| 00:01:58 | * | ftsf_ joined #nim |
| 00:06:07 | * | AnonymousCoward quit (Ping timeout: 260 seconds) |
| 00:16:54 | * | gangstacat quit (Ping timeout: 246 seconds) |
| 00:21:47 | * | gangstacat joined #nim |
| 00:28:15 | * | yglukhov joined #nim |
| 00:32:48 | * | yglukhov quit (Ping timeout: 260 seconds) |
| 00:40:50 | * | jivank quit (Ping timeout: 256 seconds) |
| 00:41:07 | * | jivank joined #nim |
| 00:45:38 | * | chemist69 quit (Ping timeout: 258 seconds) |
| 00:48:12 | * | djellemah quit (Ping timeout: 256 seconds) |
| 00:48:45 | * | chemist69 joined #nim |
| 00:50:48 | * | djellemah joined #nim |
| 01:06:51 | * | gangstacat_ joined #nim |
| 01:06:57 | * | gangstacat quit (Ping timeout: 246 seconds) |
| 01:07:04 | * | gangstacat_ is now known as gangstacat |
| 01:07:28 | * | jivank quit (Ping timeout: 256 seconds) |
| 01:07:44 | * | jivank joined #nim |
| 01:10:38 | * | yglukhov joined #nim |
| 01:15:16 | * | yglukhov quit (Ping timeout: 260 seconds) |
| 01:35:05 | * | Demos joined #nim |
| 01:38:09 | * | Demos_ quit (Ping timeout: 258 seconds) |
| 01:51:23 | * | yglukhov joined #nim |
| 01:51:29 | * | gangstacat quit (Ping timeout: 260 seconds) |
| 01:55:17 | * | Sergio965 quit (Ping timeout: 248 seconds) |
| 01:55:52 | * | yglukhov quit (Ping timeout: 260 seconds) |
| 01:57:09 | * | gangstacat joined #nim |
| 01:59:04 | * | ibk joined #nim |
| 02:08:00 | * | saml_ joined #nim |
| 02:34:09 | * | yglukhov joined #nim |
| 02:38:36 | * | FreezerburnV joined #nim |
| 02:38:48 | * | yglukhov quit (Ping timeout: 260 seconds) |
| 02:47:53 | * | chemist69 quit (Ping timeout: 256 seconds) |
| 03:01:52 | * | chemist69 joined #nim |
| 03:05:53 | * | Varriount|Mobile joined #nim |
| 03:07:25 | Varriount|Mobile | Araq: I've done some translation of the Windows header files. It looks like C2Nim is going to need some way to interpret/translate pushing pragmas |
| 03:07:42 | Varriount|Mobile | *packing pragmas |
| 03:09:09 | * | dddddd quit (Quit: Hasta otra..) |
| 03:13:41 | * | ekarlso_ quit (Ping timeout: 248 seconds) |
| 03:14:03 | * | joebo quit (Ping timeout: 250 seconds) |
| 03:15:57 | * | Dankrad quit (Ping timeout: 240 seconds) |
| 03:16:31 | * | joebo joined #nim |
| 03:16:36 | * | yglukhov joined #nim |
| 03:16:52 | * | ekarlso_ joined #nim |
| 03:21:17 | * | yglukhov quit (Ping timeout: 260 seconds) |
| 03:26:53 | * | FreezerburnV quit (Quit: FreezerburnV) |
| 03:28:42 | * | PMunch quit (Quit: leaving) |
| 03:33:09 | * | FreezerburnV joined #nim |
| 03:46:31 | * | FreezerburnV quit (Quit: FreezerburnV) |
| 03:50:39 | * | Jesin joined #nim |
| 03:58:56 | * | yglukhov joined #nim |
| 04:05:08 | * | yglukhov quit (Ping timeout: 260 seconds) |
| 04:19:33 | * | melike quit (Remote host closed the connection) |
| 04:30:53 | * | ivone quit (Remote host closed the connection) |
| 04:31:54 | * | clio joined #nim |
| 04:38:06 | * | clio quit (Remote host closed the connection) |
| 04:38:52 | * | vendethiel joined #nim |
| 04:41:17 | * | yglukhov joined #nim |
| 04:45:44 | * | yglukhov quit (Ping timeout: 260 seconds) |
| 04:47:54 | * | vendethiel quit (Ping timeout: 256 seconds) |
| 05:02:28 | * | Varriount|Mobile quit (Read error: Connection reset by peer) |
| 05:06:03 | * | saml_ quit (Remote host closed the connection) |
| 05:09:42 | * | space-wizard quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 05:10:54 | * | space-wizard joined #nim |
| 05:11:14 | * | space-wizard quit (Client Quit) |
| 05:11:45 | * | space-wizard joined #nim |
| 05:12:02 | * | space-wizard quit (Client Quit) |
| 05:12:32 | * | space-wizard joined #nim |
| 05:12:50 | * | space-wizard quit (Client Quit) |
| 05:13:20 | * | space-wizard joined #nim |
| 05:13:38 | * | space-wizard quit (Client Quit) |
| 05:14:09 | * | space-wizard joined #nim |
| 05:14:26 | * | space-wizard quit (Client Quit) |
| 05:14:57 | * | space-wizard joined #nim |
| 05:15:14 | * | space-wizard quit (Client Quit) |
| 05:19:25 | * | def- quit (Quit: -) |
| 05:23:39 | * | yglukhov joined #nim |
| 05:28:12 | * | yglukhov quit (Ping timeout: 260 seconds) |
| 05:31:25 | * | Demos_ joined #nim |
| 05:33:32 | * | def- joined #nim |
| 05:34:17 | * | Demos quit (Ping timeout: 258 seconds) |
| 05:41:21 | * | Dankrad joined #nim |
| 05:59:31 | * | Vendan quit (Quit: ZNC - http://znc.in) |
| 06:04:10 | * | yglukhov joined #nim |
| 06:08:48 | * | yglukhov quit (Ping timeout: 260 seconds) |
| 06:12:52 | * | byte512 joined #nim |
| 06:17:11 | * | yglukhov joined #nim |
| 06:26:00 | * | space-wizard joined #nim |
| 06:28:19 | * | gokr joined #nim |
| 06:31:08 | * | nsf joined #nim |
| 06:31:37 | * | Dankrad quit (Ping timeout: 240 seconds) |
| 06:34:57 | * | bjz joined #nim |
| 06:40:26 | * | bjz quit (Read error: Connection reset by peer) |
| 06:52:54 | * | bjz joined #nim |
| 07:04:42 | * | ftsf_ quit (Quit: :q!) |
| 07:05:08 | * | space-wizard quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 07:08:26 | * | yglukhov quit (Remote host closed the connection) |
| 07:37:27 | * | Sembei quit (Ping timeout: 256 seconds) |
| 07:44:41 | * | Trustable joined #nim |
| 07:45:45 | * | djiin quit (Remote host closed the connection) |
| 07:46:03 | * | djiin joined #nim |
| 07:54:57 | * | chemist69 quit (Ping timeout: 240 seconds) |
| 07:59:45 | * | chemist69 joined #nim |
| 08:06:18 | * | NhanH quit (Ping timeout: 245 seconds) |
| 08:08:17 | * | NhanH joined #nim |
| 08:13:07 | * | Andris_zbx joined #nim |
| 08:27:01 | * | yglukhov joined #nim |
| 08:28:08 | * | yglukhov quit (Remote host closed the connection) |
| 08:28:21 | * | yglukhov joined #nim |
| 08:28:31 | * | yglukhov quit (Remote host closed the connection) |
| 08:29:04 | * | yglukhov joined #nim |
| 08:39:51 | * | LeNsTR is now known as lenstr |
| 08:58:23 | * | brechtm joined #nim |
| 08:58:53 | * | nicanaca0 quit (Remote host closed the connection) |
| 09:02:46 | * | nicanaca0 joined #nim |
| 09:30:01 | * | Arrrr joined #nim |
| 09:30:54 | * | Arrrr quit (Client Quit) |
| 09:31:50 | * | Arrrr joined #nim |
| 09:43:40 | * | Sembei joined #nim |
| 09:52:55 | * | ebarrett quit (Remote host closed the connection) |
