lychee/lychee-lib/src/client.rs

533 lines
18 KiB
Rust
Raw Normal View History

Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
#![allow(
clippy::module_name_repetitions,
clippy::struct_excessive_bools,
clippy::default_trait_access,
clippy::used_underscore_binding
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
)]
2021-12-17 00:32:13 +00:00
use std::{collections::HashSet, time::Duration};
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
use check_if_email_exists::{check_email, CheckEmailInput, Reachable};
use http::{
header::{HeaderMap, HeaderValue},
StatusCode,
};
use hubcaps::{Credentials, Github};
use regex::RegexSet;
use reqwest::header;
use tokio::time::sleep;
use typed_builder::TypedBuilder;
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
use crate::{
filter::{Excludes, Filter, Includes},
quirks::Quirks,
ErrorKind, Request, Response, Result, Status, Uri,
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
};
/// Default lychee user agent
pub const DEFAULT_USER_AGENT: &str = concat!("lychee/", env!("CARGO_PKG_VERSION"));
/// Number of redirects until a request gets declared as failed
pub const DEFAULT_MAX_REDIRECTS: usize = 5;
/// Number of retries until a request gets declared as failed
pub const DEFAULT_MAX_RETRIES: u64 = 3;
/// Wait time in seconds between requests (will be doubled after every failure)
pub const DEFAULT_RETRY_WAIT_TIME: u64 = 1;
/// Total timeout per request until a request gets declared as failed
pub const DEFAULT_TIMEOUT: usize = 20;
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
/// Handles incoming requests and returns responses. Usually you would not
/// initialize a `Client` yourself, but use the `ClientBuilder` because it
/// provides sane defaults for all configuration options.
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
#[derive(Debug, Clone)]
pub struct Client {
/// Underlying reqwest client instance that handles the HTTP requests.
reqwest_client: reqwest::Client,
/// Github client.
github_client: Option<Github>,
/// Filtered domain handling.
filter: Filter,
/// Maximum number of retries
max_retries: u64,
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
/// Default request HTTP method to use.
method: reqwest::Method,
/// The set of accepted HTTP status codes for valid URIs.
accepted: Option<HashSet<StatusCode>>,
2021-09-04 01:21:54 +00:00
/// Require HTTPS URL when it's available.
require_https: bool,
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
/// Override behavior for certain known issues with URIs.
quirks: Quirks,
}
/// A link checker using an API token for Github links
/// otherwise a normal HTTP client.
#[allow(unreachable_pub)]
#[derive(TypedBuilder, Debug)]
#[builder(field_defaults(default, setter(into)))]
pub struct ClientBuilder {
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
/// Set an optional Github token.
/// This allows for more requests before
/// getting rate-limited.
github_token: Option<String>,
/// Check links matching this set of regular expressions
includes: Option<RegexSet>,
/// Exclude links matching this set of regular expressions
excludes: Option<RegexSet>,
/// Exclude all private network addresses
exclude_all_private: bool,
/// Exclude private IP addresses
exclude_private_ips: bool,
/// Exclude link-local IPs
exclude_link_local_ips: bool,
/// Exclude loopback IP addresses (e.g. 127.0.0.1)
exclude_loopback_ips: bool,
/// Don't check mail addresses
exclude_mail: bool,
/// Maximum number of redirects before returning error
#[builder(default = DEFAULT_MAX_REDIRECTS)]
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
max_redirects: usize,
/// Maximum number of retries before returning error
#[builder(default = DEFAULT_MAX_RETRIES)]
max_retries: u64,
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
/// User agent used for checking links
// Faking the user agent is necessary for some websites, unfortunately.
// Otherwise we get a 403 from the firewall (e.g. Sucuri/Cloudproxy on ldra.com).
#[builder(default_code = "String::from(DEFAULT_USER_AGENT)")]
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
user_agent: String,
/// Ignore SSL errors
allow_insecure: bool,
/// Set of allowed URI schemes (e.g. https, http).
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
/// This excludes all links from checking, which
/// don't specify any of these schemes in the URL.
schemes: HashSet<String>,
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
/// Map of headers to send to each resource.
