Fix typos #1650
Fix typos #1650
Conversation
|
approved (only needed once), to run the CI ... however, unfortunately, the PR "fixes typos" not only in comments and docstrings, but also in member variable names and similar (identifier) : could you remove any changes to identifiers, only keeping comment/docstring changes? |
|
@oberstet Done. |
|
would probably make sense to integrate codepspell in CI as well |
| @@ -404,7 +404,7 @@ def generate_token(char_groups: int, | |||
| :param chars_per_group: Number of characters per character group (or 1 to return a token with no grouping). | |||
| :param chars: Characters to choose from. Default is 27 character subset | |||
| of the ISO basic Latin alphabet (see: ``DEFAULT_TOKEN_CHARS``). | |||
| :param sep: When separating groups in the token, the separater string. | |||
| :param sep: When separating groups in the token, the separated string. | |||
| :param lower_case: If ``True``, generate token in lower-case. | |||
There was a problem hiding this comment.
this is wrong, I think the correct word would be separator (noun)
There was a problem hiding this comment.
as in (I think, but I'm non-native): separator is the single character used to split the input string into resulting individual separated string portions / parts / components**
so the former is the "tool" (verb/subject), while the latter is the resulting "material" (object)
There was a problem hiding this comment.
Yes, "separator" is correct
There was a problem hiding this comment.
while nitpicking, and looking at the type hints we have
autobahn-python/autobahn/util.py
Line 370 in 7bc85b3
is that even "correct"?
I mean sep: Optional[str] = None, allows both strings of length 1, and strings of other lengths, while actually it must be 1 char (if non None) ..
There was a problem hiding this comment.
mmh, yeah, I know that, but that's not a type declaration, but a run-time type check ...
There was a problem hiding this comment.
AFAIK the bolted on Python type system is not robust enough to handle "N-character string" in the type-hinting (only, as per above, via runtime checks).
There was a problem hiding this comment.
seems to be unavailable, at least today (well, and if the AI doesn't lie/hallucinate;)
There was a problem hiding this comment.
fwiw, Rust:
struct SingleChar(String);
impl SingleChar {
pub fn new(value: &str) -> Option<Self> {
if value.chars().count() == 1 {
Some(SingleChar(value.to_string()))
} else {
None // Returns None if the input is not exactly one character
}
}
}
fn check_single_char(input: Option<SingleChar>) {
// No runtime check is necessary; input is already guaranteed to be either
// a valid `SingleChar` or `None`.
}C++:
#include <string>
#include <optional>
#include <stdexcept>
#include <iostream>
class SingleChar {
public:
// Constructor that enforces a single-character constraint
explicit SingleChar(const std::string& s) {
if (s.size() != 1) {
throw std::invalid_argument("String must be exactly one character long.");
}
value = s;
}
// Getter to retrieve the stored character
char get() const { return value[0]; }
private:
std::string value;
};
// Function that accepts either a single character or nothing
void checkSingleChar(const std::optional<SingleChar>& input) {
// No runtime check needed; we rely on the SingleChar constructor
if (input) {
std::cout << "Valid single character: " << input->get() << std::endl;
} else {
std::cout << "No input provided." << std::endl;
}
}
int main() {
try {
auto validChar = SingleChar("a");
checkSingleChar(validChar); // Valid input
auto invalidChar = SingleChar("abc"); // Throws exception
checkSingleChar(invalidChar);
} catch (const std::invalid_argument& e) {
std::cerr << "Error: " << e.what() << std::endl;
}
checkSingleChar(std::nullopt); // No input provided
}untested .. I am too lazy .. it looks "reasonable" though ..
There was a problem hiding this comment.
I think Haskell would just be:
import Data.Char
my_fn :: Char -> Nothing
There was a problem hiding this comment.
yeah, definitely very nice example.
truth is, only looking at the type system, Haskell (and formerly ML) are unmatched in their underpinnings as they are based on typed lambda calculus (https://en.wikipedia.org/wiki/Typed_lambda_calculus)
now, the fact that Haskell not only has a solid type foundation, but also nice terse algebraic surface syntax makes it even more interesting.
on the other hand, if you'd ask me for the "best syntax" ... obviously taste comes into play .. and so on, but there is one specific argument I always was attracted to:
LISP does away with any surface syntax, and it has this powerful macro thing, which sits on top of https://en.wikipedia.org/wiki/Homoiconicity as an afterburner.
well, and then there is the real world. countless pretty good Python packages;) get stuff done. batteries included.
anyways, nice comment - kudos @meejah
oh yes, "some" spellchecker, preferably one in wide use within Sphinx / Python packages in general, I kinda regret not having setup docs with all bells and whistles, and CI test-enforce stuff like correct spelling then if we had done that from day 1, the hurdle to pull that off now is much bigger fwiw, in a new project I am working on, I did that (and more CI etc) from day 1 .. TDD .. or TDCI .. e.g. deciding to add and maintain a glossary has paid of hugely (for myself), also the other reference sections which of course are just "best practice" in advanced projects
|
|
rgd https://github.com/crossbario/autobahn-python/actions/runs/11577829463/job/32239966210?pr=1650 it seems this - no longer - works https://github.com/crossbario/autobahn-python/blame/master/docs/conf.py#L334 rgd configure: error: No modern nasm or yasm found as required. Nasm should be v2.11.01 or later (v2.13 for AVX512) and yasm should be 1.2.0 or later. wtf. why? |
Found via
codespell -L datas,ser,readd,wont,frameword.