| 09:52:56 | Calinou | dom96: oh, I've been thinking, I need to port the complete Nim feature list (that's on current website homepage, nim-lang.org) to the Jekyll prototype |
| 09:53:12 | Calinou | (to a dedicated Features page for example, or a page that's linked from Examples) |
| 09:53:27 | Calinou | not hard, maybe it can be shortened a bit, but I personally like such feature lists |
| 10:03:06 | * | chemist69 quit (Ping timeout: 250 seconds) |
| 10:05:35 | * | chemist69 joined #nim |
| 10:07:13 | * | Demos_ quit (Ping timeout: 258 seconds) |
| 10:09:03 | * | couven92 joined #nim |
| 10:19:15 | yglukhov | Is there some static_for macro to "iterate" over compile-time something and in the end produces its body repeatedly but parametrized with counter? |
| 10:20:55 | cheatfate | endragor: it looks like old asyncdispatch is pretty good with sequence of asyncCheck send() too... |
| 10:26:03 | FromGitter | <endragor> @cheatfate you mean before the fix? the problem may not occur every time you make multiple sends. but it may occur in case of app slowness or some other condition. There is no guarantee that the second send() will always return EAGAIN, because OS had time to do something between two calls. If you do 3 sends to the same socket, the results of syscalls can be: ⏎ ⏎ 1) EAGAIN ⏎ 2) OK ⏎ 3) EAGAIN |
| 10:26:03 | FromGitter | ... [https://gitter.im/nim-lang/Nim?at=5832cbbbb7cf637f2ac8e5dc] |
| 10:27:42 | FromGitter | <endragor> to simulate this you can insert some long for loop or a sleep into one of the send() callbacks, that should make it more likely to happen. |
| 10:28:17 | * | enthus1ast joined #nim |
| 10:28:49 | FromGitter | <cheatfate> @endragor nope, i'm talking about order of asyncCheck sends in output stream of data |
| 10:29:01 | FromGitter | <endragor> yes, that's what I'm talking about, too |
| 10:29:29 | FromGitter | <endragor> are you talking about the situation *before* or *after* the fix from @dom96? |
| 10:29:56 | FromGitter | <cheatfate> i'm talking about not fixed asyncdispatch |
| 10:30:37 | FromGitter | <cheatfate> because i have made test which makes many asyncCheck send calls (about 100) |
| 10:30:44 | FromGitter | <cheatfate> and all of them properly send data |
| 10:30:48 | FromGitter | <cheatfate> in right order |
| 10:31:22 | FromGitter | <cheatfate> i have tried small chunks of 128 bytes and bigger chunks about 128kb |
| 10:31:44 | FromGitter | <endragor> yeah, so what I'm saying is that not fixed asyncdispatch will keep proper order for `asyncCheck send()` most of the time. But rarely it may happen that the order is broken. Above I explained the scenario. ⏎ Yeah, a test that simply makes bunch of send() calls will not make it occur |
| 10:33:55 | FromGitter | <cheatfate> but not fixed asyncdispatch must complete all sequenced `asyncCheck send()` after first write event, am i wrong? |
| 10:35:30 | FromGitter | <endragor> I'm not sure what you mean by saying it "must complete all sequenced sends". it must invoke the first one for sure |
| 10:40:11 | * | zielmicha[m] joined #nim |
| 10:41:12 | FromGitter | <cheatfate> ```code paste, see link``` ⏎ ⏎ this is part of `poll()` and if `EvWrite` event happens and many callbacks stored inside of `data.writeCBs`, then all of them will be called after first incoming `EvWrite` event. So many `send(3)` syscalls will be made without waiting for appropriate `EvWrite`, agree? [https://gitter.im/nim-lang/Nim?at=5832cf47b7cf637f2ac8f96a] |
| 10:41:42 | * | Varriount|Mobile joined #nim |
| 10:42:17 | Varriount|Mobile | Araq: https://github.com/Varriount/Win2Nim |
| 10:43:18 | Varriount|Mobile | That script does a fairly good job of performing batch modifications on the Windows SDK files |
| 10:44:27 | Varriount|Mobile | There's still a number of additions that need to be made, but the basic framework is there. |
| 10:44:38 | * | chemist69 quit (Ping timeout: 245 seconds) |
| 10:45:52 | * | elrood joined #nim |
| 10:50:32 | FromGitter | <endragor> @cheatfate yes, and that's the problem. |
| 10:52:57 | cheatfate | endragor, but for some reason all my tests on `not fixed` asyncdispatch working good |
| 10:53:31 | Araq | Varriount|Mobile: thanks |
| 10:53:43 | * | prose[m] joined #nim |
| 10:53:43 | * | hohlerde joined #nim |
| 10:53:43 | * | MrAxilus[m] joined #nim |
| 10:53:43 | * | M-Quora joined #nim |
| 10:53:44 | * | jivank[m] joined #nim |
| 10:53:50 | * | terry[m] joined #nim |
| 10:53:51 | * | erwana[m] joined #nim |
| 10:53:51 | * | TheManiac joined #nim |
| 10:53:51 | * | Guest73656[m] joined #nim |
| 10:54:15 | * | dmi0 quit (Ping timeout: 246 seconds) |
| 10:56:42 | FromGitter | <endragor> @cheatfate because it's not a problem that's easy to track with tests. I explained the scenario where it may happen. Insert sleep() or CPU-burning code (some long `for` loops) into send() callback and you may be able to see it in tests. ⏎ The thing is - asyncdispatch was using an assumption that if one send() returned EAGAIN or partially consumed data, then the second send() will always return EAGAIN. But there is no such guarantee. |
| 10:56:42 | FromGitter | ... Anything may happen between two (or more) send() calls. |
| 10:58:37 | * | chemist69 joined #nim |
| 11:07:10 | cheatfate | i'm thinking to change send socket buffer size |
| 11:14:00 | * | Arrrr quit (Ping timeout: 260 seconds) |
| 11:20:52 | cheatfate | endragor: but anyway, less syscalls is always better |
| 11:21:14 | cheatfate | endragor: do you have made any tests on Windows? |
| 11:22:00 | FromGitter | <endragor> no |
| 11:26:44 | * | Arrrr joined #nim |
| 11:29:20 | * | arnetheduck joined #nim |
| 11:48:06 | * | dmi0 joined #nim |
| 11:56:01 | * | planhths joined #nim |
| 12:05:00 | * | planhths quit (Remote host closed the connection) |
| 12:05:25 | * | planhths joined #nim |
| 12:08:04 | * | brechtm_ joined #nim |
| 12:09:05 | * | planhths quit (Client Quit) |
| 12:09:45 | Araq | hmm we need a toString that takes a 'var string' to append to for speed |
| 12:10:22 | Araq | should we name it '$' or $-> or toString? |
| 12:10:34 | Araq | 345 $-> resultBuffer |
| 12:11:36 | * | brechtm quit (Ping timeout: 246 seconds) |
| 12:11:45 | Araq | any opinions? |
| 12:12:36 | Araq | or maybe resultBuffer &=$ 234 |
