| 00:00:35 | * | Xoc17 joined #nim |
| 00:00:38 | * | Xoc17 quit (Killed (Unit193 (Spam is not permitted on freenode.))) |
| 00:15:06 | * | nlsun joined #nim |
| 00:15:18 | * | nlsun is now known as Guest15279 |
| 00:15:45 | * | Guest15279 quit (Killed (Sigyn (Spam is off topic on freenode.))) |
| 00:24:04 | * | ddstreet14 joined #nim |
| 00:24:50 | * | ddstreet14 quit (Killed (Unit193 (Spam is not permitted on freenode.))) |
| 00:34:10 | * | wildlander quit (Quit: Konversation terminated!) |
| 00:40:26 | FromGitter | <arnetheduck> @data-man on #8185, in early versions of the patch, one would check for errors using -1, and later in your patch, it's... implementation-defined? |
| 00:41:24 | FromGitter | <data-man> @arnetheduck Yes. -1 was bad. |
| 00:42:14 | FromGitter | <arnetheduck> so what should the calling code do? compare with different values depending on what kind of stream it is? |
| 00:42:30 | * | pendo324 joined #nim |
| 00:43:30 | FromGitter | <data-man> Yes. E.g., I implemented the size for all stdlib's streams. |
| 00:43:31 | * | pendo324 quit (Killed (Sigyn (Spam is off topic on freenode.))) |
| 00:43:52 | * | arooni1 joined #nim |
| 00:44:11 | FromGitter | <arnetheduck> yeah. but imagine I'm using the interface.. isn't the idea for an interface of this kind that it behaves the same for all kinds of streams? |
| 00:44:13 | FromGitter | <data-man> Inspired by Delphi & FreePascal streams. :) |
| 00:44:21 | * | arooni1 quit (Killed (Sigyn (Spam is off topic on freenode.))) |
| 00:44:30 | * | PrimHelios joined #nim |
| 00:45:45 | FromGitter | <data-man> E.g. https://github.com/nim-lang/Nim/pull/8326 |
| 00:46:43 | FromGitter | <data-man> @arnetheduck If you has your own stream, you should implement the size for this stream. |
| 00:48:19 | FromGitter | <arnetheduck> I understand that whoever makes a stream must implement size. what I don't understand is what they're supposed to return in case of error.. since it's not -1, what should it be? conversely, if I'm calling size, how do I know what size will return, if it's an error? |
| 00:50:47 | FromGitter | <data-man> Why is there no such question to the result of getPosition? |
| 00:51:23 | FromGitter | <arnetheduck> maybe there should be? |
| 00:52:03 | FromGitter | <data-man> I don't know :) |
| 00:54:58 | FromGitter | <arnetheduck> well, that's the thing with standard libraries - they age badly, and you're left with poor choices - keep shitty code, break user code or make up new names with numbers in them.. if someone didn't think clearly when writing getPosition, that becomes and argument for doing the same when writing the next function as well (I'm guessing getPosition exists already before your patch :)) |
| 00:55:20 | * | dddddd quit (Remote host closed the connection) |
| 00:59:57 | * | arecaceae quit (Remote host closed the connection) |
| 01:00:12 | FromGitter | <data-man> I like the idea about splitting streams to readers and writers (as in Go, Rust or Zig). |
| 01:00:16 | * | arecaceae joined #nim |
| 01:01:52 | * | drazan joined #nim |
| 01:02:12 | FromGitter | <arnetheduck> I like the idea of streams being in a library so they can be iterated upon, without breaking user code that depends on a particular version of them ;) though if you break into r&w, you get another chance at designing a coherent API |
| 01:08:52 | * | stefanos82 joined #nim |
| 01:10:36 | FromGitter | <data-man> When system.nim will be splitted, readers and writers can be in the std.io. ;) |
| 01:15:05 | * | vosper quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 01:15:43 | * | Jesin joined #nim |
| 01:18:05 | * | a__b__m quit (Ping timeout: 240 seconds) |
| 01:36:54 | FromGitter | <arnetheduck> a *nimble* library, outside of the main nim distro |
| 01:38:27 | FromGitter | <arnetheduck> speaking of the std library, what's the difference between `core`, `pure`, and `std` and `packages`? |
| 01:45:01 | FromGitter | <data-man> IMHO, packages and core should be removed, std is new and has big potential. :) |
| 01:45:12 | FromGitter | <xDotDash> pure is for things written 100% in nim I believe |
| 01:45:44 | FromGitter | <data-man> Everything in the lib written in Nim :-D |
| 01:46:56 | FromGitter | <data-man> Except several *.h, sorry. |
| 01:47:51 | * | mwbrown quit (Ping timeout: 268 seconds) |
| 01:51:13 | * | mwbrown joined #nim |
| 01:51:19 | FromGitter | <arnetheduck> `pure` depends on `c` and `posix` libraries at least, so the definition would be "mostly nim except when not".. impure depends on non-nim stuff for sure though - I guess the difference lies in how exotic the dependency is :) |
| 01:53:05 | PrimHelios | perhaps pure is only dependent on C and posix, while impure is any external C lib? |
| 01:53:57 | PrimHelios | is dependent upon any external lib* |
| 01:54:57 | FromGitter | <data-man> Is libc the external lib? :) |
| 01:55:27 | * | mwbrown quit (Ping timeout: 240 seconds) |
| 01:57:00 | * | red-00120 joined #nim |
| 02:00:14 | PrimHelios | does libc count as an external lib? like 90% of C code depends on it in some way |
| 02:02:35 | * | red-00120 quit (Ping timeout: 240 seconds) |
| 02:04:37 | FromGitter | <data-man> And remove dependency from libc is the biggest task. |
| 02:08:03 | * | developers joined #nim |
| 02:12:51 | * | developers quit (Ping timeout: 240 seconds) |
| 02:16:04 | FromDiscord_ | <emekoi> do converter type classes work yet? |
| 02:19:37 | FromGitter | <kayabaNerve> Converters work |
| 02:20:10 | * | gildarts_ joined #nim |
| 02:20:20 | * | gildarts_ is now known as Guest62181 |
| 02:26:13 | * | Guest62181 quit (Ping timeout: 244 seconds) |
| 02:26:49 | * | vosper joined #nim |
| 02:38:44 | * | vosper quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 02:41:31 | * | rayman240 joined #nim |
| 02:43:45 | * | rayman22201 quit (Ping timeout: 248 seconds) |
| 02:44:57 | FromDiscord_ | <emekoi> sorry i meant converter concepts (concepts with a return in them) |
| 02:53:35 | * | chamar quit (Read error: Connection reset by peer) |
| 02:58:32 | * | Stummi26 joined #nim |
| 02:58:43 | * | Stummi26 quit (Remote host closed the connection) |
| 03:13:01 | * | endragor joined #nim |