/// This allows working around validation issues
/// on some websites.
custom_headers: HeaderMap,
/// Request method (e.g. `GET` or `HEAD`)
#[builder(default = reqwest::Method::GET)]
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
method: reqwest::Method,
/// Set of accepted return codes / status codes
accepted: Option<HashSet<StatusCode>>,
/// Response timeout per request
timeout: Option<Duration>,
2021-09-06 21:46:31 +00:00
/// Treat HTTP links as errors when HTTPS is available
2021-09-04 01:21:54 +00:00
require_https: bool,
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
}
impl Default for ClientBuilder {
fn default() -> Self {
Self::builder().build()
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
}
}
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
impl ClientBuilder {
fn build_filter(&self) -> Filter {
let includes = self.includes.clone().map(|regex| Includes { regex });
let excludes = self.excludes.clone().map(|regex| Excludes { regex });
let schemes = self.schemes.clone();
Filter {
includes,
excludes,
schemes,
// exclude_all_private option turns on all "private" excludes,
// including private IPs, link-local IPs and loopback IPs
exclude_private_ips: self.exclude_all_private || self.exclude_private_ips,
exclude_link_local_ips: self.exclude_all_private || self.exclude_link_local_ips,
exclude_loopback_ips: self.exclude_all_private || self.exclude_loopback_ips,
exclude_mail: self.exclude_mail,
}
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
}
/// The build method instantiates the client.
///
/// # Errors
///
/// Returns an `Err` if:
/// - The user agent cannot be parsed
/// - The request client cannot be created
/// - The Github client cannot be created
pub fn client(&self) -> Result<Client> {
let mut headers = self.custom_headers.clone();
headers.insert(header::USER_AGENT, HeaderValue::from_str(&self.user_agent)?);
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
headers.insert(
header::TRANSFER_ENCODING,
HeaderValue::from_static("chunked"),
);
let builder = reqwest::ClientBuilder::new()
.gzip(true)
.default_headers(headers)
.danger_accept_invalid_certs(self.allow_insecure)
.redirect(reqwest::redirect::Policy::limited(self.max_redirects));
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
let reqwest_client = (match self.timeout {
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
Some(t) => builder.timeout(t),
None => builder,
})
.build()?;
let github_token = match self.github_token {
Some(ref token) if !token.is_empty() => Some(Github::new(
self.user_agent.clone(),
Credentials::Token(token.clone()),
)?),
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
_ => None,
};
let filter = self.build_filter();
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
let quirks = Quirks::default();
Ok(Client {
reqwest_client,
github_client: github_token,
filter,
max_retries: self.max_retries,
method: self.method.clone(),
accepted: self.accepted.clone(),
2021-09-04 01:21:54 +00:00
require_https: self.require_https,
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
quirks,
})
}
}
impl Client {
/// Check a single request
///
/// # Errors
///
/// This returns an `Err` if
/// - The request cannot be parsed
/// - An HTTP website with an invalid URI format gets checked
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
pub async fn check<T, E>(&self, request: T) -> Result<Response>
where
Request: TryFrom<T, Error = E>,
ErrorKind: From<E>,
{
let Request {
uri,
source,
element: _element,
attribute: _attribute,
} = request.try_into()?;
// TODO: Allow filtering based on element and attribute
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
let status = if self.filter.is_excluded(&uri) {
Status::Excluded
} else if uri.is_file() {
self.check_file(&uri).await
} else if uri.is_mail() {
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
self.check_mail(&uri).await
} else {
2021-09-04 01:21:54 +00:00
match self.check_website(&uri).await {
Status::Ok(code) if self.require_https && uri.scheme() == "http" => {
let mut https_uri = uri.clone();
https_uri
.url
.set_scheme("https")
.map_err(|_| ErrorKind::InvalidURI(uri.clone()))?;
2021-09-04 01:21:54 +00:00
if self.check_website(&https_uri).await.is_success() {
Status::Error(Box::new(ErrorKind::InsecureURL(https_uri)))
} else {
Status::Ok(code)
}
}
s => s,
}
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
};
Ok(Response::new(uri, status, source))
}
/// Check if the given URI is filtered by the client
#[must_use]
pub fn filtered(&self, uri: &Uri) -> bool {
self.filter.is_excluded(uri)
}
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
/// Check a website URI
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
pub async fn check_website(&self, uri: &Uri) -> Status {
let mut retries: u64 = 0;
let mut wait = DEFAULT_RETRY_WAIT_TIME;
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
let mut status = self.check_default(uri).await;
while retries < self.max_retries {
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
if status.is_success() {
return status;
}
sleep(Duration::from_secs(wait)).await;
retries += 1;
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
wait *= 2;
status = self.check_default(uri).await;
}
// Pull out the heavy machinery in case of a failed normal request.
