1,521 karma · joined August 13, 2015
Userspace Linux is different in that not only are these libraries typically installed separately, they all have varying degrees of compatibility between versions.
It will track all uses of reflection during a program's execution. You can then use the resultant config as the starting point for the reflect-config.json.
2. Being able to patch dependencies is important. This isn't so much a feature to help developers, but one to help users. You could technically do this without dynamic linking, but have to make it easy to recompile a project with an updated dependency (which most package managers don't). Not being able to patch a package's dependencies puts the developer accountable for your system administration issues, which is not scalable.
I don't think the GP meant "write an image renderer" when they said "get a very simple barebones HTML renderer". Most languages have libraries for encoding/decoding images in various formats, so aside from how it would fit into document layout, this is a non-issue.
That being said, the Flutter widget model is remarkably simple to understand. And rules like "Constraints go down. Sizes go up. Parent sets position." make it pretty easy to reason about the UI. State management is a bit of a hurdle, since there are so many options, each with its own advantages and disadvantages.
All in all, this was a great choice for building something that works on both mobile and desktop, and has a fast startup time.
The top example there takes a GraphViz description and passes it into the dot command, placing the resultant SVG into the HTML output.
They are. OP has been spending 30 years not being paid to use the tools he wants.
> but that applies to every language barring joke ones
Some languages are far more effective for a domain than others. To stretch it further than Haskell, there's a reason why Verilog is used in hardware over something like C. It would be enormously frustrating to work against my tool to get something done. Programmers who focus on "getting stuff done" in these situations are succeeding despite their tools, which is commendable, but we should still push to using better tools.
> interested in the tool through which the problem is solved not the domain itself
Many factors play into why someone works at a company, whether it's cost-effective to use a certain tool, whether a tool is trendy, whether the candidate is passionate about the domain, or whether the candidate is passionate about the tooling. A lot of people aren't passionate about the domain of adtech, for example, unlike self-driving cars, but they work in these areas because they like the practice of engineering.
- Using Haskell means you can model your domain more precisely and have higher correctness guarantees.
- Using Rust means you get memory safety while maintaining the speed of C++, where the vast majority of vulnerabilities are memory safety issues.
- Using SQL means you can apply an optimizer on top of what is essentially relational algebra over tables.
In all of these cases, a less effective tool could be used, yes. But "it's just a tool" doesn't capture the fact that tools do solve domain-specific issues, and OP is not wrong to seek an employer who acknowledges the value of Haskell has to add.
But I agree, there's a difference between code that veers towards obfuscation and code that employs techniques someone just hasn't learned yet.
However, zero values throw a wrench into this. The zero value of an interface is nil, so enhancing interfaces would require you to address what happens with an uninitialized variable. One of the current proposals suggests that nil continue as the zero value.
They could introduce a totally different type, like a sealed interface, which doesn't require a zero value, but that distinguishes between different types of interfaces, and I'm not sure how that'll be received.
The NFA simulator is heavily optimized for readability, which means lots of recursion, and fewer special cases outside of small regexps. The compiler also doesn't perform certain optimizations like vectorization and emitting jump tables, which might be useful here.
There isn't a metaprogramming facility to generate equivalent Go code like in re2go: https://re2c.org/manual/manual_go.html. The best we can do is pre-declare the regexps globally as to initialize them once, but we still have to run the interpreter.
Moreover, thus far, a DFA matcher is out of a picture, as discussed here: https://github.com/golang/go/issues/11646.
Another possibility is to use something like GraalVM native-image to compile Typescript's type-checker into a binary. This might offer a speedup over its execution on V8.