_jonas> fizzie: and there's https://www2.opengroup.org/ogsys/catalog/t101
16:10:40 -!- wob_jonas has quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client).
16:42:18 -!- SopaXorzTaker has quit (Quit: Leaving).
16:43:14 <esowiki> [[Language list]] M https://esolangs.org/w/index.php?diff=55341&oldid=55338 * Soaku * (+11) Added Pepe
16:46:04 <zzo38> Another possibility for font encoding would be for the high byte of the font encoding to match the byte used after ESC ( to select that character set. Due to some things it is not work exactly, but it is close. That also allows you to define additional character sets.
16:47:42 <zzo38> (This also enables access to some of the characters that are not available in Unicode. When Unicode is selected, a different font would be used.)
16:55:18 -!- SopaXorzTaker has joined.
16:55:44 -!- imode has joined.
16:56:14 <Vorpal> zzo38: isn't the point of unicode to have all characters?
16:58:21 <zzo38> Maybe, but it doesn't. Unicode is a mess anyways. (Furthermore, the X core font system can't even properly implement astral characters; I have suggested how to fix this by use of a EnableLigatures flag, though)
17:06:08 <Vorpal> zzo38: who uses X server font rendering these days though? I thought every modern program rendered text client side, using freetype, pango and so on
17:06:30 <Vorpal> also, X will be replaced by wayland within a few years
17:08:44 <Vorpal> so do I, but I bet it will only be for a couple more years
17:11:55 <zzo38> I do not intend to replace it; I still want to use xterm and bitmap fonts and Xaw and so on
17:12:55 <Vorpal> eh, there will be xwayland to run X programs under wayland
17:13:10 -!- imode has quit (Ping timeout: 245 seconds).
17:13:53 <zzo38> Will Wayland even work without a desktop environment installed like X does?
17:15:14 <zzo38> X also supports a default background stipple (although you need the -retro option, but I think it should be the default); I don't know if Wayland requires a full picture instead
17:15:38 <Vorpal> you could just take a picture of it?
17:16:06 <zzo38> I could, but it just takes up a lot of memory
17:16:41 <Vorpal> also it is rather ugly I think
17:17:05 <Vorpal> anyway I assume you can run with a window manager in wayland
17:18:12 <zzo38> I don't know if it supports window managers like X does because I haven't checked
17:18:52 <zzo38> And it isn't ugly; it is good. There is also the X cursor shapes, I don't think Wayland supports that either
17:21:19 <zzo38> (You probably can use larger cursors instead, and I like the small monochrome ones)
17:22:01 <zzo38> (There is also X resource manager, although you can use that independently of X.)
17:23:47 -!- wob_jonas has joined.
17:24:10 <wob_jonas> "<Vorpal> X will be replaced by wayland within a few years" => I don't believe that. Most of the things people say about wayland aren't actually valid advantages over X.
17:24:39 <wob_jonas> Vorpal: the part about X server-side font rendering versus in-client pango mapped stuff is more valid of course.
17:24:55 <zzo38> Yes and I agree; I think most of them are disadvantages over X.
17:25:30 <zzo38> I wonder if there is a way to run Wayland programs in X so that you can use a X window manager with Wayland programs.
17:27:46 <zzo38> The bitmap fonts used by X look better, especially at small sizes.
17:27:54 <wob_jonas> zzo38: I'm especially annoyed by the supposed security advantages that wayland would have if you use untrusted clients, because those are literally impossible
17:28:35 <Vorpal> wob_jonas: several of the devs working on X say it is an unmaintainable code base
17:29:00 <Vorpal> also it is not secure at all, one program can easily read input for another program
17:29:13 <zzo38> You could use a proxy server for security
17:29:22 <wob_jonas> zzo38: that's not a good excuse, because the client-side font rendering pipeline (freetype, fontconfig, pango, ...) can render X bitmap fonts among other formats, it's just that the fontconfig settings have an option that is set to hide them by default on some distributions.
17:30:20 <wob_jonas> "<Vorpal> also it is not secure at all, one program can easily read input for another program" => that's the claim that annoys me, when people want their flashy half-transparent borderless overlay toolbar windows, and at the same time they try to pretend that it's somehow clear to the user which program he's giving input to.
17:30:50 <Vorpal> wob_jonas: I hate semi-transparent, it just makes it harder to read everything
17:31:25 <zzo38> I have made some ideas about how to make X version 12, which compared with version 11 incorporates some features of extensions in core (such as XkbBell) and drops most of the other ones, as well as add some other stuff (such as the EnableLigatures flag for font loading, the ExtensionControl request, and HoldAndModify visuals).
17:31:52 <Vorpal> wob_jonas: but look at for example the XTEST extension (used legitimately mostly)
17:32:50 <zzo38> Unlike X11, the first byte of the X12 protocol is always "X", so you can easily implement both versions on the same server for compatibility if you want to do.