| 03:18:19 | * | BitPuffin quit (Remote host closed the connection) |
| 03:30:50 | * | Cthalupa quit (Ping timeout: 268 seconds) |
| 03:33:25 | * | ipv629 joined #nim |
| 03:35:08 | * | ipv629 quit (Remote host closed the connection) |
| 03:37:29 | * | govg joined #nim |
| 03:38:13 | * | govg quit (Killed (Sigyn (Spam is off topic on freenode.))) |
| 03:39:31 | * | vosper joined #nim |
| 03:59:58 | FromGitter | <Quelklef> Any convention for naming helper functions? |
| 04:05:30 | * | atk21 joined #nim |
| 04:05:43 | * | atk21 quit (Killed (Sigyn (Spam is off topic on freenode.))) |
| 04:27:24 | FromGitter | <timotheecour> `FooImpl` => lots of such names in stdlib |
| 04:27:44 | FromGitter | <Quelklef> P sure that's for macros |
| 04:35:45 | * | stefanos82 quit (Quit: Quitting for now...) |
| 04:43:52 | * | leorize quit (Ping timeout: 256 seconds) |
| 04:47:41 | * | vosper quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 05:07:03 | * | vosper joined #nim |
| 05:19:04 | * | nsf joined #nim |
| 05:19:48 | * | miran joined #nim |
| 05:42:49 | * | vosper quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 05:44:41 | * | vosper joined #nim |
| 05:51:45 | Araq | the "Pure" vs "not pure" distinction is about DLLs (shared objects) |
| 05:52:26 | Araq | DLLs not part of the OS that you would need to ship with your binary |
| 05:53:10 | Araq | libc is usually kinda there and is considered part of the OS |
| 05:54:03 | Araq | I don't understand why 'getPosition' is bad today, essentially all IO/Stream procs can throw an exception on error |
| 05:55:21 | * | leorize joined #nim |
| 05:57:33 | Araq | 'getPosition' doesn't return a hint of a position, 'size' returns a size *hint* as maybe the stream has no way of telling its size, it's a stream after all. apples and oranges. |
| 05:59:14 | FromGitter | <data-man> @Araq Any example for stream without a size? |
| 06:00:09 | Araq | a socket? |
| 06:01:32 | Araq | any protocol that uses a sentinel can't tell the message's size before having read the sentinel |
| 06:02:27 | * | vosper quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 06:02:39 | FromGitter | <data-man> Socket has current size of recieved data |
| 06:02:57 | FromGitter | <data-man> (Or can be added) |
| 06:03:49 | Araq | iirc the OS's socket APIs don't offer that feature |
| 06:04:49 | Araq | getPosition # number of bytes we have "consumed" in some sense. |
| 06:05:00 | Araq | size # number of bytes still to consume. |
| 06:05:05 | * | miran quit (Ping timeout: 240 seconds) |
| 06:05:23 | Araq | spot the difference? you might not know 'size', you can always know 'getPosition' |
| 06:05:38 | Araq | one refers to the past, the other to the future. |
| 06:07:07 | FromGitter | <data-man> Anyway we haven't SocketStream as the Stream successor |
| 06:07:36 | * | BurningPrincess3 joined #nim |
| 06:09:09 | * | BurningPrincess3 quit (Remote host closed the connection) |
| 06:12:45 | Araq | well Stream is an interface |
| 06:13:08 | Araq | and it should allow a SocketStream implementation |
| 06:15:05 | FromGitter | <data-man> Sure, and this implementation can returns all what you want. |
| 06:21:20 | * | vosper joined #nim |
| 06:32:01 | * | vosper quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 06:33:04 | Araq | and what can it return for "I don't know" |
| 06:33:22 | Araq | ? |
| 06:34:41 | FromGitter | <data-man> It's impossible, you know everything! :) |
| 06:37:50 | * | casdr23 joined #nim |
| 06:38:51 | * | cspar joined #nim |
| 06:38:59 | * | casdr23 quit (Killed (Sigyn (Spam is off topic on freenode.))) |
| 06:48:35 | FromGitter | <data-man> PicoC also migrated to gitlab: https://gitlab.com/zsaleeba/picoc |
| 06:49:23 | * | Our joined #nim |
| 06:49:35 | * | Our quit (Killed (Sigyn (Spam is off topic on freenode.))) |
| 06:50:16 | leorize | I really don't understand this entire migration trend |
| 06:50:19 | * | dgwana joined #nim |
| 06:50:27 | leorize | if you're afraid that github may kill off the repo |
| 06:50:42 | leorize | the better way is to have muliple copies everywhere |
| 06:51:48 | FromGitter | <data-man> I tried gitlab and I like it more than github. :) |
| 06:52:16 | FromGitter | <gogolxdong> how to format nim code after pasting? |
| 06:52:26 | FromGitter | <gogolxdong> in vim. |
| 06:53:49 | FromGitter | <gogolxdong> or how to paste nim code from windows vscode to vim? |
| 06:54:27 | leorize | if you have a good vim plugin |
| 06:54:28 | leorize | select the block |
| 06:54:34 | leorize | press = |
| 06:54:54 | leorize | if not, Ctrl+V |
| 06:54:59 | FromGitter | <gogolxdong> which nim plugin , using https://github.com/zah/nim.vim |
| 06:55:41 | leorize | I don't know, I made my own :P Try if it's work |
| 06:57:50 | FromGitter | <gogolxdong> right click is the default paste key in xshell, but it produces incremental indentations for each line. |
| 06:57:52 | FromGitter | <mpcjanssen> or if it's formatted in the source do :set paste before pasting |
| 06:59:56 | FromGitter | <gogolxdong> sorry,middle click , and @mpcjanssen works. |
| 07:00:09 | FromGitter | <gogolxdong> thanks. |
| 07:10:14 | * | Vladar joined #nim |
| 07:10:14 | * | leorize quit (Read error: Connection reset by peer) |
| 07:10:50 | * | dgwana quit (Quit: Leaving) |
| 07:14:05 | * | ng0 joined #nim |
| 07:15:35 | * | leorize joined #nim |
| 07:28:34 | * | darkmagic joined #nim |
| 07:28:41 | * | darkmagic quit (Killed (Sigyn (Spam is off topic on freenode.))) |
| 07:34:53 | * | Cthalupa joined #nim |
| 07:40:31 | * | miran joined #nim |
| 07:44:02 | * | PMunch joined #nim |
| 07:48:45 | norok2 | does nim have a sort of "killer app"? |
| 07:49:16 | * | vosper joined #nim |
| 07:49:33 | * | rayman22201 joined #nim |
| 07:50:35 | ldlework | norok2: macros |
| 07:50:41 | leorize | norok2: like an app that kill processes? |
| 07:50:55 | ldlework | leorize: hehe |
| 07:51:01 | norok2 | :'D |
| 07:51:25 | norok2 | an application where it really shines |
| 07:51:41 | norok2 | like I don't know, rust has firefox |
| 07:52:01 | leorize | nimbus :P |
| 07:53:01 | Araq | Reel Valley was written in Nim |
| 07:53:29 | * | rayman240 quit (Ping timeout: 260 seconds) |
| 07:53:34 | FromGitter | <gogolxdong> what's the equivalent of ioctl in winlean.nim? |
