clap/src/app/mod.rs

836 lines
30 KiB
Rust

#[allow(dead_code)]
mod settings;
#[macro_use]
mod macros;
mod parser;
mod meta;
pub use self::settings::AppSettings;
use std::env;
use std::io::{self, BufRead, BufWriter, Write};
use std::path::Path;
use std::process;
use std::ffi::OsString;
use std::borrow::Borrow;
#[cfg(feature = "yaml")]
use yaml_rust::Yaml;
use args::{Arg, AnyArg, ArgGroup, ArgMatches, ArgMatcher};
use app::parser::Parser;
use errors::Error;
use errors::Result as ClapResult;
/// Used to create a representation of a command line program and all possible command line
/// arguments.
///
/// Application settings are set using the "builder pattern" with `.get_matches()` being the
/// terminal method that starts the runtime-parsing process and returns information about
/// the user supplied arguments (or lack there of).
///
/// There aren't any mandatory "options" that one must set. The "options" may also appear in any
/// order (so long as `.get_matches()` is the last method called).
///
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// let matches = App::new("myprog")
/// .author("Me, me@mail.com")
/// .version("1.0.2")
/// .about("Explains in brief what the program does")
/// .arg(
/// Arg::with_name("in_file").index(1)
/// )
/// .after_help("Longer explaination to appear after the options when \
/// displaying the help information from --help or -h")
/// .get_matches();
///
/// // Your program logic starts here...
/// ```
#[allow(missing_debug_implementations)]
pub struct App<'a, 'b>(Parser<'a, 'b>) where 'a: 'b;
impl<'a, 'b> App<'a, 'b> {
/// Creates a new instance of an application requiring a name (such as the binary). The name
/// will be displayed to the user when they request to print version or help and usage
/// information. The name should not contain spaces (hyphens '-' are ok).
///
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// let prog = App::new("myprog")
/// # .get_matches();
/// ```
pub fn new<S: Into<String>>(n: S) -> Self { App(Parser::with_name(n.into())) }
/// Creates a new instace of `App` from a .yml (YAML) file. The YAML file must be properly
/// formatted or this function will panic!(). A full example of supported YAML objects can be
/// found in `examples/17_yaml.rs` and `examples/17_yaml.yml`.
///
/// In order to use this function you must compile with the `features = ["yaml"]` in your
/// settings for `[dependencies.clap]` table of your `Cargo.toml`
///
/// Note, due to how the YAML objects are built there is a convienience macro for loading the
/// YAML file (relative to the current file, like modules work). That YAML object can then be
/// passed to this function.
///
/// # Examples
///
/// ```ignore
/// # use clap::App;
/// let yml = load_yaml!("app.yml");
/// let app = App::from_yaml(yml);
/// ```
#[cfg(feature = "yaml")]
pub fn from_yaml<'y>(mut yaml: &'y Yaml) -> App<'y, 'y, 'y, 'y, 'y, 'y> {
// We WANT this to panic on error...so expect() is good.