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
// This could be a Github URL and we run into the rate limiter.
if let Some(github_uri) = uri.gh_org_and_repo() {
if let Some(client) = &self.github_client {
match client.repo(github_uri.owner, github_uri.repo).get().await {
Ok(repo) => {
if repo.private {
// Simplify checking private repos by
// assuming the path exists if the main repo exists
// E.g. `github.com/org/private/issues` would exist
// because `github.com/org/private` exists
return Status::Ok(StatusCode::OK);
} else {
if github_uri.endpoint.is_some() {
// The URI returned a non-200 status code in a
// normal request and now we find that the repo
// exists, so the full URI (including the
// additional endpoint) must be invalid.
return Status::Ok(StatusCode::NOT_FOUND);
}
}
}
Err(e) => {
return e.into();
}
}
} else {
return ErrorKind::MissingGitHubToken.into();
}
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
}
status
}
async fn check_default(&self, uri: &Uri) -> Status {
let request = match self
.reqwest_client
.request(self.method.clone(), uri.as_str())
.build()
{
Ok(r) => r,
Err(e) => return e.into(),
};
let request = self.quirks.apply(request);
match self.reqwest_client.execute(request).await {
Ok(ref response) => Status::new(response, self.accepted.clone()),
Err(e) => e.into(),
}
}
/// Check a file URI
pub async fn check_file(&self, uri: &Uri) -> Status {
2021-09-06 21:46:31 +00:00
if let Ok(path) = uri.url.to_file_path() {
if path.exists() {
return Status::Ok(StatusCode::OK);
}
}
ErrorKind::InvalidFilePath(uri.clone()).into()
}
/// Check a mail address
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
pub async fn check_mail(&self, uri: &Uri) -> Status {
let input = CheckEmailInput::new(vec![uri.as_str().to_owned()]);
let result = &(check_email(&input).await)[0];
if let Reachable::Invalid = result.is_reachable {
ErrorKind::UnreachableEmailAddress(uri.clone()).into()
} else {
Status::Ok(StatusCode::OK)
}
}
}
/// A convenience function to check a single URI
/// This is the most simple link check and avoids having to create a client manually.
/// For more complex scenarios, look into using the [`ClientBuilder`] instead.
#[allow(clippy::missing_errors_doc)]
pub async fn check<T, E>(request: T) -> Result<Response>
where
Request: TryFrom<T, Error = E>,
ErrorKind: From<E>,
{
let client = ClientBuilder::builder().build().client()?;
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
Ok(client.check(request).await?)
}
#[cfg(test)]
mod test {
2021-07-04 23:35:36 +00:00
use std::{
fs::File,
time::{Duration, Instant},
};
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
use http::{header::HeaderMap, StatusCode};
use reqwest::header;
2021-07-04 23:35:36 +00:00
use tempfile::tempdir;
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
use super::ClientBuilder;
use crate::{mock_server, test_utils::get_mock_client_response, Uri};
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
#[tokio::test]
async fn test_nonexistent() {
let mock_server = mock_server!(StatusCode::NOT_FOUND);
let res = get_mock_client_response(mock_server.uri()).await;
assert!(res.status().is_failure());
}
#[tokio::test]
async fn test_nonexistent_with_path() {
let res = get_mock_client_response("http://127.0.0.1/invalid").await;
assert!(res.status().is_failure());
}
#[tokio::test]
async fn test_exponential_backoff() {
let mock_server = mock_server!(StatusCode::NOT_FOUND);
let start = Instant::now();
let res = get_mock_client_response(mock_server.uri()).await;
let end = start.elapsed();
assert!(res.status().is_failure());
// on slow connections, this might take a bit longer than nominal backed-off timeout (7 secs)
assert!(end.as_secs() >= 7);
assert!(end.as_secs() <= 8);
}
#[tokio::test]
async fn test_github() {
let res = get_mock_client_response("https://github.com/lycheeverse/lychee").await;
assert!(res.status().is_success());
}
#[tokio::test]
async fn test_github_nonexistent() {
let res = get_mock_client_response("https://github.com/lycheeverse/not-lychee").await;
assert!(res.status().is_failure());
}
#[tokio::test]
async fn test_youtube() {
// This is applying a quirk. See the quirks module.