| 07:54:54 | leorize | does windows even have ioctl? |
| 07:54:59 | norok2 | @Araq you mean this: https://www.facebook.com/reelvalley/ |
| 07:56:02 | Araq | Windows tends to have more specialized APIs than just "ioctl" |
| 07:56:57 | Araq | norok2: yes |
| 07:57:40 | FromGitter | <gogolxdong> yes, it takes more effort to program on windows, which beyond my scope. |
| 07:58:22 | norok2 | was it done with the JS back-end? |
| 07:58:26 | FromGitter | <gogolxdong> need a windows expert. |
| 07:58:40 | leorize | gogolxdong: or use msys2 :P |
| 07:59:33 | FromGitter | <gogolxdong> what is that? |
| 07:59:37 | leorize | norok2: yes |
| 07:59:47 | leorize | gogolxdong: a better cygwin |
| 08:00:18 | FromGitter | <gogolxdong> I have mingw ,does it help? |
| 08:00:29 | Araq | norok2: it uses emscripten and the JS backend, I think in the end emscripten had better performance, but it worked with both |
| 08:01:30 | * | leorize quit (Read error: Connection reset by peer) |
| 08:03:09 | * | gmpreussner_ quit (Quit: kthxbye) |
| 08:04:00 | * | gmpreussner joined #nim |
| 08:07:15 | FromGitter | <gogolxdong> I'm in msys2, can I write linux or posix code to compile windows executable? |
| 08:10:21 | * | rubdos24 joined #nim |
| 08:10:21 | * | rubdos24 quit (K-Lined) |
| 08:12:45 | FromGitter | <timotheecour> @araq for https://github.com/nim-lang/Nim/pull/8555 (doAssert, assert now print full path), which aspects don’t you like ? |
| 08:13:42 | Araq | the FAKEFILE idea, the fact that it requires yet another rarely used public system.nim symbol |
| 08:14:18 | FromGitter | <timotheecour> `lineInfoToString` or `assertFailMsg` ? |
| 08:14:31 | Araq | all of it |
| 08:14:52 | Araq | the compiler should use absolute paths and testament should strip them away |
| 08:15:05 | FromGitter | <timotheecour> `lineInfoToString ` is not rarely used ; it’s also used in macros. it avoids duplicating code to print (file, line,col) so it’s all in 1 place. |
| 08:16:06 | FromGitter | <timotheecour> this matters because (in subsequent PR) i want to enable user to customize it (eg for ppl using sublimetext, file:line:col is much better) ; so it’s best if it’s refactored in as few places instead of having code duplication |
| 08:17:50 | FromGitter | <timotheecour> what’s bad about FAKEFILE idea? i think it’s cleanest way to handle this: it avoids hacks I’ve seen which have to be updated when test changes (because they hardcode line numbers etc) or that’d need to be updated, say, if test file moves. |
| 08:19:45 | * | rayman22201 quit (Ping timeout: 248 seconds) |
| 08:21:38 | Araq | it's yet another subtle thing, "Why does my doAssert() suddenly produce FAKEFILE? It breaks my tests" |
| 08:22:50 | FromGitter | <timotheecour> well i can improve the name FAKEFILE to something w a very clear name; in any case it’s trivial to search and UNIQUE so anyone can find where it came from |
| 08:23:45 | Araq | every step is reasonably simple and has clear benefits. and in the end you get another unittest.nim |
| 08:24:44 | Araq | or a system.nim file with over 4000 lines, not counting its includes |
| 08:25:48 | * | leorize joined #nim |
| 08:26:58 | FromGitter | <timotheecour> others have complained about `assert/doAssert` not giving enough context on failure; improving this makes sense IMO |
| 08:27:38 | * | abm joined #nim |
| 08:29:42 | * | vosper quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 08:29:55 | * | heinrich5991 quit (Ping timeout: 268 seconds) |
| 08:30:10 | * | Guest94149 joined #nim |
| 08:32:17 | FromGitter | <timotheecour> > or a system.nim file with over 4000 lines, not counting its includes ⏎ ⏎ that’s a consequence of lack of mutial imports within a project ; that limitation may (hopefully) disappear, in which case we’ll just have 1 definition for assert+friends etc. |
| 08:32:43 | * | heinrich5991 joined #nim |
| 08:33:24 | Araq | no, it's not a consequence of a lack of mutual dependencies |
| 08:33:44 | Araq | it's a consequence of "I like to use this without having to write an 'import' " |
| 08:34:30 | Araq | C++ did that better, assert is in its header file that you need to #include |
| 08:35:35 | * | Guest94149 quit (Ping timeout: 240 seconds) |
| 08:37:30 | FromGitter | <timotheecour> I dont’ follow: system.nim does call assert in a number of places, how would that even work without mutual imports and with an assert defined outside of system.nim ? |
| 08:41:10 | Araq | system.nim doesn't call assert much |
| 08:41:20 | Araq | it uses sysAssert and gcAssert |
| 08:44:05 | FromGitter | <timotheecour> and assert.. cf items, mitems etc. and the fact we have several of these variants (assert, sysAssert), is another symptom of that. |
| 08:44:38 | Araq | it's not. |
| 08:44:52 | Araq | -d:useSysAssert is different from --assertions:on |
| 08:47:55 | * | sins- joined #nim |
| 08:48:47 | * | sins- quit (Killed (Unit193 (Spam is not permitted on freenode.))) |
| 08:49:47 | Araq | and different subsystem using different switches is a good practice IME. in a sense we need to make custom asserts easier to write, not assert easier to use |
| 08:50:18 | Araq | (but we can do both, of course) |
| 08:50:21 | * | rayman22201 joined #nim |
| 08:50:30 | FromGitter | <timotheecour> Ok for sysAssert. but general idea is: it’s easier to implement functions when we already have access to other functions to implement them. right now the workaround is to use a lot of magic procs to bootstrap system.nim, but with mutual imports a lot of these would not be necessary and could be in modules |
| 08:53:51 | Araq | I know of few magics that could be implemented without magics. |
| 08:54:53 | FromGitter | <timotheecour> they could but if system.nim needs them, then the only way out of this conundrum is mutual imports. |
| 08:56:10 | FromGitter | <timotheecour> having stuff like seq[T], etc in normal nim files would be a big improvement. Happy to help on mutual import support implementation if/when you’re convinced this can be beneficial |
| 08:57:19 | FromGitter | <timotheecour> (basically i’d rather spend time implementing than coming up with argumentation in favor) |