let mut is_sc = None;
let mut a = if let Some(name) = yaml["name"].as_str() {
App::new(name)
} else {
let yaml_hash = yaml.as_hash().unwrap();
let sc_key = yaml_hash.keys().nth(0).unwrap();
is_sc = Some(yaml_hash.get(sc_key).unwrap());
App::new(sc_key.as_str().unwrap())
};
yaml = if let Some(sc) = is_sc {
sc
} else {
yaml
};
if let Some(v) = yaml["version"].as_str() {
a = a.version(v);
}
if let Some(v) = yaml["author"].as_str() {
a = a.author(v);
}
if let Some(v) = yaml["bin_name"].as_str() {
a = a.bin_name(v);
}
if let Some(v) = yaml["about"].as_str() {
a = a.about(v);
}
if let Some(v) = yaml["after_help"].as_str() {
a = a.after_help(v);
}
if let Some(v) = yaml["usage"].as_str() {
a = a.usage(v);
}
if let Some(v) = yaml["help"].as_str() {
a = a.help(v);
}
if let Some(v) = yaml["help_short"].as_str() {
a = a.help_short(v);
}
if let Some(v) = yaml["version_short"].as_str() {
a = a.version_short(v);
}
if let Some(v) = yaml["settings"].as_vec() {
for ys in v {
if let Some(s) = ys.as_str() {
a = a.setting(s.parse().ok().expect("unknown AppSetting found in YAML file"));
}
}
}
if let Some(v) = yaml["args"].as_vec() {
for arg_yaml in v {
a = a.arg(Arg::from_yaml(&arg_yaml.as_hash().unwrap()));
}
}
if let Some(v) = yaml["subcommands"].as_vec() {
for sc_yaml in v {
a = a.subcommand(SubCommand::from_yaml(&sc_yaml));
}
}
if let Some(v) = yaml["arg_groups"].as_vec() {
for ag_yaml in v {
a = a.group(ArgGroup::from_yaml(&ag_yaml.as_hash().unwrap()));
}
}
a
}
/// Sets a string of author(s) and will be displayed to the user when they request the help
/// information with `--help` or `-h`.
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// App::new("myprog")
/// .author("Me, me@mymain.com")
/// # ;
/// ```
pub fn author<S: Into<&'b str>>(mut self, author: S) -> Self {
self.0.meta.author = Some(author.into());
self
}
/// Overrides the system-determined binary name. This should only be used when absolutely
/// neccessary, such as the binary name for your application is misleading, or perhaps *not*
/// how the user should invoke your program.
///
/// **NOTE:** This command **should not** be used for SubCommands.
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// App::new("myprog")
/// .bin_name("my_binary")
/// # ;
/// ```
pub fn bin_name<S: Into<String>>(mut self, name: S) -> Self {
self.0.meta.bin_name = Some(name.into());
self
}
/// Sets a string briefly describing what the program does and will be displayed when
/// displaying help information.
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// App::new("myprog")
/// .about("Does really amazing things to great people")
/// # ;
/// ```
pub fn about<S: Into<&'b str>>(mut self, about: S) -> Self {
self.0.meta.about = Some(about.into());
self
}
/// Adds additional help information to be displayed in addition to and directly after
/// auto-generated help. This information is displayed **after** the auto-generated help
/// information. This additional help is often used to describe how to use the arguments,
/// or caveats to be noted.
///
/// # Examples
///
/// ```no_run
/// # use clap::App;
/// App::new("myprog")
/// .after_help("Does really amazing things to great people")
/// # ;
/// ```
pub fn after_help<S: Into<&'b str>>(mut self, help: S) -> Self {
self.0.meta.more_help = Some(help.into());
self
}
/// Sets a string of the version number to be displayed when displaying version or help
/// information.
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// App::new("myprog")
/// .version("v0.1.24")
/// # ;
/// ```
pub fn version<S: Into<&'b str>>(mut self, ver: S) -> Self {
self.0.meta.version = Some(ver.into());
self
}
/// Sets a custom usage string to override the auto-generated usage string.
///
/// This will be displayed to the user when errors are found in argument parsing, or when you
/// call `ArgMatcher::usage()`
///
/// **NOTE:** You do not need to specify the "USAGE: \n\t" portion, as that will
/// still be applied by `clap`, you only need to specify the portion starting
/// with the binary name.
///
/// **NOTE:** This will not replace the entire help message, *only* the portion
/// showing the usage.
///
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// App::new("myprog")
/// .usage("myapp [-clDas] <some_file>")
/// # ;
/// ```
pub fn usage<S: Into<&'b str>>(mut self, usage: S) -> Self {
self.0.meta.usage_str = Some(usage.into());
self
}
/// Sets a custom help message and overrides the auto-generated one. This should only be used
/// when the auto-generated message does not suffice.