| 12:12:55 | * | planhths joined #nim |
| 12:15:53 | gokr | In Smalltalk such patterns often use the ... blablaOn: naming style. Like "streamOn:" or "printOn:" - though the argument is meant to be a stream, but almost the same I guess. |
| 12:16:20 | gokr | "$->" looks a bit... convoluted. |
| 12:18:35 | flyx | I think `$` is fine as infix, myString $ someValue |
| 12:19:06 | flyx | or maybe we can just overload `&` ? |
| 12:23:16 | Arrrr | i'd choose toString, it won't be too popular anyway |
| 12:25:35 | Arrrr | what's the problem with add? |
| 12:27:56 | Araq | Arrrr: there is no problem, just a minor fear that 'add' weakens the type checking too much |
| 12:28:15 | Araq | str.add(someInt) # hmm |
| 12:29:33 | Araq | I think I prefer x &=$ y which is just x &= $y but faster :-) |
| 12:29:49 | flyx | I think `&` standing for concatenation would be reasonable. what else would be the concatenation of a string and an int if not the int's representation concatenated to the string |
| 12:30:27 | Araq | flyx: '&' yields a string though, '&=' yields void |
| 12:30:40 | Araq | and the new operation would yield void too |
| 12:30:41 | flyx | ah yes |
| 12:30:47 | gokr | "x &=$ y" ... oh my. ;) |
| 12:31:14 | Araq | gokr: problem is that &=$ gets the wrong precedence |
| 12:31:25 | Araq | &$= would get the proper precedence ... |
| 12:31:54 | flyx | what about overloading &= |
| 12:31:59 | Araq | and at this point we shorten &$= to &= which is the universal alias for 'add' |
| 12:32:12 | Araq | so let's overload 'add' |
| 12:40:07 | elrood | seems like this should be a reasonably easy and simple optimization on some intermediate representation level, wouldn't it? does your special casing really generate a benefit compared to buf &= $foo ? |
| 12:46:30 | Araq | you can optimize &= $ to &=$ with a TR macro but that still means you need the &=$ as a primitive available |
| 12:47:03 | Araq | these primitives cannot be produced by inlining and specialization with today's compiler technology |
| 12:47:42 | Araq | or at least I can not envision how it would work, the transformation is not trivial |
| 12:50:53 | * | dmi0 quit (Ping timeout: 260 seconds) |
| 12:52:16 | * | lenstr quit (Excess Flood) |
| 12:53:06 | * | LeNsTR joined #nim |
| 12:57:44 | * | djiin quit (Ping timeout: 250 seconds) |
| 13:02:31 | elrood | hm, i'm curious how your code and the generated ir and assembly are going to look like then, would've thought this was merely about saving some allocations and copying, but i'm probably missing something important then |
| 13:05:36 | flyx | `$` is a procedure, so that optimization would be interprocedural. *even if* it could be implemented, compiler usually stay on intraprocedural level for optimization. |
| 13:10:05 | elrood | i am not convinced that's been true for any reasonably sophisticated compiler implementation at least since the mid-90s. inlining and specialization on datatypes isn't that esoteric and rare |
| 13:10:49 | * | brechtm_ quit (Remote host closed the connection) |
| 13:11:55 | * | brechtm joined #nim |
| 13:17:58 | * | Arrrr quit (Ping timeout: 245 seconds) |
| 13:18:45 | flyx | I would guess that a `$` implementation's IR is typically too complex for a compiler to consider it for inlining, but I may be wrong |
| 13:25:48 | elrood | i might be wrong just as well. we'd probably have to look at the generated code to get an idea of what's really happening and possible |
| 13:27:21 | flyx | assuming someone does some strange or clever stuff in `$`, I also don't think an algorithm can correctly transform it |
| 13:27:52 | * | couven92 quit (Read error: Connection reset by peer) |
| 13:28:19 | flyx | let's look at some arbitrary example I just invented: https://gist.github.com/flyx/1b55d79b94ef9aec07e519f234abd948 |
| 13:28:57 | flyx | do you think the compiler can transform the generated code of this so that it appends to an existing string? |
| 13:30:16 | flyx | it would need some complex knowledge about what happens here to do that. I don't think current analysis methods in compilers can provide that |
| 13:38:03 | * | chemist69 quit (Ping timeout: 258 seconds) |
| 13:40:04 | elrood | frankly, i don't see a problem. your function is pure, easily inlineable, operates on a fixed range of integer inputs, returns a 4-char string.. could you rephrase the question? |
| 13:41:13 | elrood | the practical solution would be to let your compiler do its work and see, though. a human's opinion is only so useful in this regard |
| 13:42:47 | * | planhths quit (Ping timeout: 260 seconds) |
| 13:43:08 | Araq | elrood: look at my patch and see for yourself. |
| 13:43:23 | Araq | it changes semantics from "return and append" to "append inplace" |
| 13:44:03 | flyx | elrood: assume the proc will be inlined. the originally generated IR consists of allocating a string on the heap and doing some operations on it, the ultimately returning a pointer to it [together with some dope]. so you think a compiler would be able to rewrite the algorithm to append to another string instead? |
| 13:44:59 | Araq | it implies changing the indexing from result[i] to result[i+base] where 'let base = result.len' (aka consider the existing length) |
| 13:45:30 | Araq | I claim that no compiler will do this heavy rewrite as an optimization. |
| 13:45:42 | * | flyx supports that claim |
| 13:46:39 | elrood | flyx, i'd think it should, yes, but as said, my opinion is not very useful here. we should see what our compilers of choice come up with |
| 13:46:55 | elrood | will have a look later, thanks, Araq |
| 13:47:08 | flyx | the compiler would need to generate code to possibly expand the existing string dependent on its capacity. so, it would need to generate implicit allocation code. |
| 13:47:15 | elrood | bbl |
| 13:56:06 | * | chemist69 joined #nim |
| 14:14:34 | FromGitter | <gogolxdong> How to let cpu execute machine code directly?Do you have some material to recommend? |
| 14:19:07 | cheatfate | gogolxdong: what does it mean directly? |
| 14:23:24 | * | arnetheduck quit (Ping timeout: 260 seconds) |