| 08:59:41 | Araq | well you can follow my araq-devel branch on the new seqs and what it takes to make them library based |
| 09:01:02 | Araq | cyclic deps are the least of my worries |
| 09:04:57 | * | leorize quit (Ping timeout: 240 seconds) |
| 09:10:41 | FromGitter | <timotheecour> lack of cyclic deps is actually made worse by fact we dont have easy way to do cross-file fwd declarations (as done in C, C++) ; it results in link errors unless user is careful to avoid these. This could be an intermediate solution : supporting cross-module fwd decls (and avoiding link errors when compiler knows where these symbols are defined) |
| 09:12:17 | Araq | I don't get linker errors in Nim, I don't know what you're talking about |
| 09:17:03 | miran | `nre` is faster than `re`? (i'm only splitting strings based on some regex pattern) |
| 09:17:32 | Araq | usually it's slower |
| 09:20:34 | * | TheLemonMan joined #nim |
| 09:23:37 | Araq | but again |
| 09:23:56 | Araq | the compiler should use absolute paths and testament should strip them away. Simplest solution that can work. |
| 09:25:40 | TheLemonMan | Araq, given a symbol `x` of type `var T` what type should `T(x)` evaluate to? `var T` or just `T` ? |
| 09:28:00 | Araq | `T` |
| 09:28:52 | Araq | your FAKEFILE doesn't mean "I don't know the filename" it's a distinction between relative and absolute paths afaict |
| 09:29:41 | * | October joined #nim |
| 09:31:33 | FromGitter | <timotheecour> FAKEFILE avoids all the extra complications in having to parse testament output for patterns that look like filenames, and replace them with their base names ; definitely more complex than my trivial FAKEFILE idea |
| 09:33:16 | Araq | that's testament's job though. it already parses the compiler's/program's output. |
| 09:34:05 | Araq | and nimsuggest's tester has the same problem too. nimsuggest must return absolute paths for stupid editors but for testing absolute files suck |
| 09:34:42 | * | October quit (Ping timeout: 256 seconds) |
| 09:34:47 | Araq | push the complexity to where it belongs. |
| 09:35:04 | FromGitter | <timotheecour> Eg: |
| 09:36:01 | FromGitter | <timotheecour> (/Users/timothee/git_clone/nim/Nim/tests/ambsym/tambsym3.nim) |
| 09:37:42 | FromGitter | <timotheecour> it’s a form of mocking, which is commonly usd in testing frameworks. Nothing new in my proposal |
| 09:41:49 | Araq | don't mock me. |
| 09:43:43 | FromGitter | <Bennyelg> Hey, I want to create something like the familliar javascript method ⏎ someList T .forEach(e => echo(e)) |
| 09:43:57 | FromGitter | <Bennyelg> I am struggling :| anyone can advice ? |
| 09:44:06 | miran | `mapIt`? |
| 09:44:26 | FromGitter | <Bennyelg> No, I dont want to use those implementation I want to learn by creating one myself |
| 09:45:02 | miran | see how `mapIt` is implemented and base your own on that? |
| 09:50:48 | FromGitter | <Bennyelg> deduplicate not working for custom types? ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b6d5ff82a8e6c6083a3e34c] |
| 09:51:24 | FromGitter | <timotheecour> @araq in a similar vein, we run tests with hints off instead of trying to parse the output with hints and filter them out. That would be foolish, just as it’s pointless for testament to parse which part of output corresponds to file name and clean these up |
| 09:52:35 | Araq | you just proposed we parse gcc's/clang's error messages and map them to our own |
| 09:53:07 | Araq | and yet testament should parse the compiler's output? even though it already does that and it's in fact a big part of its job? |
| 09:53:16 | Araq | *should not |
| 09:53:21 | * | Taylor_ joined #nim |
| 09:53:35 | FromGitter | <GULPF> @Bennyelg for ref types you need to define `==` yourself, otherwise they will use referential equality |
| 09:53:49 | FromGitter | <Bennyelg> Oh, now it's all clear |
| 09:54:01 | FromGitter | <Bennyelg> fantastish ;] |
| 09:54:37 | FromGitter | <timotheecour> you mean https://github.com/nim-lang/Nim/issues/8602 ? that’s just because there’s no simpler solution in that case (i can close that issue , not a big annoyance) |
| 09:55:01 | * | clyybber joined #nim |
| 09:55:21 | * | PrimHelios quit (Ping timeout: 240 seconds) |
| 09:56:31 | Araq | and testing absolute filenames is valuable anyway, Nim swallows the leading '/' sometimes and that could have been detected by the tests too |
| 09:57:08 | Araq | we can give testament a regex filter before it does its comparisons |
| 09:57:31 | FromGitter | <timotheecour> sure, but that should be dealt with in either integration test or unittest, not a justificaiton for that |
| 09:58:58 | FromGitter | <timotheecour> maybe I’m just missing the motivations behind your thinking of why mocking file:line:col is bad, given that it demonstrably leads to simpler code |
| 09:59:38 | FromGitter | <Bennyelg> @GULPF Work like a charm@ thanks |
| 10:05:10 | Araq | your PR is not simple. |
| 10:07:00 | * | cspar quit (Ping timeout: 256 seconds) |
| 10:07:35 | FromGitter | <timotheecour> all the FAKEFILE logic is there: ⏎ ⏎ ```code paste, see link``` ⏎ ⏎ `+ passing `-d:lineInfoToStringSquashForTesting` to a testament file (which could just be done by default later)` ... [https://gitter.im/nim-lang/Nim?at=5b6d63e7a6af14730b170417] |
| 10:09:50 | TheLemonMan | Araq, enforcing that seems to break quite some code (eg: https://github.com/nim-lang/Nim/blob/devel/lib/system/sysio.nim#L144 and its usage of `.string` to get a `var string`) |
| 10:10:03 | * | thk127 joined #nim |
| 10:10:10 | * | thk127 quit (Killed (Sigyn (Spam is off topic on freenode.))) |
| 10:16:00 | Araq | TheLemonMan: type conversions from distinct or subtype conversions keep the l-value property intact |
| 10:16:28 | Araq | and have nothing to do with 'var T' afaik |
| 10:17:24 | TheLemonMan | oh well, that's why I asked |
| 10:17:41 | TheLemonMan | a `var T` is a l-value |
| 10:18:07 | Araq | yes but a 'var x: string; x' is a l-value that is not a 'var T' |
| 10:18:39 | Araq | l-value is a property of an expression, not of a type |
| 10:19:12 | Araq | I don't want system.lineInfoToString |
| 10:25:22 | Araq | or system.$ for arrays, or system.assert, these should all be moved from system into different modules |