///
/// This will be displayed to the user when they use the default `--help` or `-h`
///
/// **NOTE:** This replaces the **entire** help message, so nothing will be auto-generated.
///
/// **NOTE:** This **only** replaces the help message for the current command, meaning if you
/// are using subcommands, those help messages will still be auto-generated unless you
/// specify a `.help()` for them as well.
///
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// App::new("myapp")
/// .help("myapp v1.0\n\
/// Does awesome things\n\
/// (C) me@mail.com\n\n\
///
/// USAGE: myapp <opts> <comamnd>\n\n\
///
/// Options:\n\
/// -h, --helpe Dispay this message\n\
/// -V, --version Display version info\n\
/// -s <stuff> Do something with stuff\n\
/// -v Be verbose\n\n\
///
/// Commmands:\n\
/// help Prints this message\n\
/// work Do some work")
/// # ;
/// ```
pub fn help<S: Into<&'b str>>(mut self, help: S) -> Self {
self.0.meta.help_str = Some(help.into());
self
}
/// Sets the short version of the `help` argument without the preceding `-`.
///
/// By default `clap` automatically assigns `h`, but this can be overridden
///
/// **NOTE:** Any leading `-` characters will be stripped, and only the first
/// non `-` chacter will be used as the `short` version
///
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// App::new("myprog")
/// // Using an uppercase `H` instead of the default lowercase `h`
/// .help_short("H")
/// # ;
pub fn help_short<S: AsRef<str> + 'b>(mut self, s: S) -> Self {
self.0.help_short(s.as_ref());
self
}
/// Sets the short version of the `version` argument without the preceding `-`.
///
/// By default `clap` automatically assigns `V`, but this can be overridden
///
/// **NOTE:** Any leading `-` characters will be stripped, and only the first
/// non `-` chacter will be used as the `short` version
///
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// App::new("myprog")
/// // Using a lowercase `v` instead of the default capital `V`
/// .version_short("v")
/// # ;
pub fn version_short<S: AsRef<str>>(mut self, s: S) -> Self {
self.0.version_short(s.as_ref());
self
}
/// Enables Application level settings, passed as argument
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg, AppSettings};
/// App::new("myprog")
/// .setting(AppSettings::SubcommandRequired)
/// .setting(AppSettings::WaitOnError)
/// # ;
/// ```
pub fn setting(mut self, setting: AppSettings) -> Self {
self.0.set(setting);
self
}
/// Enables multiple Application level settings, passed as argument
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg, AppSettings};
/// App::new("myprog")
/// .settings( &[AppSettings::SubcommandRequired,
/// AppSettings::WaitOnError])
/// # ;
/// ```
pub fn settings(mut self, settings: &[AppSettings]) -> Self {
for s in settings {
self.0.set(*s);
}
self
}
/// Adds an argument to the list of valid possibilties manually. This method allows you full
/// control over the arguments settings and options (as well as dynamic generation). It also
/// allows you specify several more advanced configuration options such as relational rules
/// (exclusions and requirements).
///
/// The only disadvantage to this method is that it's more verbose, and arguments must be added
/// one at a time. Using `Arg::from_usage` helps with the verbosity, and still allows full
/// control over the advanced configuration options.