| 14:24:11 | FromGitter | <gogolxdong> I have a group of machine code corresponding to their assembly codes ,input these codes onto bare-metal and let cpu execute them. |
| 14:25:56 | FromGitter | <gogolxdong> such as let screen show hello world with colors and twinkle. |
| 14:27:31 | cheatfate | gogolxdong: abstract algorithm will looks like, allocate some memory, copy your machine code to memory, transfer code execution to that memory via, `jmp`, `call`, `ret` or any other target platform specific instruction |
| 14:28:30 | cheatfate | also you need to make `comeback` to your code, via `ret`, `jmp` or any other target platform specific instruction |
| 14:29:52 | cheatfate | but between allocate and copy there can be some kind of tricks which depends on target platform, for x86 memory page must be marked to have execution flags |
| 14:31:44 | cheatfate | also bare-metal have some kind of `bootloader` |
| 14:32:07 | cheatfate | which after doing some initialization procedures transfers code execution to some know specific address |
| 14:32:38 | cheatfate | and your machine codes must be placed in memory at this address, so `bootloader` will transfer execution to your code |
| 14:34:12 | FromGitter | <gogolxdong> This is a legacy issue which many modern advanced language compiled to c, then compiled to assembly , then compiled to machine code , then let cpu execute via what you describe above. I am conceiving a possibility to be compiled to machine code and make cpu execute directly without system. |
| 14:35:37 | * | PMunch joined #nim |
| 14:37:21 | cheatfate | gogolxdong: `system`, you mean operation system? |
| 14:37:31 | FromGitter | <gogolxdong> yes |
| 14:37:51 | cheatfate | gogolxdong: what type of cpu? |
| 14:38:01 | FromGitter | <gogolxdong> x86 |
| 14:38:33 | FromGitter | <gogolxdong> and arm,compatible |
| 14:41:11 | cheatfate | gogolxdong: http://zoo.cs.yale.edu/classes/cs422/2010/lec/l3-hw |
| 14:43:45 | FromGitter | <gogolxdong> Is this your homework? |
| 14:45:07 | cheatfate | gogolxdong: this was just proper `google` request, is it helps? |
| 14:45:50 | FromGitter | <gogolxdong> LUX ET VERITAS is Yale motto |
| 14:46:06 | FromGitter | <gogolxdong> yes, marked |
| 14:47:12 | * | planhths joined #nim |
| 14:57:16 | * | Senketsu joined #nim |
| 14:57:45 | * | jackv quit (Quit: ZNC - 1.6.0 - http://znc.in) |
| 15:00:44 | * | LeNsTR quit (Changing host) |
| 15:00:44 | * | LeNsTR joined #nim |
| 15:06:21 | * | jackv joined #nim |
| 15:09:06 | * | FreezerburnV joined #nim |
| 15:12:50 | * | FreezerburnV quit (Client Quit) |
| 15:15:00 | * | FreezerburnV joined #nim |
| 15:21:33 | * | chemist69 quit (Ping timeout: 256 seconds) |
| 15:24:16 | * | [CBR]Unspoken quit (Ping timeout: 256 seconds) |
| 15:25:26 | * | [CBR]Unspoken joined #nim |
| 15:26:57 | * | pregressive joined #nim |
| 15:27:23 | * | chemist69 joined #nim |
| 15:39:03 | * | Gonzih quit (Read error: Connection reset by peer) |
| 15:41:41 | * | PMunch quit (Quit: leaving) |
| 16:07:20 | * | planhths quit (Remote host closed the connection) |
| 16:07:49 | * | planhths joined #nim |
| 16:26:51 | * | FreezerburnV quit (Quit: FreezerburnV) |
| 16:43:48 | * | couven92 joined #nim |
| 16:48:47 | * | FreezerburnV joined #nim |
| 16:49:49 | * | FreezerburnV quit (Client Quit) |
| 16:50:12 | * | FreezerburnV joined #nim |
| 16:59:46 | * | Andris_zbx quit (Remote host closed the connection) |
| 17:07:44 | * | planhths quit (Ping timeout: 258 seconds) |
| 17:09:38 | * | chemist69 quit (Ping timeout: 245 seconds) |
| 17:11:17 | * | Varriount|Mobile quit (Ping timeout: 260 seconds) |
| 17:24:22 | * | FreezerburnV quit (Quit: FreezerburnV) |
| 17:27:50 | Calinou | hmm, Crystal site changed a bit: https://crystal-lang.org/ |
| 17:27:59 | * | chemist69 joined #nim |
| 17:35:51 | Calinou | how is the Gitbook effort for Nim going by the way? |
| 17:36:13 | Calinou | I remember someone was working on porting/creating Nim documentation in a GitBook, which offers a nice interface for reading, already used by Rust and Crystal |
| 17:40:04 | dom96 | Their sponsors are front and center |
| 17:40:06 | dom96 | we need that too |
| 17:40:22 | dom96 | Like I said, the front page should show panels with various highlights |
| 17:40:34 | dom96 | similar to how the top sponsors are presented on the crystal site |
| 17:43:35 | * | space-wizard joined #nim |
| 17:45:05 | dom96 | As a side note: I think we will need to consider moving away from bountysource |
| 17:45:16 | dom96 | The site is run by volunteers and has become incredibly unstable. |
| 17:45:38 | * | yglukhov_ joined #nim |
| 17:45:52 | dom96 | Oh, this appears to be the reason "<spunky> [23:53:44] Cofounder had a baby, skeleton staff atm" |
| 17:49:20 | * | yglukhov quit (Ping timeout: 256 seconds) |
| 17:50:14 | * | yglukhov_ quit (Ping timeout: 265 seconds) |
| 17:50:47 | * | Salewski joined #nim |
| 17:52:41 | Salewski | Dom, in my editor I have for each socket.connect() a newSocket() just before: https://github.com/ngtk3/NEd/blob/master/ned.nim#L1359 |
| 17:53:26 | Salewski | Is that good practice? Or may it be better to resue the same socket, so move newSocket at start of the proc? |
| 17:54:34 | * | yglukhov joined #nim |
| 17:56:46 | dom96 | Salewski: It's necessary. You can't reuse sockets. |
| 17:57:11 | Salewski | OK, thanks. |
| 17:57:45 | * | Varriount|Mobile joined #nim |
| 17:57:51 | * | Varriount|Mobile quit (Read error: Connection reset by peer) |
| 17:58:41 | * | Varriount|Mobile joined #nim |
| 17:59:05 | * | yglukhov quit (Ping timeout: 256 seconds) |
| 17:59:06 | * | Varriount|Mobile quit (Read error: Connection reset by peer) |
| 17:59:26 | * | Varriount|Mobile joined #nim |
| 17:59:27 | * | Varriount|Mobile quit (Read error: Connection reset by peer) |
| 18:01:03 | * | Varriount|Mobile joined #nim |
| 18:01:42 | * | kulelu88 joined #nim |
| 18:01:43 | * | gokr quit (Ping timeout: 268 seconds) |
| 18:02:36 | * | Varriount|Phone joined #nim |
| 18:02:38 | * | Varriount|Mobile quit (Read error: Connection reset by peer) |
| 18:05:37 | * | Varriount|Phone quit (Read error: Connection reset by peer) |
| 18:05:43 | * | [ui] joined #nim |
| 18:05:56 | * | Varriount|Phone joined #nim |