| 10:26:10 | Araq | the existing stuff has the strategic advantage of existing though, but I can fight for that it's not getting worse all the time |
| 10:33:52 | FromGitter | <timotheecour> ok so would you accept PR with these 2 changes: ⏎ ⏎ 1) `lineInfoToString becomes private` ⏎ 2) keep pre-existing code duplication in macros.nim [https://gitter.im/nim-lang/Nim?at=5b6d6a1049932d4fe4c14175] |
| 10:43:27 | * | dddddd joined #nim |
| 10:45:07 | Araq | why is that so important? |
| 10:45:27 | Araq | improving testament seems more valuable |
| 10:49:08 | * | leorize joined #nim |
| 10:57:29 | FromGitter | <timotheecour> given 2 approaches, I’m always in favor of the simplest one. my `FAKEFILE` approach is definitely simpler than adding more logic to testament. |
| 10:58:24 | Araq | it's not simpler. you're adding code to system.nim which affects everybody, adding code to testament affects few. |
| 10:59:14 | FromGitter | <timotheecour> not if I make `lineInfoToString` private, then it doesn’t add complexity to system.nim |
| 10:59:19 | Araq | or maybe it's "simpler" but "simple" is not the only criterion out there |
| 11:00:17 | FromGitter | <tim-st> Araq: here is an example that only fails when the proc is called, that's the example I meant yesterday: |
| 11:00:26 | FromGitter | <tim-st> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b6d7049a3a93d242245e4d2] |
| 11:15:08 | * | Taylor_ quit (Quit: Leaving) |
| 11:34:40 | * | j-fish joined #nim |
| 11:34:52 | * | j-fish is now known as Guest64500 |
| 11:35:35 | * | Guest64500 quit (Killed (Unit193 (Spam is not permitted on freenode.))) |
| 11:36:19 | * | fsamareanu13 joined #nim |
| 11:36:28 | * | fsamareanu13 quit (Killed (Sigyn (Spam is off topic on freenode.))) |
| 11:37:41 | * | f16 joined #nim |
| 11:37:58 | * | f16 quit (Killed (Unit193 (Spam is not permitted on freenode.))) |
| 11:39:22 | * | rockcavera quit (Remote host closed the connection) |
| 12:06:06 | * | graingert8 joined #nim |
| 12:09:13 | * | jdhorwitz quit (Quit: Connection closed for inactivity) |
| 12:11:35 | * | graingert8 quit (Ping timeout: 240 seconds) |
| 12:12:19 | Araq | that's an implicit generic |
| 12:12:36 | Araq | because of the openArray[enum] |
| 12:13:42 | FromGitter | <tim-st> yes, I found out here too: https://github.com/nim-lang/Nim/issues/8603 |
| 12:14:14 | * | noonien joined #nim |
| 12:20:40 | * | francisl joined #nim |
| 12:21:32 | * | PrimHelios joined #nim |
| 12:25:04 | * | MyMind quit (Read error: Connection reset by peer) |
| 12:25:56 | * | MyMind joined #nim |
| 12:41:31 | * | cryptocat1094 joined #nim |
| 12:42:07 | * | PrimHelios quit (Remote host closed the connection) |
| 12:42:35 | * | PrimHelios joined #nim |
| 12:45:41 | * | PrimHelios quit (Remote host closed the connection) |
| 12:47:27 | * | francisl quit (Ping timeout: 240 seconds) |
| 12:49:39 | * | floppydh quit (Quit: WeeChat 2.2) |
| 12:49:50 | * | floppydh joined #nim |
| 12:58:54 | * | francisl joined #nim |
| 13:16:17 | * | ehmry quit (Ping timeout: 248 seconds) |
| 13:17:06 | * | endragor quit (Remote host closed the connection) |
| 13:18:04 | * | ehmry joined #nim |
| 13:24:39 | * | dgwana joined #nim |
| 13:27:21 | * | abm quit (Ping timeout: 240 seconds) |
| 13:35:10 | * | endragor joined #nim |
| 13:39:30 | * | endragor quit (Ping timeout: 256 seconds) |
| 13:46:56 | * | Sveta joined #nim |
| 13:47:15 | * | Sveta quit (Killed (Sigyn (Spam is off topic on freenode.))) |
| 13:56:31 | FromGitter | <kayabaNerve> Is it seqs or tables that must be initialized to a power of 2? |
| 13:57:19 | FromGitter | <kayabaNerve> Tables. Got it. |
| 13:59:45 | * | ng0 quit (Quit: Alexa, when is the end of world?) |
| 14:00:48 | FromGitter | <data-man> What? Why "must be"? |
| 14:03:15 | FromGitter | <kayabaNerve> No idea. I just not they must be initialized SIZE WISE to a power of 2, even if you don't use it. |
| 14:03:16 | FromGitter | <kayabaNerve> https://github.com/nim-lang/Nim/blob/master/lib/pure/collections/tables.nim#L318 |
| 14:03:46 | * | floppydh quit (Quit: WeeChat 2.2) |
| 14:03:47 | FromGitter | <kayabaNerve> Here's the better link: https://nim-lang.org/docs/tables.html#initTable,int |
| 14:03:53 | FromGitter | <tim-st> Tables, but in theory const tables could be replaced by minimal perfect hash functions that would use nearly size of items space |
| 14:04:14 | FromGitter | <tim-st> (internally without the dev to do it) |
| 14:13:36 | FromDiscord_ | <awr> hello |
| 14:14:08 | FromGitter | <data-man> Hello! |
| 14:14:16 | FromDiscord_ | <awr> how is everyone |
| 14:15:48 | FromGitter | <data-man> so-so (personally about me) :-D |
| 14:16:09 | FromDiscord_ | <awr> @tim-st are const tables supported? last time i tried doing that it didn't work but this was probably a year ago |
| 14:17:09 | FromGitter | <tim-st> yes, afaik they are. but they use the space like desribed above. maybe it's a better idea to just use an array[k, v] and hope gcc creates a minimal perfect hash function |
| 14:17:21 | FromGitter | <tim-st> maybe someone can findout if that is done^^ |
| 14:25:48 | * | PMunch quit (Quit: Leaving) |
| 14:29:55 | * | ZetaDot joined #nim |
| 14:31:22 | dom96 | yeah, the main examples all work |
| 14:31:28 | dom96 | You may see some deprecation warnings though |
| 14:33:40 | federico3 | any success story with py2nim? |
| 14:33:56 | miran | dom96: are you talking to someone we don't see? |
| 14:36:15 | * | Jesin quit (Remote host closed the connection) |
| 14:36:23 | dom96 | unregistered users are muted due to spammers |
| 14:36:31 | dom96 | and so I can only see them |
| 14:36:55 | ZetaDot | ah ok, where can I register? |
| 14:37:45 | miran | ZetaDot: now you're visible (if it was you before that we didn't see) |
| 14:38:15 | TheLemonMan | federico3, I'm using nimpy for some numpy/matplotlib/seaborn stuff |
| 14:38:43 | ZetaDot | yes it was me, ok so it seems I am ‘registered' |
| 14:39:01 | miran | TheLemonMan: i would be very interested to see that - i often use numpy and matplotlib |
| 14:39:15 | miran | can you share the code? |
| 14:40:31 | * | vivus joined #nim |
| 14:40:39 | TheLemonMan | miran, not quite, but I just added support for named arguments to it the other day so you can just write everything as it were plain old python (with more discard's) |