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// App::new("myprog")
/// // Adding a single "flag" argument with a short and help text, using Arg::with_name()
/// .arg(
/// Arg::with_name("debug")
/// .short("d")
/// .help("turns on debugging mode")
/// )
/// // Adding a single "option" argument with a short, a long, and help text using the less
/// // verbose Arg::from_usage()
/// .arg(
/// Arg::from_usage("-c --config=[CONFIG] 'Optionally sets a config file to use'")
/// )
/// # ;
/// ```
pub fn arg<A: Borrow<Arg<'a, 'b>> + 'a>(mut self, a: A) -> Self {
self.0.add_arg(a.borrow());
self
}
/// Adds multiple arguments to the list of valid possibilties by iterating over a Vec of Args
///
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// App::new("myprog")
/// .args(
/// vec![Arg::from_usage("[debug] -d 'turns on debugging info"),
/// Arg::with_name("input").index(1).help("the input file to use")]
/// )
/// # ;
/// ```
pub fn args(mut self, args: &[Arg<'a, 'b>]) -> Self {
for arg in args {
self.0.add_arg(arg);
}
self
}
/// A convienience method for adding a single basic argument (one without advanced
/// relational rules) from a usage type string. The string used follows the same rules and
/// syntax as `Arg::from_usage()`
///
/// The downside to using this method is that you can not set any additional properties of the
/// `Arg` other than what `Arg::from_usage()` supports.
///
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// App::new("myprog")
/// .arg_from_usage("-c --conf=<config> 'Sets a configuration file to use'")
/// # ;
/// ```
pub fn arg_from_usage(mut self, usage: &'a str) -> Self {
self.0.add_arg(&Arg::from_usage(usage));
self
}
/// Adds multiple arguments at once from a usage string, one per line. See `Arg::from_usage()`
/// for details on the syntax and rules supported.
///
/// Like `App::arg_from_usage()` the downside is you only set properties for the `Arg`s which
/// `Arg::from_usage()` supports. But here the benefit is pretty strong, as the readability is
/// greatly enhanced, especially if you don't need any of the more advanced configuration
/// options.
///
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// App::new("myprog")
/// .args_from_usage(
/// "-c --conf=[config] 'Sets a configuration file to use'
/// [debug]... -d 'Sets the debugging level'
/// <input> 'The input file to use'"
/// )
/// # ;
/// ```
pub fn args_from_usage(mut self, usage: &'a str) -> Self {
for l in usage.lines() {
self.0.add_arg(&Arg::from_usage(l.trim()));
}
self
}
/// Adds an ArgGroup to the application. ArgGroups are a family of related arguments. By
/// placing them in a logical group, you make easier requirement and exclusion rules. For
/// instance, you can make an ArgGroup required, this means that one (and *only* one) argument
/// from that group must be present. Using more than one argument from an ArgGroup causes a
/// failure (graceful exit).
///
/// You can also do things such as name an ArgGroup as a confliction, meaning any of the
/// arguments that belong to that group will cause a failure if present.
///
/// Perhaps the most common use of ArgGroups is to require one and *only* one argument to be
/// present out of a given set. For example, lets say that you were building an application
/// where one could set a given version number by supplying a string using an option argument,
/// such as `--set-ver v1.2.3`, you also wanted to support automatically using a previous
/// version numer and simply incrementing one of the three numbers, so you create three flags
/// `--major`, `--minor`, and `--patch`. All of these arguments shouldn't be used at one time
/// but perhaps you want to specify that *at least one* of them is used. You can create a
/// group
///
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, ArgGroup};
/// # App::new("app")
/// .args_from_usage("--set-ver [ver] 'set the version manually'
/// --major 'auto increase major'
/// --minor 'auto increase minor'
/// --patch 'auto increase patch")
/// .group(ArgGroup::with_name("vers")
/// .args(&["ver", "major", "minor","patch"])
/// .required(true))
/// # ;
pub fn group(mut self, group: ArgGroup<'a>) -> Self {
self.0.add_group(group);
self
}
/// Adds a ArgGroups to the application. ArgGroups are a family of related arguments. By
/// placing them in a logical group, you make easier requirement and exclusion rules. For
/// instance, you can make an ArgGroup required, this means that one (and *only* one) argument
/// from that group must be present. Using more than one argument from an ArgGroup causes a
/// failure (graceful exit).
///
/// You can also do things such as name an ArgGroup as a confliction, meaning any of the
/// arguments that belong to that group will cause a failure if present.