| 18:05:56 | dom96 | Salewski: Why are you putting so much code in a single file? |
| 18:06:21 | dom96 | You should honestly consider refactoring it into multiple modules, and also splitting the functions up a bit too |
| 18:08:08 | Salewski | Dom, I see no real benefit in splitting it into modules currently, and even have no real idea what part I may split. |
| 18:08:31 | Salewski | I may indeed refactor some functions. |
| 18:09:40 | Salewski | For your info, I think I found the reason for random gtk related crashes: It is that GTK is not thread save, and I did not care enough. |
| 18:10:11 | * | gangstacat quit (Quit: Ĝis) |
| 18:10:23 | Salewski | I think I fixed that, no chrashes for a few hours... |
| 18:11:41 | Salewski | Unfortunately nimsuggest can still crash in very rare situations, but I have already submitted minimal examples for that crashes, so I hope Araq will fix that soon... |
| 18:18:12 | * | byte512 quit (Ping timeout: 260 seconds) |
| 18:18:25 | * | Sembei quit (Ping timeout: 244 seconds) |
| 18:18:30 | * | ibk quit (Quit: Connection closed for inactivity) |
| 18:19:03 | * | Dankrad joined #nim |
| 18:19:21 | * | gangstacat joined #nim |
| 18:29:20 | Salewski | Does Google search for IRC logs works releable? Yesterday I forgott gcc option to save assembler listing, it is "save-temps" from http://irclogs.nim-lang.org/02-11-2016.html |
| 18:30:20 | Salewski | But google search "site:irclogs.nim-lang.org searchterm" seems to ignore many terms. |
| 18:30:27 | * | yglukhov joined #nim |
| 18:32:36 | * | brson joined #nim |
| 18:34:00 | * | gokr joined #nim |
| 18:34:30 | * | yglukhov quit (Ping timeout: 246 seconds) |
| 18:36:44 | * | Salewski quit (Quit: Salewski) |
| 18:36:54 | * | Salewski joined #nim |
| 18:42:45 | * | gokr quit (Ping timeout: 248 seconds) |
| 18:44:58 | * | gokr joined #nim |
| 18:46:19 | * | Gonzih joined #nim |
| 18:55:29 | * | Dankrad quit (Ping timeout: 265 seconds) |
| 19:02:30 | * | vendethiel joined #nim |
| 19:02:30 | * | Varriount|Phone quit (Read error: Connection reset by peer) |
| 19:02:40 | * | Varriount|Phone joined #nim |
| 19:04:49 | * | yglukhov joined #nim |
| 19:05:06 | * | Dankrad joined #nim |
| 19:08:25 | * | brechtm quit (Ping timeout: 260 seconds) |
| 19:09:21 | * | enthus1ast quit (Ping timeout: 256 seconds) |
| 19:24:58 | * | Ven_ joined #nim |
| 19:32:38 | * | Varriount|Phone quit (Read error: Connection reset by peer) |
| 19:32:51 | * | Salewski left #nim (#nim) |
| 19:40:55 | Calinou | dom96: ok, good to note for the site (sponsors centered) |
| 19:47:10 | * | Varriount|Phone joined #nim |
| 19:49:48 | * | Varriount|Phone quit (Read error: Connection reset by peer) |
| 19:51:42 | * | Ven_ quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 19:51:56 | * | libman joined #nim |
| 19:53:13 | * | PMunch joined #nim |
| 20:01:13 | * | pregressive quit () |
| 20:03:07 | * | Sembei joined #nim |
| 20:04:53 | * | Pisuke quit (Ping timeout: 256 seconds) |
| 20:10:26 | * | gokr quit (Ping timeout: 256 seconds) |
| 20:11:27 | * | [ui] quit (Quit: Connection closed for inactivity) |
| 20:27:15 | * | cheatfate quit (Ping timeout: 268 seconds) |
| 20:31:16 | Calinou | hmm, on Windows, the 0.15.2 64-bit .zip is not able to detect my 64-bit MinGW-w64 |
| 20:31:18 | Calinou | (version 6.2.0) |
| 20:31:25 | Calinou | can I specify its path manually? |
| 20:36:43 | Araq | Calinou: I don't think so, but all finish.exe is does is to patch your nim.cfg anyway |
| 20:36:50 | Calinou | ah |
| 20:38:02 | * | planhths joined #nim |
| 20:43:03 | * | planhths quit (Remote host closed the connection) |
| 20:43:36 | * | planhths joined #nim |
| 20:44:35 | * | planhths quit (Client Quit) |
| 20:46:46 | * | Varriount|Phone joined #nim |
| 20:48:52 | * | planhths joined #nim |
| 20:52:42 | * | gokr joined #nim |
| 20:56:38 | * | dddddd joined #nim |
| 20:56:47 | kulelu88 | Araq: how you do feel about OOP? |
| 21:02:24 | * | couven92 quit (Quit: Client disconnecting) |
| 21:02:35 | Araq | pretty much the same way I feel about FP. When performance is a concern, it's a liability. |
| 21:03:08 | libman | nim's license freedom score went from 88.918% (345 of 388 packages have license fitting copyfree.org definition) on Oct 31 to 88.491% (346 of 391) just now. I for one blame Global Warming. :P |
| 21:04:19 | libman | How much of my license nagging is the Nim community willing to tolerate? Because I can be a lot more annoying if I tried harder. ;) |
| 21:05:04 | kulelu88 | libman: you can annoy as much as you like, but asking the limited community on IRC to bully 3rd-party devs to unlicense their code will just shrink the community |
| 21:06:27 | kulelu88 | Araq: have you written python code that was OOP or procedural? (in the past) |
| 21:07:09 | libman | I definitely want to avoid anything that might "shrink the community". I'm hoping being the most "don't worry about the legal crap" language ecosystem would be a positive point for Nim. |
| 21:07:46 | kulelu88 | depriving people of using the GPL cause it exists, doesn't make much sense |
| 21:07:55 | * | planhths quit (Ping timeout: 265 seconds) |
| 21:08:17 | Araq | kulelu88: procedural code until I noticed Python's scoping rules are severely lacking for procedural, so I switched to packing stuff into a 'class' |
| 21:08:25 | libman | kulelu88: Please clarify your last statement. |
| 21:08:26 | Araq | very roughly speaking. |
| 21:09:22 | Araq | libman: I improved performance of JSON generation today. dig out some benchmark and take a look. |
| 21:09:24 | * | Demos joined #nim |
| 21:09:28 | kulelu88 | libman: does the GPL violate the copyfree.org definition ? |
| 21:09:39 | libman | Araq: yaaay! :D |
| 21:09:57 | libman | kulelu88: GPL is like the #1 worst offender. |
| 21:10:17 | Araq | that's more productive than arguing about Stallman being an Andrew Ryan |
| 21:10:35 | libman | I think Nim's approach to OOP (especially UFCS) is ideal. |
| 21:10:47 | kulelu88 | Araq: you should drop by in #gnu if you don't want to do any work for a day |
| 21:12:19 | libman | "I chose language X because it comes with the greatest quantity and diversity of legalese in its library ecosystem" - said no one ever. |
