19:07:13 <richlowe> do any of the modern cscope-y replacements do "find assignments to" but of a type rather than a symbol?  and if so, which?
19:24:53 <richlowe> I feel certain libclang knows, but I have no idea of what if anything is the good consumer :)
19:39:09 <toasterson> some lsp probably
19:39:47 <toasterson> helix-editor uses clangd and lldb for C
19:40:07 <toasterson> so helix editor should work for you
19:45:08 <richlowe> does it actually do the search I want too?
19:46:03 <toasterson> it can do find usages. are you looking for other symbol type searches?
19:46:12 <richlowe> I'm looking for assignments to a specific type
19:46:36 <toasterson> so type conversions?
19:46:56 <rmustacc> Not quite.
19:47:03 <richlowe> no, any assignment to a symbol of a given type.
19:47:26 <toasterson> like when a type is used?
19:48:09 <rmustacc> Just imagine a type called a foo_t. You're looking for all assignments in C where the computed type of the left hand size from the compiler is a foo_t.
19:48:35 <rmustacc> Whether someone has a global, a local, or access it through a struct or union member.
19:49:34 <toasterson> Hmmmm, is that very C specific? Other languages that would be either usage of contructor or when you build the variable of that type. Which IIRC is part of a usage search of the type itself
19:50:31 <richlowe> it's kinda C specific, but clang has to know, so I was hoping something that consumed libclang for these purposes exposed it 
19:50:37 <richlowe> and also that someone had used one and knew which was good.
19:51:23 <toasterson> i think most lsp or very c specific things are hidden deep beneath the surface and the abstracted editor functions
20:02:37 <rmustacc> That's not that C specific. A similar question is often asked in rust, java, etc.
20:13:06 <jlevon> richlowe: uh, maybe coccinelle/spatch?!
20:14:44 <richlowe> possibly, but also aaaargh
20:14:54 <richlowe> jlevon:i-believe-i-did-bob.jpg
20:15:35 <jlevon> I don't think it's a query you can make in LSP at least.
21:42:57 <jclulow> In Rust I think if you pick a type you can at least ask all the places that type is used in your software to define a variable or whatever.  I'm not sure about assignments like that though
21:44:27 <richlowe> Rust's ownership model would make this far easier, because there would be a concrete place the data _is_ and someone who owns that.
21:44:48 <richlowe> if all your pointers are into the heap, you're just going to find "every possible use of this type", which is what I'm trying to avoid, there are lots and lots.
21:45:54 <jclulow> mmm
22:09:00 <richlowe> on the upside I found an excuse to C99-ize something.
22:09:02 <richlowe> which always makes me happy
22:09:15 <richlowe> please forget I said "excuse" when it comes to code review though.
22:19:57 <alanc> clearly you meant to say "a strong and valid need"