| 14:40:43 | * | Jesin joined #nim |
| 14:42:32 | TheLemonMan | miran, https://gist.github.com/LemonBoy/132cf570228b10bb968179b24d073827 a small snippet I wrote the other day |
| 14:43:16 | miran | wow, this is quite elegant/simple! i didn't expect this would work |
| 14:44:22 | miran | i used to `execCmd(python foo.py)` just to run some matplotlib |
| 14:44:41 | miran | this is much nicer and powerful |
| 14:45:30 | TheLemonMan | I have some rough DSL that outputs gnuplot files for my graphing needs |
| 14:45:51 | TheLemonMan | at some point I said "screw everything" and got back to matlab heh |
| 14:46:22 | miran | matlab, ugh |
| 14:47:02 | FromDiscord_ | <awr> is there a way I can check if a proc is defined for a type at compiletime |
| 14:47:59 | * | orliesaurus joined #nim |
| 14:48:08 | * | orliesaurus quit (K-Lined) |
| 14:50:19 | * | dgwana quit (Quit: Leaving) |
| 14:52:16 | * | planetmaker25 joined #nim |
| 14:53:25 | TheLemonMan | awr, use a concept maybe? |
| 14:53:31 | * | planetmaker25 quit (Read error: Connection reset by peer) |
| 14:53:36 | federico3 | TheLemonMan: I'm talking about the py2nim tool |
| 14:53:49 | miran | TheLemonMan: would something like `let (fig, ax) = plt.subplots(); ax.plot(n); fig.tight_layout()` work? |
| 14:58:23 | TheLemonMan | miran, sure |
| 14:58:45 | TheLemonMan | you have to add `.to((PyObject, PyObject))` after the subplots |
| 15:00:34 | FromGitter | <Vindaar> @TheLemonMan @miran: woah, I also had no idea nimpy was this advanced. That's awesome! |
| 15:00:43 | * | ZetaDot quit (Quit: ZetaDot) |
| 15:02:30 | TheLemonMan | that's because the Nim Evangelism Force is too weak |
| 15:06:31 | * | bathtub_shark21 joined #nim |
| 15:07:22 | * | bathtub_shark21 quit (Remote host closed the connection) |
| 15:11:00 | * | nsf quit (Quit: WeeChat 2.2) |
| 15:14:09 | FromGitter | <kayabaNerve> What's the fastest way to cut the last 5 elements of the seq off? sequtils.delete or... |
| 15:14:50 | Araq | setLen |
| 15:15:21 | FromGitter | <kayabaNerve> Thanks |
| 15:16:00 | * | monoxane6 joined #nim |
| 15:16:28 | * | monoxane6 quit (Remote host closed the connection) |
| 15:22:23 | * | michagogo18 joined #nim |
| 15:22:42 | * | michagogo18 quit (K-Lined) |
| 15:25:09 | * | paperManu left #nim (#nim) |
| 15:37:32 | * | Trustable joined #nim |
| 15:41:40 | * | NimBot joined #nim |
| 15:43:35 | * | miran quit (Ping timeout: 240 seconds) |
| 15:44:16 | * | rayman218 joined #nim |
| 15:45:23 | * | rayman22201 quit (Ping timeout: 260 seconds) |
| 15:47:37 | FromGitter | <kayabaNerve> Can someone here who knows what Merkle Trees are (even binary trees, hash the two below it until you're at the top), mind reviewing my code? |
| 15:47:56 | * | ZetaDot joined #nim |
| 15:50:14 | * | DarkArctic joined #nim |
| 16:07:46 | * | Vladar quit (Remote host closed the connection) |
| 16:09:25 | * | TheLemonMan quit (Quit: "It's now safe to turn off your computer.") |
| 16:14:28 | FromGitter | <Varriount> @kayabaNerve Do Merkle trees only grow in Germany? |
| 16:14:55 | FromGitter | <arnetheduck> @dom96, for choosenim, is it possible to activate it only in a particular shell, and not globally? ie it seems that choosenim changes links in `.nimble/bin` which means I can't run two nims side by side in two terminals (easily).. is there maybe a flag for this? notably, this is how python venv's work |
| 16:15:18 | Araq | Kanada is also affected, I think |
| 16:15:58 | dom96 | arnetheduck: AFraid not |
| 16:17:49 | * | yglukhov[i] quit (Remote host closed the connection) |
| 16:21:39 | FromGitter | <arnetheduck> would it be considered as a feature req? worth jotting down in github? |
| 16:22:02 | * | francisl quit (Ping timeout: 255 seconds) |
| 16:22:27 | * | francisl joined #nim |
| 16:26:45 | FromGitter | <kayabaNerve> @Varriount Lol |
| 16:29:18 | FromGitter | <kaushalmodi> @arnetheduck Not sure if the per shell thing can work. |
| 16:29:43 | FromGitter | <kaushalmodi> Though, this might be possible.. what if choosenim creates symlinks for all versions? |
| 16:30:05 | FromGitter | <kaushalmodi> Like `nim_0180`, `nim_devel` |
| 16:30:14 | * | HYP3RBOR3A_ joined #nim |
| 16:30:29 | FromGitter | <kaushalmodi> `nim` will point to whatever user last ran the choosenim to pick. |
| 16:30:55 | * | HYP3RBOR3A_ quit (Remote host closed the connection) |
| 16:31:11 | dom96 | choosenim doesn't create symlinks |
| 16:31:25 | dom96 | so yes, it's perfectly possible |
| 16:31:28 | dom96 | Sure, make a feature request |
| 16:32:26 | FromGitter | <kaushalmodi> @arnetheduck Can you create that request? (Anyways choosenim doesn't work for me on RHEL 6 :P) |
| 16:33:25 | FromGitter | <kaushalmodi> Just thinking out loud here.. would those symlinks work? Would also need to ensure that the right version of stdlib, docs, cfg, etc is picked.. |
| 16:33:28 | federico3 | https://github.com/nim-lang/Nim/issues/8437 |
| 16:33:30 | FromGitter | <arnetheduck> @kaushalmodi of course it can work - instead of modifying a binary in '.nimble/bin', it can update `$PATH` to point to `.nimble/nim-version-XXX/bin`that contains the specific version - or like python, create a folder specific for the environment/project and update path to there... |
| 16:35:00 | FromGitter | <kaushalmodi> I use GNU Stow that symlinks all the files from my versioned Nim installations to the "stowed" dir that's added to PATH |
| 16:35:16 | federico3 | adding translation to Nim docs sounds good. I wonder how to translate the library docs tho |
| 16:35:36 | FromGitter | <kaushalmodi> So when I switch to v0.18.0, all files from that version get used via symlinks |
| 16:35:38 | FromGitter | <arnetheduck> this also neatly takes care of one of the biggest deficiencies of nimble: it's global package registry, where version matching regularly gets messed up |
| 16:37:35 | FromGitter | <kaushalmodi> Also, food for thought.. I don't use Windows, so not sure if the symlinks would work there. |
| 16:42:16 | FromGitter | <arnetheduck> so as dom96 says, there are no symlinks involved.. I used the term casually, but what happens in reality is that choosenim relies on `.nimble/bin` being in your `PATH`.. what it could do is update `PATH` instead of that binary.. |