| 21:12:20 | kulelu88 | I tried convincing libman to return to civil society and become an internet millionaire but he chose not to :P |
| 21:13:57 | libman | My choice of Nim vs Go has become an epic philosophical crusade. I'll write about that later. |
| 21:14:42 | Araq | "I chose Nim because I'm not a sheep, nor an ape." |
| 21:15:18 | Araq | (^ don't take this too seriously) |
| 21:15:34 | libman | For me this is like marriage, but without no-fault divorce. I'd hate to become Internet's #1 Nim advocate only to be banished for my license nagging... |
| 21:16:11 | ldlework | Well you should certainly start writing articles rather than nagging. |
| 21:16:14 | ldlework | in realtime |
| 21:16:19 | kulelu88 | I find it odd that the Nim community isn't as big as it *could* be, considering how it always get great exposure on Hacker News (and possibly reddit too) |
| 21:16:31 | libman | ldlework: you are absolutely right about that. |
| 21:16:39 | ldlework | kulelu88: its kinda unstable right? |
| 21:17:19 | kulelu88 | Unless HN really is just noise and "flavour of the moment" |
| 21:17:19 | * | Matthias247 joined #nim |
| 21:18:04 | kulelu88 | has there been a "killer app" written in Nim yet? The Kubernetes kind ? |
| 21:18:22 | libman | I have a couple of killer apps in mind. ;) |
| 21:18:35 | dom96 | kulelu88: if there was one you'd know about it :) |
| 21:19:19 | libman | I hope dom96 and Araq will weigh in on Nim community's policy toward licenses and license nagging. |
| 21:19:53 | kulelu88 | libman: in order to continue nagging on licenses, I think we need to issue a code of conduct that is GPL |
| 21:20:06 | kulelu88 | no GPL is too soft. AGPL |
| 21:20:13 | libman | As far as I'm concerned, `nimble`s recommendation of MIT license is a huge benefit for Nim. |
| 21:20:49 | dom96 | libman: I would love to improve Nimble so that it is more aware of the different licenses out there. |
| 21:21:11 | Araq | libman: yeah good point, but I think that's exactly how much we should enforce this issue. A nice default for Nimble. |
| 21:21:15 | dom96 | This doesn't mean that it should quit with an "GPL sucks" error when people pick GPL |
| 21:21:16 | kulelu88 | dom96: when you get accepted into Y Combinator, will your startup use Nim ? |
| 21:21:37 | dom96 | But something impartial that provides more information about each license would be super nice |
| 21:21:45 | * | Demos_ joined #nim |
| 21:21:52 | dom96 | (If you want to add something like that to `nimble init` then please do so :)) |
| 21:22:02 | kulelu88 | dom96: I'd avoid doing that. don't 'guide' people to a license. Let them decide |
| 21:22:04 | libman | dom96: I have a license detection tool in Nim that I'm in the process of refactoring. I'll get back to you on that. |
| 21:22:40 | dom96 | kulelu88: Avoid what? Giving more info about licenses in Nimble? |
| 21:22:52 | dom96 | I wouldn't consider that "guiding" |
| 21:23:09 | libman | If I wrote `nimble init`, it would contain much code to electroshock people who choose restrictive licenses on compatible Samsung laptops. ;) |
| 21:23:15 | kulelu88 | dom96: I've never heard of a package manager providing license info before |
| 21:23:50 | dom96 | Perhaps I'm not making myself clear. |
| 21:23:58 | dom96 | I'm not entirely sure if my idea would work in a CLI app |
| 21:24:07 | Calinou | a "nimble license" command would be nice, it would tell you the license of a package (or your dependencies if you're in a project) |
| 21:24:11 | * | Demos quit (Ping timeout: 258 seconds) |
| 21:24:13 | Calinou | (maybe that exists already) |
| 21:24:16 | dom96 | But I'm thinking that `nimble init` should give you a choice of licenses with a short description of each |
| 21:24:17 | Calinou | I believe eg. Symfony has that |
| 21:24:34 | dom96 | or maybe some sort of ASCII table that compares them :) |
| 21:24:39 | libman | Of the package ecosystems I've analyzed, after Nim, Haskell does the second best job with package license metadata. Ruby, Node, FreeBSD ports are all very haphazard. |
| 21:26:15 | Calinou | is there a simpler syntax for creating arrays? http://nim-lang.org/docs/tut1.html#advanced-types-arrays |
| 21:26:25 | kulelu88 | dom96: when you get accepted into Y Combinator, will your startup use Nim ? |
| 21:27:20 | dom96 | kulelu88: hah, where did that question come from? |
| 21:27:30 | dom96 | Are you expecting me to be accepted into YC? |
| 21:27:31 | libman | My `lice` tool (currently broken) detects pretty well from among 757 license types scraped from Gentoo Portage. The goal is to have one command that tells you all your "legal obligations" for all dependencies (including base OS components, ports, and language libraries). |
| 21:27:38 | kulelu88 | I asked it above, but you probably missed it |
| 21:27:45 | dom96 | Yeah, I did, sorry. |
| 21:27:53 | dom96 | Good that you repeated it |
| 21:28:14 | kulelu88 | I hope you are dom96 . |
| 21:28:24 | dom96 | Anyway, answer is: of course. |
| 21:28:43 | kulelu88 | Unless you dream of working in a silo job for X-bank improving their Java backend |
| 21:29:20 | dom96 | Yeah... don't want that. |
| 21:29:39 | dom96 | I think I will try my hand at working freelance |
| 21:29:43 | dom96 | or remotely somewhere |
| 21:29:46 | dom96 | That would be nice. |
| 21:30:30 | dom96 | Of course, a startup would be nice too and I do consider it my ultimate goal. |
| 21:30:30 | libman | How about this deal: Nim has an "official recommendation" for using Copyfree licenses in all libraries (like two spaces instead of tabs), and have `nimble` issue a one-line warning about uncopyfree licenses (I'll provide the code). |
| 21:31:03 | kulelu88 | yeah, no that won't work libman |
| 21:31:41 | dom96 | libman: I would prefer for Nim and Nimble to remain neutral on this manner. |
| 21:32:00 | dom96 | It's better to educate people and let them come to their own conclusion. |
| 21:32:39 | dom96 | In addition, creating such a recommendation might prove a bit too divisive. |
| 21:33:04 | Calinou | so I don't understand http://nim-lang.org/docs/tut1.html#advanced-types-arrays because it sounds inflexible to me |
