I admire hobbyists who build their own kernels, but the size of modern systems is intimidating. Even something like a Bluetooth stack is bigger than historic kernels, and tends to require multiple people.
I think it's just the cost of compatibility? Tech collaboration is hard.
miniblog.
Sure, your templating language might be good, but is it so good that the manual includes a warning for addiction? 🙃
I've released difftastic 0.33!
I'm really pleased with the results in almost every case I've tried: it's handling nesting way better. This Scala example has improved substantially over 0.32.
Oh, and some bug fixes and I've added CMake support :)
I've realised that the Haskell 'slogan' "avoid success at all costs" is ambiguous.
It can be parsed as "avoid success, at all costs" or "avoid success-at-all-costs". I'm not always sure which meaning is intended.
Fascinating talk on applying deep learning to detecting cheaters in CS:GO https://www.youtube.com/watch?v=kTiP0zKF9bc
The presenter discusses how they get machine-readable data out of matches, and how they still keep a human in the loop (ML just feeds the human analysis component).
My favourite pull requests are PRs that are obviously correct after you've read the description.
I've been learning some CMake, and it's a weird language. "ON" is used for booleans, it puts code in .txt files, but it still has a lispy macro system.
I suspect the vast majority of people (including me) learn CMake from a large preexisting build. That will bias impressions.
@krinkle@mastodon.technology I don't know of anyone using difftastic by default, including me! (I use it >50% of the time though.)
I toyed with background colours but I didn't find anything that I really liked. Contrast is hard, and depends on the user's theme.
Background colours look bad with syntax highlighting (e.g. red background with blue comment text).
It also looked silly due to ignoring whitespace between symbols.
It's an interesting space and I'm still experimenting :)
Do you think language/API docs websites should support comments?
Pros: Adds context, familiar format, engages community
Cons: Can disincentivise contributions, comments might promote bad practices
I've only seen PHP and clojuredocs offer this.
Diffing comments is really hard in a syntactic diff tool like difftastic. I was already making changed words bold, but it's harder to spot.
I've started experimenting with underlines to make the specific words more prominent. What do you think?
@friend Ooh, that's a brilliant example! More fine grained types definitely allows you to offer a smaller, more relevant set of possibilities in completion.
Today I learnt about `git apply --reject`, which applies as much of a patch as it can, and leaves the remaining conflicting changes in foo.rej files: https://web.archive.org/web/20130706020737/http://devblog.4teamwork.ch/blog/2013/06/05/git-partially-applying-a-patch/
Choosing the right amount of content to put on each docs page is surprisingly hard.
I've seen info/man pages as a single large HTML page and it eventually slows down the browser. A short page with a short scrollbar is less intimidating and you're more likely to read to the end.
Are there examples of IDE services deliberately leveraging elaborate type systems?
I'm struggling to think of what a richer type system enables in the IDE space. The existence of void* in C, interface{} in Go, or a lack of generics, doesn't seem to have limited possibilities.
I've released difftastic 0.32! A small release, but well worth upgrading:
* Difftastic is now smarter at discarding obviously unchanged content, producing much faster diffs in a bunch of cases
* Fixed a crash on Windows when a whole file is added/removed
Smart contract security assessments are surprisingly readable: https://certificate.quantstamp.com/full/nomad
It's clear what's being audited, the issues identified are clearly categorised, and the maintainer responses are shown.
@takloufer@fosstodon.org Wow, I did not know that! Thanks for sharing, this is a very googleable term.
Rust is exploring ideas that allow it to generalise over effects, so you don't need foo and foo_async.
It's an unusual name, and potentially complicates error messages, but it's good problem to examine.
https://blog.rust-lang.org/inside-rust/2022/07/27/keyword-generics.html
I haven't found a nice way of reporting "N/N tasks done" in progress interfaces.
Users don't like seeing "Baking cakes 10/10 (100%)", especially if the last cake is slow. It's weird to them to see 100% when you haven't finished, but mathematically it's logical.
If I created a brand new programming language today, I'd be very tempted to write a bootstrap compiler in JS.
This would make it much easier to provide web playgrounds for people to write small programs and play with the language.
Showing 641-660 of 7,506 posts