let res = get_mock_client_response("https://www.youtube.com/watch?v=NlKuICiT470&list=PLbWDhxwM_45mPVToqaIZNbZeIzFchsKKQ&index=7").await;
assert!(res.status().is_success());
let res = get_mock_client_response("https://www.youtube.com/watch?v=invalidNlKuICiT470&list=PLbWDhxwM_45mPVToqaIZNbZeIzFchsKKQ&index=7").await;
assert!(res.status().is_failure());
}
#[tokio::test]
async fn test_non_github() {
let mock_server = mock_server!(StatusCode::OK);
let res = get_mock_client_response(mock_server.uri()).await;
assert!(res.status().is_success());
}
#[tokio::test]
async fn test_invalid_ssl() {
let res = get_mock_client_response("https://expired.badssl.com/").await;
assert!(res.status().is_failure());
// Same, but ignore certificate error
let res = ClientBuilder::builder()
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
.allow_insecure(true)
.build()
.client()
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
.unwrap()
.check("https://expired.badssl.com/")
.await
.unwrap();
assert!(res.status().is_success());
}
2021-07-04 23:35:36 +00:00
#[tokio::test]
async fn test_file() {
let dir = tempdir().unwrap();
let file = dir.path().join("temp");
File::create(file).unwrap();
let uri = format!("file://{}", dir.path().join("temp").to_str().unwrap());
let res = get_mock_client_response(uri).await;
assert!(res.status().is_success());
}
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
#[tokio::test]
async fn test_custom_headers() {
// See https://github.com/rust-lang/crates.io/issues/788
let mut custom = HeaderMap::new();
custom.insert(header::ACCEPT, "text/html".parse().unwrap());
let res = ClientBuilder::builder()
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
.custom_headers(custom)
.build()
.client()
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
.unwrap()
.check("https://crates.io/crates/lychee")
.await
.unwrap();
assert!(res.status().is_success());
}
#[tokio::test]
async fn test_exclude_mail() {
let client = ClientBuilder::builder()
.exclude_mail(false)
.exclude_all_private(true)
.build()
.client()
.unwrap();
assert!(!client.filtered(&Uri {
url: "mailto://mail@example.org".try_into().unwrap()
}));
let client = ClientBuilder::builder()
.exclude_mail(true)
.exclude_all_private(true)
.build()
.client()
.unwrap();
assert!(client.filtered(&Uri {
url: "mailto://mail@example.org".try_into().unwrap()
}));
}
2021-09-04 01:21:54 +00:00
#[tokio::test]
async fn test_require_https() {
let client = ClientBuilder::builder().build().client().unwrap();
let res = client.check("http://example.org").await.unwrap();
assert!(res.status().is_success());
// Same request will fail if HTTPS is required
let client = ClientBuilder::builder()
.require_https(true)
.build()
.client()
.unwrap();
let res = client.check("http://example.org").await.unwrap();
assert!(res.status().is_failure());
}
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
#[tokio::test]
async fn test_timeout() {
// Note: this checks response timeout, not connect timeout.
// To check connect timeout, we'd have to do something more involved,
// see: https://github.com/LukeMathWalker/wiremock-rs/issues/19
let mock_delay = Duration::from_millis(20);
let checker_timeout = Duration::from_millis(10);
assert!(mock_delay > checker_timeout);
let mock_server = mock_server!(StatusCode::OK, set_delay(mock_delay));
let client = ClientBuilder::builder()
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
.timeout(checker_timeout)
.build()
.client()
Major refactor of codebase (#208) - The binary component and library component are separated as two packages in the same workspace. - `lychee` is the binary component, in `lychee-bin/*`. - `lychee-lib` is the library component, in `lychee-lib/*`. - Users can now install only the `lychee-lib`, instead of both components, that would require fewer dependencies and faster compilation. - Dependencies for each component are adjusted and updated. E.g., no CLI dependencies for `lychee-lib`. - CLI tests are only moved to `lychee`, as it has nothing to do with the library component. - `Status::Error` is refactored to contain dedicated error enum, `ErrorKind`. - The motivation is to delay the formatting of errors to strings. Note that `e.to_string()` is not necessarily cheap (though trivial in many cases). The formatting is no delayed until the error is needed to be displayed to users. So in some cases, if the error is never used, it means that it won't be formatted at all. - Replaced `regex` based matching with one of the following: - Simple string equality test in the case of 'false positivie'. - URL parsing based test, in the case of extracting repository and user name for GitHub links. - Either cases would be much more efficient than `regex` based matching. First, there's no need to construct a state machine for regex. Second, URL is already verified and parsed on its creation, and extracting its components is fairly cheap. Also, this removes the dependency on `lazy-static` in `lychee-lib`. - `types` module now has a sub-directory, and its components are now separated into their own modules (in that sub-directory). - `lychee-lib::test_utils` module is only compiled for tests. - `wiremock` is moved to `dev-dependency` as it's only needed for `test` modules. - Dependencies are listed in alphabetical order. - Imports are organized in the following fashion: - Imports from `std` - Imports from 3rd-party crates, and `lychee-lib`. - Imports from `crate::*` or `super::*`. - No glob import. - I followed suggestion from `cargo clippy`, with `clippy::all` and `clippy:pedantic`. Co-authored-by: Lucius Hu <lebensterben@users.noreply.github.com>
2021-04-14 23:24:11 +00:00
.unwrap();
let res = client.check(mock_server.uri()).await.unwrap();
assert!(res.status().is_timeout());
}
}