///
/// Perhaps the most common use of ArgGroups is to require one and *only* one argument to be
/// present out of a given set. For example, lets say that you were building an application
/// where one could set a given version number by supplying a string using an option argument,
/// such as `--set-ver v1.2.3`, you also wanted to support automatically using a previous
/// version numer and simply incrementing one of the three numbers, so you create three flags
/// `--major`, `--minor`, and `--patch`. All of these arguments shouldn't be used at one time
/// but perhaps you want to specify that *at least one* of them is used. You can create a
/// group
///
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, ArgGroup};
/// # App::new("app")
/// .args_from_usage("--set-ver [ver] 'set the version manually'
/// --major 'auto increase major'
/// --minor 'auto increase minor'
/// --patch 'auto increase patch")
/// .group(ArgGroup::with_name("vers")
/// .args(&["ver", "major", "minor","patch"])
/// .required(true))
/// # ;
pub fn groups(mut self, groups: &[ArgGroup<'a>]) -> Self {
for g in groups {
self = self.group(g.into());
}
self
}
/// Adds a subcommand to the list of valid possibilties. Subcommands are effectively sub apps,
/// because they can contain their own arguments, subcommands, version, usage, etc. They also
/// function just like apps, in that they get their own auto generated help, version, and
/// usage.
///
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg, SubCommand};
/// # App::new("myprog")
/// .subcommand(SubCommand::with_name("config")
/// .about("Controls configuration features")
/// .arg_from_usage("<config> 'Required configuration file to use'"))
/// // Additional subcommand configuration goes here, such as other arguments...
/// # ;
/// ```
pub fn subcommand(mut self, subcmd: App<'a, 'b>) -> Self {
self.0.add_subcommand(subcmd);
self
}
/// Adds multiple subcommands to the list of valid possibilties by iterating over a Vec of
/// `SubCommand`s
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg, SubCommand};
/// # App::new("myprog")
/// .subcommands( vec![
/// SubCommand::with_name("config").about("Controls configuration functionality")
/// .arg(Arg::with_name("config_file").index(1)),
/// SubCommand::with_name("debug").about("Controls debug functionality")])
/// # ;
/// ```
pub fn subcommands(mut self, subcmds: Vec<App<'a, 'b>>) -> Self {
for subcmd in subcmds.into_iter() {
self.0.add_subcommand(subcmd);
}
self
}
/// Prints the full help message to `io::stdout()` using a `BufWriter`
///
/// # Examples
/// ```no_run
/// # use clap::App;
/// # use std::io;
/// let app = App::new("myprog");
/// app.print_help();
/// ```
pub fn print_help(&self) -> ClapResult<()> {
let out = io::stdout();
let mut buf_w = BufWriter::new(out.lock());
self.write_help(&mut buf_w)
}
/// Writes the full help message to the user to a `io::Write` object
///
/// ```no_run
/// # use clap::App;
/// # use std::io;
/// let mut app = App::new("myprog");
/// let mut out = io::stdout();
/// app.write_help(&mut out).ok().expect("failed to write to stdout");
/// ```
pub fn write_help<W: Write>(&self, w: &mut W) -> ClapResult<()> {
self.0.write_help(w)
}
/// Starts the parsing process. Called on top level parent app **ONLY** then recursively calls
/// the real parsing function for all subcommands
///
/// # Panics
///
/// If any arguments contain invalid unicode characters. If this is not desired it is
/// recommended to use the `*_safe()` or `*_lossy()` versions of this method.
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// let matches = App::new("myprog")
/// // Args and options go here...
/// .get_matches();
/// ```
pub fn get_matches(self) -> ArgMatches<'a> {
// Start the parsing
self.get_matches_from(&mut env::args_os())
}
/// Starts the parsing process. Called on top level parent app **ONLY** then recursively calls
/// the real parsing function for all subcommands
///
/// **NOTE:** This method WILL NOT exit when `--help` or `--version` (or short versions) are
/// used. It will return an error, where the `kind` is a `ErrorKind::HelpDisplayed`
/// or `ErrorKind::VersionDisplayed` respectively. You must call `error.exit()` or
/// perform a `std::process::exit` yourself.