17:33:05 <Vorpal> wob_jonas: but if things like snaps or other generic prepackaged app formats are to succeed you need to be able to limit access and have security boundaries, like for example Android does. Yes an app can request being able to see other apps, and the used should be informed to be able to determine if it is legitimate
17:33:45 <zzo38> Vorpal: Then perhaps you should use proxy servers like I suggested, then.
17:33:50 <wob_jonas> Vorpal: that barely works in practice even on android, and android runs apps in full-screen, and even there an app could mimic the interface of another app to confuse the user.
17:34:28 <Vorpal> wob_jonas: it can mimic, can't read input sent to another program though
17:35:06 <wob_jonas> Vorpal: right. but that gets much more confused if you don't have all apps in full-screen all the time, but in windows.
17:35:46 <Vorpal> does it though, now that android allows PiP as well as split-screen
17:37:58 <Vorpal> anyway, if the X code base is considered unmaintainable by the people who develop it, then something clearly needs to be done. They have selected to begin a new project, and are probably going to leave X behind, not develop new features for it, and eventually stop supporting it at all. Sure, it is open source so someone else can pick it up, but I doubt that will happen (beyond some enterprise distro keeping it alive with security
17:37:58 <Vorpal> fixes for another 5-10 years)
17:39:24 <Vorpal> wob_jonas: I don't see what X does better than wayland inherently (as opposed to things that it currently does better due to being more mature)
17:39:24 -!- Phantom_Hoover has joined.
17:39:24 -!- Phantom_Hoover has quit (Changing host).
17:39:24 -!- Phantom_Hoover has joined.
17:44:22 -!- LKoen has joined.
17:48:05 <esowiki> [[Totally Accurate Malbolge]] M https://esolangs.org/w/index.php?diff=55342&oldid=41794 * Kaa-kun * (+0)
17:48:36 <esowiki> [[Totally Accurate Malbolge]] M https://esolangs.org/w/index.php?diff=55343&oldid=55342 * Kaa-kun * (+0) more caps
17:49:08 <HackEso> aglist 593,594,595: b_jonas shachaf
18:23:46 -!- wob_jonas has quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client).
18:31:44 <shachaf> int-e: So you can think of the intensional and extensional perspective of the same lattice, as you mentioned above
18:33:37 <shachaf> if A,C are indistinguishable then that sorting algorithm produces the same result on the extensional total order.
18:43:32 -!- contrapumpkin has quit (Ping timeout: 245 seconds).
18:52:04 -!- SopaXorzTaker has quit (Remote host closed the connection).
18:59:55 -!- atslash has quit (Quit: This computer has gone to sleep).
19:00:30 -!- Slereah has quit (Ping timeout: 252 seconds).
19:01:20 -!- Guest6451 has joined.
19:59:03 -!- arseniiv has quit (Ping timeout: 256 seconds).
20:29:21 -!- Camto has joined.
20:52:06 -!- ais523 has joined.
21:30:51 -!- user24 has quit (Quit: Leaving).
21:38:19 -!- Camto has quit (Quit: Page closed).
22:01:12 -!- xkapastel has quit (Quit: Connection closed for inactivity).
22:15:44 <esowiki> [[Along and Across]] https://esolangs.org/w/index.php?diff=55344&oldid=55340 * Challenger5 * (+303)
22:17:16 -!- user24 has joined.
22:26:44 -!- Phantom_Hoover has quit (Remote host closed the connection).
22:27:48 -!- LKoen has quit (Remote host closed the connection).
22:45:24 -!- Sgeo has joined.
22:46:19 -!- tromp has quit (Remote host closed the connection).
22:46:46 -!- Sgeo_ has quit (Ping timeout: 252 seconds).
22:48:50 -!- Sgeo_ has joined.
22:51:08 -!- Sgeo has quit (Ping timeout: 252 seconds).
23:03:32 -!- Galaxtone has quit (Ping timeout: 245 seconds).
23:06:21 -!- impomatic has quit (Ping timeout: 264 seconds).
23:23:48 -!- xkapastel has joined.
23:26:52 -!- LKoen has joined.
23:31:55 -!- user24 has quit (Ping timeout: 245 seconds).
23:34:06 -!- tromp has joined.
23:34:48 <zzo38> Should they add a "goto case" and "goto default" commands into C?
23:35:40 <zzo38> (The "goto case" is followed by a numeric expression, which does not have to be constant; then it can check which case (which is constant) is matching, or default if there isn't any match)
23:42:42 -!- tromp has quit (Ping timeout: 245 seconds).
23:44:45 -!- user24 has joined.
23:47:01 -!- user24 has quit (Client Quit).
23:48:59 -!- boily has joined.
23:49:07 <fungot> boily: mhm noise ' cause i was getting a little juicier well
23:49:15 <boily> fungot: oh yeah juicy.
23:49:15 <fungot> boily: and uh i don't know i think
23:49:43 <boily> @tell oerjan bonsϿirjan. a little song: https://youtu.be/4vgcYBwyw28
23:52:00 -!- LKoen has quit (Quit: “It’s only logical. First you learn to talk, then you learn to think. Too bad it’s not the other way round.”).