| 16:43:26 | Araq | 1. symlinks don't work well on windows indeed. |
| 16:43:48 | Araq | 2. --nimbleDir can be project specific, it doesn't have to be a global package registry |
| 16:43:56 | dom96 | arnetheduck: I think you misunderstand how choosenim works too |
| 16:44:05 | dom96 | choosenim creates a proxy executable in ~/.nimble/bin |
| 16:44:13 | dom96 | This executable reads ~/.choosenim/current |
| 16:44:22 | dom96 | and then launches the specified toolchain |
| 16:44:44 | dom96 | In what way does version matching get messed up in Nimble? |
| 16:46:47 | FromGitter | <arnetheduck> @dom96 right.. same end result though |
| 16:52:11 | FromGitter | <arnetheduck> @dom96 it's a global registry - it means that running nimble on project a, then project b will change which packages are used for project a.. example: you're compiling a with stable version 1.0, then for project b, you want to try out `#head`... when you try to compile a again, you'll find that `#head` is matched even though you didn't change anything about a.. |
| 16:53:08 | FromGitter | <arnetheduck> so what versions it matches all depends on what happens to sit in the `pkgs` folder which is global.. it means you cannot `nimble develop` some package comfortably because it breaks all the other projects you're trying to build |
| 16:53:17 | dom96 | Sure, Nimble isn't feature complete yet. The plan is to implement lockfiles to solve this |
| 16:54:43 | FromGitter | <arnetheduck> does that plan include never ever ever depending on global state? |
| 16:55:38 | FromGitter | <arnetheduck> http://cewing.github.io/training.python_web/html/presentations/venv_intro.html points this out in a succinct way: ⏎ For every package ⏎ installed in the ⏎ system Python, the ⏎ gods kill a kitten ... [https://gitter.im/nim-lang/Nim?at=5b6dc38aa3a93d242247f299] |
| 16:57:33 | * | Jesin quit (Remote host closed the connection) |
| 17:00:03 | * | wildlander joined #nim |
| 17:02:21 | * | Hoolootwo joined #nim |
| 17:02:57 | * | Hoolootwo quit (Remote host closed the connection) |
| 17:05:02 | * | Jesin joined #nim |
| 17:07:37 | Araq | as I said, you can use --nimbleDir |
| 17:14:36 | FromGitter | <arnetheduck> and I can use my own shell script to update path to a full nimble+nim installation with aliases that replace nim, nimble etc to add `--nimbleDir` and all the other flags! oh wait, I just implemented choosenim... |
| 17:17:22 | * | vosper joined #nim |
| 17:20:17 | Araq | maybe, but if the script is short enough... |
| 17:22:22 | * | ng0 joined #nim |
| 17:22:58 | Araq | I thought shell scripts are universally regarded as the pinnacle of human achievement. |
| 17:24:21 | FromGitter | <arnetheduck> well, the point of a tool like `choosenim` and `nimble` is to take care of those things conveniently in a way that promotes a good workflow about which you can write tutorials for newcomers.. just like global variables in code, having them depend on global state is.. problematic and causes unnecessary friction when trying to use nim |
| 17:25:15 | * | TBloemink22 joined #nim |
| 17:25:40 | * | TBloemink22 quit (Killed (Sigyn (Spam is off topic on freenode.))) |
| 17:26:19 | FromGitter | <arnetheduck> it's the same story repeated all over again: windows3.11 didn't have isolated memory spaces for apps, python2 didn't have isolated package repos, C overused global vars in basic functions leading to races and reentrancy problems... I guess history is not that great of a teacher ;) |
| 17:34:27 | Araq | sure, actually you don't have to convince me, I wrote 'nawabs' fwiw |
| 17:36:26 | Araq | lol "Is this valid Nim code or a typo?" |
| 17:36:53 | FromGitter | <arnetheduck> https://github.com/dom96/choosenim/issues/78 |
| 17:37:24 | Araq | "Is that art or can I trash it?" |
| 17:37:43 | * | themagician quit () |
| 17:44:17 | * | miran joined #nim |
| 17:44:38 | FromGitter | <dom96> Offtopic: I'm calling it now, this is where UIs are heading https://news.ycombinator.com/item?id=17733515. I've thought about creating a UI toolkit with a minimal implementation of HTML5/CSS3. |
| 17:44:50 | FromGitter | <dom96> Eventually someone will make a good enough version of this so it's used for many desktop apps. |
| 17:56:04 | FromGitter | <zetashift> Flutter works kind of like that no? Instead of C++ it uses dart tho |
| 17:56:15 | FromGitter | <zetashift> and skia as the backend renderer |
| 17:58:54 | * | jgh- joined #nim |
| 18:00:51 | jgh- | hey guys, i'm messing around with macros a bit and i made one like this: `checkSame(x: typed): bool = sameType(getType(float32), x.getType)` i get an error `got <int literal(1)> expected bool`. If I change the return type to `int` I get `got bool expected int` |
| 18:01:39 | jgh- | i found a workaround is to set return type to int, cast to int prior to returning, and do if checkType(v) == 1: instead of if checkType(v): |
| 18:02:30 | jgh- | is there a way to have it emit bool instead of an int? Seems strange that it can't do that. |
| 18:03:05 | Araq | please create an issue for it |
| 18:06:11 | jgh- | ok |
| 18:12:16 | FromGitter | <kungtotte> Dominik: I think we'll see a shared-lib version of Electron before we'll see any real shift in the UI space. Ultimately |
| 18:12:32 | jgh- | https://github.com/nim-lang/Nim/issues/8606 |
| 18:13:30 | * | marduk1917 joined #nim |
| 18:15:16 | * | clyybber quit (Quit: good night) |
| 18:16:20 | * | marduk1917 quit (K-Lined) |
| 18:17:24 | Araq | ty |
| 18:22:34 | * | jgh- quit (Remote host closed the connection) |
| 18:27:17 | * | jgh- joined #nim |
| 18:29:17 | * | jgh- quit (Remote host closed the connection) |
| 18:30:18 | * | themagician joined #nim |
| 18:41:24 | * | yglukhov[i] joined #nim |
| 18:46:59 | * | cspar joined #nim |
| 18:54:40 | * | yglukhov[i] quit (Remote host closed the connection) |
| 19:07:07 | * | Cthalupa quit (Quit: ZNC 1.6.6+deb1ubuntu0.1 - http://znc.in) |
| 19:08:12 | * | nsf joined #nim |
| 19:12:42 | * | Cthalupa joined #nim |
