#avr Logs

Feb 06 2019

12:26 PM gsi__ is now known as gsi_
04:14 PM rue_bed: I hear the marlin source is a real mess
04:15 PM rue_bed: not tried to navigate it
04:42 PM cehteh: wtf forgotten to initialize a variable after i changed a for() loop to a while() loop but gcc didnt complain even with -Wall -Wextra -Werror
04:47 PM specing_ is now known as specing
05:39 PM cehteh: https://github.com/avr-rust/rust aynone of you played with that yet?
06:07 PM vmt: ..jesus.
06:08 PM cehteh: is dead
06:08 PM vmt: if you truly need move semantics and whatever the fuck else metaprogramming, especially for *AVRS*, you're doing something horribly wrong
06:08 PM vmt: horribly, fucking, wrong.
06:09 PM cehteh: i dont get that
06:10 PM vmt: not just move semantics but the whole hold-my-cock-when-i-piss memory management
06:10 PM vmt: when -- while
06:11 PM vmt: which is the biggest selling point of rust
06:11 PM cehteh: whcih is more an abstract concept
06:11 PM vmt: i guess the division here is that some people like to write code of functionality, while some enjoy writing code about writing code... metaprogramming
06:12 PM cehteh: i mean i havent used rust on avr (or somewhere else) yet
06:12 PM cehteh: but the memory model gives some gurantees for correctness without sacrifying performance, so why not
06:13 PM cehteh: of course when you try to put in each and every feature it'll certainly bloat
06:13 PM vmt: it really brings little to the table if you know what you're doing. and yes, this is an assumption, for sure. but avr programs do not exactly require hugely complex memory management.
06:13 PM cehteh: yeah, i never ever used malloc on avr :D
06:13 PM vmt: and even then, programmers should be held to a standard rather than invent more ways in which they can horribly suck. segfaults aren't just a net negative.
06:14 PM cehteh: dunno if rust is possible without heap allocations
06:14 PM vmt: well, given platforms without mmus don't segfault...
06:14 PM vmt: semantics.
06:14 PM cehteh: but tracking ownership and mutability of data by the compiler is nevertheless a nice feature at least in some cases
06:15 PM vmt: not exactly required in avrs, or anywhere else actually, for that matter
06:16 PM cehteh: 'required' is nothing. theoretically you can add a few buttons and swtiches to an avr and hack raw opcodes into the flash
06:16 PM cehteh: its all about more productivity (or the lack of it :D)
06:17 PM cehteh: polprog posted few days ago: https://i1.jbzdy.pl/contents/2019/02/a40d06bfd484e37ad3c68a55cf123032.jpg
06:17 PM vmt: "productivity" more often than not nowadays (is / possibly always has been) the excuse for poor quality
06:17 PM cehteh: or more complexity :)
06:17 PM vmt: and this boils down to the software industry seriously lacking in quality standards
06:17 PM cehteh: yeah
06:18 PM cehteh: but standards need tools to enforce them
06:18 PM vmt: the prerequirement for learning c++ is being able to recite the design standards book by heart
06:18 PM vmt: sure, but tools do not mean developing new fad languages every six months
06:19 PM cehteh: that was why ada was created ... and rust, imo not entirely wrong, and i just asked out of curiosity if anyone tried that, that was not a suggestion
06:19 PM cehteh: C++ is a horrible language
06:19 PM vmt: also other prereqs for c++ include, but are not limited to: placing a bas-relief of bjrn stringsoup on the mantle and sacrificing raw pointers on it every night
06:20 PM cehteh: for some (undoubtful gains) you need shitloads of more complexity
06:20 PM vmt: fucking "raw pointer"
06:20 PM vmt: rawness intensifies.
06:22 PM cehteh: well sometimes i wish C would be more stringent, less autimatic type coercions, having a real typedef which results in a new type etc
06:22 PM vmt: also, the industry only gets fragmentated more like this.
06:22 PM cehteh: (and real consts, of various explicit flavors)
06:23 PM vmt: just like commie loonixists, for example, bugs on software are often filed into distro-specific bugtrackers rather than to the actual authors themselves
06:23 PM cehteh: and less crap wich the compiler could deduce like __restricted
06:23 PM vmt: hmm. i don't see that being an actual limitation
06:24 PM vmt: use it or don't. it has to be incorporated into your design.
06:25 PM vmt: as well as a typedef which would result in a new type, i hardly see it as a nuisance
06:25 PM vmt: the lack of it, that is
06:28 PM cehteh: well C is as it is anyway
11:34 PM day_ is now known as day