| 21:33:11 | Calinou | it tells me to create a IntArray type for eg. a array with 10 slots |
| 21:33:17 | kulelu88 | getting into code politics is the last thing a smaller language community needs |
| 21:33:24 | Calinou | what if I want an array with 15 slots later in the program? do I need to create another type? how am I going to call it? |
| 21:34:34 | dom96 | Calinou: Yes, a dynamic array is called a sequence in Nim: seq[string] |
| 21:34:36 | * | Demos joined #nim |
| 21:34:48 | Calinou | I saw that |
| 21:35:21 | dom96 | kulelu88: indeed |
| 21:37:21 | kulelu88 | dom96: are you guys keeping track of nim in production? |
| 21:37:33 | * | Demos_ quit (Ping timeout: 268 seconds) |
| 21:37:58 | dom96 | not really. |
| 21:38:09 | Calinou | hmm, Nim on Windows compiles my program but fails to link it, "Access denied" |
| 21:38:53 | * | Kingsquee joined #nim |
| 21:39:47 | libman | How about saying something like: "Hint: Nim and 79% of nimble packages use the MIT license. Please consider this for your libraries to keep things as simple as possible for everyone." |
| 21:40:39 | * | nsf quit (Quit: WeeChat 1.6) |
| 21:40:41 | * | ryanhowe joined #nim |
| 21:42:24 | kulelu88 | libman: I don't know if you're joking or not, but pushing the copyfree agenda here is no different to Richard Stallman popping by and saying "please use the GPL" |
| 21:43:02 | * | xet7 quit (Quit: Leaving) |
| 21:44:40 | libman | The difference would be who makes a persuasive argument why it would be to Nim's benefit to get the licensing culture right from the beginning. |
| 21:44:41 | * | Varriount|Phone quit (Read error: Connection reset by peer) |
| 21:44:54 | * | Varriount|Phone joined #nim |
| 21:47:20 | kulelu88 | libman: I am trying to understand your motives here. By 'forcing' people through subtle means to adopt 'copyfree' licensing, aren't you depriving them of the very freedom you want them to be liberated into? |
| 21:47:31 | libman | Many languages (ex Ruby, Mono) had to change licenses, and it wasn't pretty. A minute of time spent dealing with legalese is a minute not spent coding. |
| 21:48:02 | * | xet7 joined #nim |
| 21:48:23 | dom96 | libman: Tell you what. If you can get Armin Ronacher (who's a copyfree proponent IIRC) to become a Nim developer (this means he begins writing projects in Nim and writes at least one blog post about why Nim is awesome) then I will add that to Nimble :P |
| 21:49:07 | kulelu88 | that sounds like a rough deal for you dom96 |
| 21:50:18 | dom96 | lol |
| 21:50:20 | libman | kulelu88: I am very specific about terminology, and there's no "forcing" here. People should be free to use Nim with a "Earth Is Flat, Jews Have Tails, Heil Hitler" license (contrary to laws in some countries) - free speech. |
| 21:50:53 | def- | Calinou: sounds like a permissions problem for that directory and the user compiling |
| 21:51:03 | kulelu88 | you are forcing them though. RE: "Hint: Nim and 79% of nimble packages use the MIT license. Please consider this for your libraries to keep things as simple as possible for everyone." |
| 21:51:25 | kulelu88 | "please consider this" |
| 21:51:29 | libman | The word you're looking for is "nudging" - https://en.wikipedia.org/wiki/Nudge_theory |
| 21:52:35 | Calinou | def-: going to run Nim in WSL for now... it works there |
| 21:52:52 | libman | A lot less "force"ful than the "no tabulators" rule (which made me switch to the other side of that Holy War after 20 years). |
| 21:53:48 | Araq | Error: unhandled exception: An attempt was made to complete a Future more than once. |
| 21:53:58 | Araq | ^ what's that about? |
| 21:54:04 | kulelu88 | libman: if you wrote more code to improve the compiler than code that checks if licensing is correct, do you think that your efforts may become more valued ? |
| 21:54:20 | libman | dom96: I can only promise to increase my efforts at promoting Nim. I cannot promise changing a specific person's mind. |
| 21:54:20 | dom96 | Araq: future.complete() # ok; future.complete() # You're doing something wrong, future is already completed. |
| 21:55:15 | Varriount|Phone | Calinou: You get a permissions error if the executable file is locked (such as if it is currently running) |
| 21:55:34 | Calinou | I can't see why it would be running... I installed that MinGW long time ago |
| 21:55:47 | dom96 | Calinou: Can you gist the output? |
| 21:56:03 | * | ryanhowe quit (Quit: WeeChat 1.4) |
| 21:57:04 | Varriount|Phone | Calinou: When you got the error, were you *re*compiling a program? |
| 21:57:45 | Calinou | https://gist.github.com/Calinou/47fb6c3db6b21025ad8845216db75d2b |
| 21:57:51 | libman | kulelu88: it's not like I've been "cheating on Nim" with another language (unless shell script counts, LOL). I was just going through a phase where I wasn't coding at all, and now I'm about to start some "killer app" projects. |
| 21:57:51 | Calinou | Varriount|Phone: yes, but deleting nimcache does not solve it |
| 21:58:18 | kulelu88 | libman: like a 'killer' http lib? |
| 21:58:25 | Varriount|Phone | O_o |
| 21:59:03 | libman | "Where we're going, we don't need roads." |
| 21:59:55 | Varriount|Phone | Calinou: That's odd... I would expect "Access Denied", not "Refused" |
| 22:00:10 | Calinou | Varriount|Phone: in French it just means "Access denied" |
| 22:00:28 | Varriount|Phone | Ah, you're translating? |
| 22:00:42 | dom96 | Calinou: And it compiled successfully once already? |
| 22:00:51 | elrood | Calinou, have you checked whether there is a main.exe in your directory already and if you can delete that file manually? |
| 22:01:02 | dom96 | what happens when you run `nim c -o:foo.exe main.nim`? |
| 22:01:31 | Calinou | dom96: nope, never |
| 22:01:50 | Varriount|Phone | elrood: The Nim compiler isn't able to write out all the C files |
| 22:01:52 | Calinou | same error with that command, fails at linking |
| 22:02:02 | * | bjz quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…) |
| 22:02:13 | Calinou | it never writes out that .exe by the way |
| 22:02:16 | Varriount|Phone | Or it can't execute the compiler |
| 22:02:16 | Araq | dom96: "Stack trace to moment of secondary completion:" is mostly all within nim's stdlib |
