!HCjHPBLFfoFpYNgwdE:matrix.org

Neovim dev

166 Members
Discussion about Neovim development and adjacent topics13 Servers

Load older messages


SenderMessageTime
2 Mar 2024
@lewis6991:matrix.orglewis6991I think that diagnostic is quite new01:50:00
@clason:matrix.orgclasonluals is like a box of chocolates...08:41:50
@amanda:graven.devAmandaIt's poisonous for your dog?09:03:16
@clason:matrix.orgclasonand bad for your skin09:04:33
@amanda:graven.devAmandaDoes that mean dundar can't use luals? 🤔09:04:58
@bfredl:matrix.orgbfredlye, CATS only09:09:59
@clason:matrix.orgclason
In reply to @mariasolos:matrix.org
lol is my brain broken?
nope, the types are
09:29:19
@clason:matrix.orgclason someone forgor to mark namespace with a ? (again) 09:29:36
@clason:matrix.orgclasonhttps://github.com/neovim/neovim/blob/ee4bbc3af260d166cdeb4634616a978c3d51e33d/runtime/lua/vim/diagnostic.lua#L106509:31:32
@clason:matrix.orgclason(same for severity below)09:32:42
@clason:matrix.orgclasonpeople used to be quite lax about annotating fields as optional, before luals decided to develop a hissy fit about them09:33:10
@clason:matrix.orgclason * (same for severity below) 09:33:43
@clason:matrix.orgclason(and before we exposed these to luals)09:37:55
@clason:matrix.orgclason * (and before we exposed these to luals in your and Lewis' Type Crusade)09:38:16
@lewis6991:matrix.orglewis6991tbf, that one was my fault09:44:23
@lewis6991:matrix.orglewis6991even though the field is in the base class, I added it to the child classes for documentation09:44:42
@lewis6991:matrix.orglewis6991 bfredl: I'd like to do something similar to https://github.com/neovim/neovim/pull/27644 for the API keysets, however since the keysets are in another file it means the documentation would become separate to the function def. Have you got any ideas on how we could do it. Would it be possible to move each keyset near to the function where it is used? 09:49:05
@bfredl:matrix.orgbfredl keysets can really be in any api/*.h file now 09:49:35
@lewis6991:matrix.orglewis6991 Or shall I just ignore the keysets directly and just use @class in the API C files? 09:49:39
@bfredl:matrix.orgbfredl it is just lazyness/convention that they are in api/keysets.h 09:49:53
@lewis6991:matrix.orglewis6991 for example there's a load of docs for nvim_buf_set_extmark in api/extmark.c. Ideally I wouldn't want to move that documentation 09:50:44
@lewis6991:matrix.orglewis6991 * for example there's a load of docs for nvim_buf_set_extmark in api/extmark.c. Ideally I wouldn't want to move that documentation far from where it is now 09:50:54
@bfredl:matrix.orgbfredl it still needs to be in a .h file, but it could be in api/extmark.h , if that helps. 09:51:53
@lewis6991:matrix.orglewis6991hmm ok09:52:10
@lewis6991:matrix.orglewis6991better than nothing I suppose09:52:14
@lewis6991:matrix.orglewis6991 the alternative is to just keep the documentation where it is now, but in a separate comment block using @class and @field 09:52:45
@lewis6991:matrix.orglewis6991and ignore the fact the actual type is defined somewhere else09:53:07
@lewis6991:matrix.orglewis6991 bfredl: one last thing, which order of fields matters most? Order in the keyset, or order in the docstring? In https://github.com/neovim/neovim/pull/27704 i've maintained the order in the docstring 10:13:33
@bfredl:matrix.orgbfredl
In reply to @lewis6991:matrix.org
bfredl: one last thing, which order of fields matters most? Order in the keyset, or order in the docstring? In https://github.com/neovim/neovim/pull/27704 i've maintained the order in the docstring
likely the latter
10:18:33
@bfredl:matrix.orgbfredlkeyset dicts are not "stored" in memory in bulk, so I wouldn't worry about struct packing. better to just make it what's look nice for the user10:19:19

There are no newer messages yet.


Back to Room ListRoom Version: 10