| 19:23:46 | zacharycarter[m] | TheLemonMan: Thank you for working on - https://github.com/nim-lang/Nim/issues/8535 - I didn't have time to take your recommendation in and really study cgthreadvars. I'm noticing though, that threadvar access still isn't correct for OSX |
| 19:24:12 | zacharycarter[m] | I'm still able to produce an error like - `error: use of undeclared identifier 'jobAllocator_pIfSFgeWayCCdkwxetqviA' |
| 19:24:13 | zacharycarter[m] | objectInit((&jobAllocator_pIfSFgeWayCCdkwxetqviA), (&NTI_DmbOJrP8XPh21ejU3vT4EA_));` |
| 19:24:26 | zacharycarter[m] | when jobAllocator is a threadvar |
| 19:25:00 | zacharycarter[m] | although I think the generated C code now does look different |
| 19:25:31 | * | zacharycarter[m] sent a long message: < https://matrix.org/_matrix/media/v1/download/matrix.org/vEQGkRewZQGlSHtgcKshXtKM > |
| 19:25:46 | zacharycarter[m] | I'm not sure if you were done with the code changes or not - if not - is there any way I can assist? |
| 19:29:06 | * | yglukhov[i] joined #nim |
| 19:36:58 | * | rayman218 quit (Ping timeout: 260 seconds) |
| 19:37:26 | * | rayman22201 joined #nim |
| 19:42:20 | * | abm joined #nim |
| 19:56:51 | * | RichiH10 joined #nim |
| 19:57:29 | * | RichiH10 quit (Killed (Sigyn (Spam is off topic on freenode.))) |
| 20:01:58 | * | cryptocat1094 quit (Quit: later) |
| 20:09:29 | * | TheLemonMan joined #nim |
| 20:12:01 | * | DarkArctic quit (Read error: Connection reset by peer) |
| 20:12:29 | zacharycarter[m] | it'd be super nice if that ultralight lib had a C99 API |
| 20:12:47 | TheLemonMan | zacharycarter[m], I still have to have a look at the test failures, if you have more code that's broken please add it to the original ticket |
| 20:17:00 | zacharycarter[m] | TheLemonMan: I'll try to create something reproducable - it's a rather large project that it's failing in atm, so I'll work on it now and see if I can figure out what's causing it |
| 20:18:20 | TheLemonMan | zacharycarter[m], I think that some chunks of the C file where the error is are enough (and the corrisponding nim code ofc :) |
| 20:19:10 | * | noonien quit (Quit: Connection closed for inactivity) |
| 20:19:21 | zacharycarter[m] | okay cool - I will get that going now |
| 20:33:40 | zacharycarter[m] | TheLemonMan: I hope this helps - https://github.com/zacharycarter/frag |
| 20:34:12 | zacharycarter[m] | Please ping me and let me know if I can help to reproduce anything |
| 20:34:15 | zacharycarter[m] | or upload more examples / try out fixes |
| 20:38:13 | TheLemonMan | zacharycarter[m], compiles fine here w/ my patch |
| 20:38:31 | TheLemonMan | I just had to bump the TLS size (yikes) as the compiler kindly suggested |
| 20:38:56 | zacharycarter[m] | hrm |
| 20:39:27 | zacharycarter[m] | weird |
| 20:39:43 | zacharycarter[m] | let me re-install Nim and make sure I have your patch applied correctly |
| 20:45:49 | dom96 | zacharycarter[m]: it'd also be nice if ultralight wasn't proprietary :) |
| 20:45:59 | dom96 | The C++ API is usable from Nim thanks to Nim's awesomness |
| 20:48:03 | zacharycarter[m] | well I know Nim can bind to C++ - I just have never managed to do it successfully / tried hard enough - but I plan on trying this feature out soon |
| 20:48:06 | * | SuchWow21 joined #nim |
| 20:48:18 | zacharycarter[m] | I didn't realize ultralight was proprietary - yes that sucks too |
| 20:48:42 | * | SuchWow21 quit (Remote host closed the connection) |
| 20:49:36 | zacharycarter[m] | TheLemonMan: this is def my fault - my bad, I only copied a portion of your change - my internet is super slow atm but I will confirm asap that your fix also compiles on my system |
| 20:58:05 | * | miran quit (Ping timeout: 244 seconds) |
| 21:04:46 | * | francisl quit (Quit: francisl) |
| 21:04:57 | * | TheSilentLink26 joined #nim |
| 21:05:08 | * | TheSilentLink26 quit (Killed (Unit193 (Spam is not permitted on freenode.))) |
| 21:05:21 | zacharycarter[m] | TheLemonMan: change looks good :thumbsup: |
| 21:05:43 | TheLemonMan | zacharycarter[m], and I just fixed the tests that were red |
| 21:06:32 | zacharycarter[m] | 😂 |
| 21:07:52 | zacharycarter[m] | thanks so much for fixing this! |
| 21:08:28 | TheLemonMan | np, I hate when bugs break your flow :) |
| 21:09:48 | zacharycarter[m] | amen! it's so nice that you're becoming such a Nim compiler wiz! the community just keeps getting better and better! I can't wait to see what's in store for the future of Nim 😀 |
| 21:12:39 | zacharycarter[m] | I'm trying to learn more about compilers work before I take that dive - looking at your fix - there's no way I would have arrived at that |
| 21:19:47 | * | yglukhov[i] quit (Remote host closed the connection) |
| 21:19:57 | * | vosper quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 21:20:59 | * | Sabotender18 joined #nim |
| 21:21:57 | * | Sabotender18 quit (Killed (Sigyn (Spam is off topic on freenode.))) |
| 21:32:58 | * | vosper joined #nim |
| 21:33:07 | * | Trustable quit (Remote host closed the connection) |
| 21:40:42 | FromDiscord_ | <treeform> zacharycarter, I bound to C++ api, the auto wrap tools did not work but manually I made progress. |
| 21:48:42 | * | nsf quit (Quit: WeeChat 2.2) |
| 22:19:51 | * | smt quit (Read error: Connection reset by peer) |
| 22:21:54 | * | rayman282 joined #nim |
| 22:24:23 | * | rayman22201 quit (Ping timeout: 260 seconds) |
| 22:34:18 | * | rayman282 quit (Ping timeout: 260 seconds) |
| 22:37:58 | * | Jesin quit (Remote host closed the connection) |
| 22:43:31 | * | TheLemonMan quit (Quit: "It's now safe to turn off your computer.") |
| 22:48:16 | * | Jesin joined #nim |
| 22:49:25 | * | cottongin29 joined #nim |
| 22:50:41 | * | smt joined #nim |
| 22:53:51 | * | cottongin29 quit (Ping timeout: 240 seconds) |
| 23:17:31 | * | Pisuke joined #nim |
| 23:19:46 | * | MyMind quit (Ping timeout: 268 seconds) |
| 23:21:24 | * | siinus`26 joined #nim |
| 23:21:40 | FromGitter | <kayabaNerve> TIL Nim cross-file inheritance can stop you from accessing members. |
| 23:22:10 | FromGitter | <kayabaNerve> I wanted descendants to have access to the members but not any file, so I had to create getters/setters for the class itself to use... |
| 23:22:57 | * | siinus`26 quit (Remote host closed the connection) |
| 23:47:50 | * | ng0 quit (Quit: Alexa, when is the end of world?) |