Posts by nothien@mstdn.io
 (DIR) Post #A0QwNZgKKWItAsQGmG by nothien@mstdn.io
       2020-10-22T20:22:08Z
       
       0 likes, 0 repeats
       
       @sir The Rust ecosystem (i.e. Cargo) is a total fuckup, I agree.  But the language itself is pretty good.  I use C(99) for most of my projects, but I often wish I had one or the other Rust language feature which would make my life easier.  Right now, the drawbacks (of Rust) are simply far too much.
       
 (DIR) Post #A0dL5iqn1oTazT1SG8 by nothien@mstdn.io
       2020-10-28T19:51:57Z
       
       0 likes, 0 repeats
       
       @sir I have very few issues with it, mainly syntax (function pointers) and namespacing (some function names get very very long).  Also the preprocessor system sucks (macros much?), and I wish there was a good replacement for it out there that integrated properly with the language.  Also sometimes private fields would be nice, so that I can expose a structure's size but not its fields.  I wish there was also a super simple generics system.
       
 (DIR) Post #A0dL5jU8fVx4xVQsam by nothien@mstdn.io
       2020-10-28T19:58:37Z
       
       0 likes, 0 repeats
       
       @sir What would be wrong with a very simple namespacing system, which used e.g. double underscore as separators, didn't allow double underscores in other names (except for special functions), and then made exactly 0 other changes?
       
 (DIR) Post #A0dLudvZ5j88UUXLVI by nothien@mstdn.io
       2020-10-28T21:07:36Z
       
       0 likes, 0 repeats
       
       @sir Oh yeah, another feature I want is member functions, which should be doable if namespacing is in place.
       
 (DIR) Post #A0dLuqQqbFHdGzGUvg by nothien@mstdn.io
       2020-10-28T21:21:17Z
       
       0 likes, 0 repeats
       
       @sir I would rather write foo.bar(baz) than Foo_bar(foo, baz) or worse my_lib_foo_bar(foo, baz)
       
 (DIR) Post #A0dLv8JM7pzCixtlke by nothien@mstdn.io
       2020-10-28T21:33:35Z
       
       0 likes, 0 repeats
       
       @sir Vanity?  All I'm saying is that one is longer to write and easier to misspell.  There are simple ways to implement member functions properly, and I'm saying that why not use the simpler system that offloads boilerplate to the compiler.  I'm not even engaging you on this arbitrary vanity front.
       
 (DIR) Post #A0dLvIdviCNwlfbLCy by nothien@mstdn.io
       2020-10-28T21:45:07Z
       
       0 likes, 0 repeats
       
       @sir and how often does ABI 1:1 correspondence matter?  Sure, it's a useful feature, but you know the names of the types you're using (I hope), so you can from them derive the ABI name if you need to.  Even without that, the benefits outweigh the costs.
       
 (DIR) Post #A0enRzAe1Rhxqtfp5M by nothien@mstdn.io
       2020-10-29T14:43:11Z
       
       0 likes, 0 repeats
       
       @toast which/what rc
       
 (DIR) Post #A0enmDWeMQPoAHc6U4 by nothien@mstdn.io
       2020-10-29T14:45:39Z
       
       0 likes, 0 repeats
       
       @toast and what more do you want from it?
       
 (DIR) Post #A0eo1gWSph5URPtzpw by nothien@mstdn.io
       2020-10-29T14:49:52Z
       
       0 likes, 0 repeats
       
       @toast sounds like every other crappy shellIs cofesh your replacement project?
       
 (DIR) Post #A0ffwmHDPIHkMcsXdQ by nothien@mstdn.io
       2020-10-29T17:20:33Z
       
       0 likes, 0 repeats
       
       @sir How do you implement printf() without generics or C's horrible variadic argument system?
       
 (DIR) Post #A0ffwmsn9aLKFASYCm by nothien@mstdn.io
       2020-10-29T17:22:46Z
       
       0 likes, 0 repeats
       
       @sir That's not a bad idea, actually.  As a side note, tagged unions should definitely be a first-class feature (something like a more flexible version of Rust's enums).