5 Jan 2024 |
max397 | then I could revive parts of it | 14:34:58 |
max397 | just use a regex | 14:35:05 |
champignoom | What I'm mentally comparing with is taskwarrior | 14:35:31 |
champignoom | and taskwarrior-tui | 14:35:45 |
max397 | never used any of these | 14:35:53 |
champignoom | Which I believe is sufficiently straightforward to be simulated by the existing neorg infrastructures | 14:36:28 |
champignoom | until the birth of gtd module ofc | 14:37:19 |
max397 | | 14:39:36 |
champignoom | Yeah something like that | 14:40:58 |
champignoom | But ideally compatible with the future gtd spec | 14:41:29 |
max397 | well just not sure if that makes sense to implement in neovim anymore since performance already was a problem with these things when I first implemented them so perhaps you should use the breeze thing for this | 14:41:51 |
max397 | and write the whole api in rust | 14:41:55 |
max397 | which in turn needs norgpolis | 14:42:02 |
boltless | 1. to get any kind of GTD client 2. which needs some backend GTD system to store & sync all the GTD data 3. which needs DB system (called NorgBerg) 4. which needs some stable parser | 14:44:39 |
boltless | …and yeah, I’m still working on parser for literally 5 months | 14:46:08 |
boltless | I totally understand what you are thinking as norg-pandoc’s developer. But as you know, that’s just dirty quick fix and requires tons of work to be able to make something working At least, we are really close to complete the parser right now than before. | 14:49:39 |
boltless | theoretically I think I can complete it in this month if I won’t maintain it never ever again | 14:50:34 |
max397 | meaning it would work but ugly code? | 14:51:02 |
champignoom | well ... | 14:51:05 |
boltless | yeap, it will become another shit like v1 parser | 14:51:27 |
champignoom | you know | 14:51:41 |
max397 | well v1 just isn't up to date with current syntax and rly buggy | 14:51:46 |
max397 | additionally to being ugly code ofc | 14:52:51 |
boltless | I’m talking exactly that. we need spec change right now for some edge cases (like how to hande link injection) and parser does works for most cases, it is still quite buggy as quite a lot of hacks tree-sitter doesn’t expect is going on | 14:53:08 |
boltless | e.g. I’m spending time to avoid parsing stuffs from lexer (which was core reason why v1 is buggy) | 14:54:36 |
boltless | so I can fix the v1, but I think it will still be something hard to maintain and still have same problem from before resolving some known edge cases with known bad design is not a way to go imo. | 14:56:13 |
champignoom | yeah but | 14:57:20 |
champignoom | until something good coming in the next moment | 14:57:31 |
champignoom | we need something working | 14:57:35 |
champignoom | and the next-moment has already been months | 14:57:47 |