| 22:03:30 | elrood | Varriount|Phone, are you sure? doesn't look like it to me, wouldn't the error pop up before the linking step then? |
| 22:03:45 | Varriount|Phone | Calinou: Can you give us the output with the verbosity option turned up? |
| 22:05:39 | Calinou | https://gist.github.com/Calinou/ac03a17b19c42dca895d302ac489377e |
| 22:05:53 | Calinou | so yeah, my MinGW path might not be great, it contains a space |
| 22:06:41 | dom96 | Araq: want me to take a look? gist it |
| 22:09:46 | Varriount|Phone | Calinou: Hm |
| 22:10:50 | Varriount|Phone | It's not GCC's linker, because the Nim compiler is the thing throwing the error. |
| 22:11:11 | Varriount|Phone | At least, If I recall correctly |
| 22:11:37 | Calinou | this is latest Nim 0.15.2, 64-bit, on Windows 10 AU |
| 22:11:47 | Calinou | the MinGW is also 64-bit |
| 22:11:59 | Calinou | I'd try Clang but Clang on Windows kinda looks messy |
| 22:13:02 | Varriount|Phone | Calinou: Could you check the permissions on the various directories to make sure you have write access? |
| 22:13:29 | Varriount|Phone | Also, what happens if you try this with an administrator command prompt? |
| 22:14:24 | Calinou | just tried with administrator command prompt |
| 22:14:32 | Calinou | I get some GUI telling me that libwinpthread-1.dll couldn't be found, twice |
| 22:14:42 | Calinou | those appear just before the linking step |
| 22:15:05 | Calinou | I guess I'll try another MinGW installation, which one do you recommend that's 64-bit and relatively recent? MSYS2? |
| 22:15:15 | Varriount|Phone | That might work |
| 22:15:21 | Araq | just use the one we ship with Nim |
| 22:15:35 | Araq | http://nim-lang.org/download/mingw64.zip |
| 22:16:10 | Calinou | ah, good |
| 22:16:32 | Araq | dom96: seems to be caused by this: |
| 22:16:35 | Araq | if future.failed: |
| 22:16:35 | Araq | resultFut.fail(future.readError()) |
| 22:16:40 | elrood | Araq, not seeing a lot of an improvement for your string.add performance patch to be honest, looks like it mainly shifts the work from allocating memory for a new string to allocating memory for growing an existing one here. perhaps the situation is different for unoptimized builds or the js backend though. if it works beneficially for you, great |
| 22:17:21 | Calinou | Araq: it works, thanks |
| 22:17:44 | Araq | elrood: well I did no benchmarks, but if the buffer is already longish it likely has enough space for the tiny stuff that's appended |
| 22:18:04 | Araq | saving GC runs |
| 22:18:07 | elrood | yeah, in that case it's an improvement for sure |
| 22:18:13 | dom96 | Araq: I doubt it |
| 22:18:19 | dom96 | Araq: Can I see the stack trace? |
| 22:19:17 | * | Trustable quit (Remote host closed the connection) |
| 22:19:55 | Calinou | Araq: what about bundling MinGW in the new .zips? |
| 22:19:56 | Araq | dom96: https://gist.github.com/Araq/33711e0bd25fe59bcf00f79581b5af2e |
| 22:20:04 | Calinou | so that it's already in dist/mingw, and with nim.cfg set accordingly |
| 22:20:40 | Araq | Calinou: the idea is that you update Nim more frequently than mingw and share mingw installations and bandwidth |
| 22:21:10 | dom96 | Calinou: You can use the installer, it will install MinGW as well for you. |
| 22:21:18 | elrood | Calinou, besides, unless mingw has dramatically reduced in size in the last years, that'd blow up the zip unproportionally |
| 22:21:45 | Araq | elrood: we actually have a "minimal" mingw ... |
| 22:21:59 | Araq | we trimmed it down by a hacky tool |
| 22:22:01 | * | Sentreen quit (Ping timeout: 256 seconds) |
| 22:22:12 | * | bjz joined #nim |
| 22:22:13 | Araq | not sure it was a good idea |
| 22:22:18 | dom96 | Araq: The usage of 'all' looks suspicious |
| 22:22:23 | dom96 | Perhaps it's to blame |
| 22:23:27 | Araq | waitFor all(matchmake(), matchmake()) # matchmake() returns a new future afaict |
| 22:23:43 | Araq | what's suspicious about that? |
| 22:25:51 | dom96 | I mean, the `all` procedure handles exceptions incorrectly. |
| 22:26:10 | dom96 | or I should say: they handle failed futures incorrectly |
| 22:26:18 | dom96 | s/they/it/ |
| 22:27:13 | Araq | oh? how about they work properly instead? :P |
| 22:27:38 | dom96 | what? |
| 22:28:07 | Araq | shouldn't it handle failures correctly? |
| 22:28:14 | Araq | what does it even mean? |
| 22:29:35 | dom96 | yes> |
| 22:29:45 | dom96 | I mean: its got a bug |
| 22:29:48 | dom96 | *it's |
| 22:30:04 | Araq | oh never mind, I'm tired |
| 22:30:24 | Araq | I mean, yes there is a bug here, most likely, but actually it doesn't concern me |
| 22:34:23 | * | bjz quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…) |
| 22:34:55 | * | Sentreen joined #nim |
| 22:41:09 | * | yglukhov quit (Remote host closed the connection) |
| 22:41:09 | * | Varriount|Phone quit (Read error: Connection reset by peer) |
| 22:41:23 | dom96 | indeed, looks buggy to me https://github.com/nim-lang/Nim/blob/devel/lib/pure/includes/asyncfutures.nim#L266 |
| 22:42:51 | * | Varriount|Phone joined #nim |
| 22:45:21 | * | Varriount|Phone quit (Read error: Connection reset by peer) |
| 22:45:44 | dom96 | Araq: try now |
| 22:45:44 | * | Varriount|Phone joined #nim |
| 22:45:53 | dom96 | I pushed a fix |
| 22:51:21 | * | Varriount|Phone quit (Ping timeout: 265 seconds) |
| 22:54:27 | Calinou | in Nim, how do I assign (overwrite) an existing sequence index? |
| 22:57:07 | Araq | dom96: sorry, will do later, as I said, it's not important for me right now |
| 22:57:41 | dom96 | Calinou: x[0] = value |
| 22:58:00 | Calinou | dom96: tried, and the compiler says it "can't be assigned to" |
| 22:58:18 | dom96 | use 'var' |
| 22:58:21 | dom96 | not 'let' |
| 22:59:15 | Calinou | ah, nevermind, there was some other issue with my code |
| 23:21:09 | * | Sergio965 joined #nim |
| 23:24:32 | * | brechtm joined #nim |
| 23:29:10 | * | brechtm quit (Ping timeout: 268 seconds) |
| 23:38:40 | * | Demos quit (Read error: Connection reset by peer) |
| 23:41:37 | * | yglukhov joined #nim |
| 23:46:10 | * | yglukhov quit (Ping timeout: 250 seconds) |
| 23:52:56 | * | Demos joined #nim |
| 23:56:53 | * | cheatfate joined #nim |
| 23:59:49 | * | desophos joined #nim |