///
/// **NOTE:** This method should only be used when is absolutely necessary to handle errors
/// manually.
///
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// let matches = App::new("myprog")
/// // Args and options go here...
/// .get_matches_safe()
/// .unwrap_or_else( |e| { panic!("An error occurs: {}", e) });
/// ```
pub fn get_matches_safe(self) -> ClapResult<ArgMatches<'a>> {
// Start the parsing
self.get_matches_from_safe(&mut env::args_os())
}
/// Starts the parsing process. Called on top level parent app **ONLY** then recursively calls
/// the real parsing function for all subcommands
///
/// **NOTE:** The first argument will be parsed as the binary name.
///
/// **NOTE:** This method should only be used when absolutely necessary, such as needing to
/// parse arguments from something other than `std::env::args()`. If you are unsure, use
/// `App::get_matches()`
///
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// let arg_vec = vec!["my_prog", "some", "args", "to", "parse"];
///
/// let matches = App::new("myprog")
/// // Args and options go here...
/// .get_matches_from(arg_vec);
/// ```
pub fn get_matches_from<I, T>(mut self, itr: I) -> ArgMatches<'a>
where I: IntoIterator<Item = T>,
T: Into<OsString>
{
self.get_matches_from_safe_borrow(itr).unwrap_or_else(|e| {
// Otherwise, write to stderr and exit
self.maybe_wait_for_exit(e);
})
}
/// Starts the parsing process. Called on top level parent app **ONLY** then recursively calls
/// the real parsing function for all subcommands
///
/// **NOTE:** This method WILL NOT exit when `--help` or `--version` (or short versions) are
/// used. It will return an error, where the `kind` is a `ErrorKind::HelpDisplayed`
/// or `ErrorKind::VersionDisplayed` respectively. You must call `error.exit()` or
/// perform a `std::process::exit` yourself.
///
/// **NOTE:** The first argument will be parsed as the binary name.
///
/// **NOTE:** This method should only be used when absolutely necessary, such as needing to
/// parse arguments from something other than `std::env::args()`. If you are unsure, use
/// `App::get_matches_safe()`
///
/// **NOTE:** This method should only be used when is absolutely necessary to handle errors
/// manually.
///
/// **NOTE:** Invalid unicode characters will result in an `Err` with type
/// `ErrorKind::InvalidUtf8`
///
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// let arg_vec = vec!["my_prog", "some", "args", "to", "parse"];
///
/// let matches = App::new("myprog")
/// // Args and options go here...
/// .get_matches_from_safe(arg_vec)
/// .unwrap_or_else( |e| { panic!("An error occurs: {}", e) });
/// ```
pub fn get_matches_from_safe<I, T>(mut self, itr: I) -> ClapResult<ArgMatches<'a>>
where I: IntoIterator<Item = T>,
T: Into<OsString>
{
self.get_matches_from_safe_borrow(itr)
}
/// Starts the parsing process without consuming the `App` struct `self`. This is normally not
/// the desired functionality, instead prefer `App::get_matches_from_safe` which *does*
/// consume `self`.
///
/// **NOTE:** The first argument will be parsed as the binary name.
///
/// **NOTE:** This method should only be used when absolutely necessary, such as needing to
/// parse arguments from something other than `std::env::args()`. If you are unsure, use
/// `App::get_matches_safe()`
///
/// **NOTE:** This method should only be used when is absolutely necessary to handle errors
/// manually.
///
/// **NOTE:** Invalid unicode characters will result in an `Err` with type
/// `ErrorKind::InvalidUtf8`
///
/// # Examples
///
/// ```no_run
/// # use clap::{App, Arg};
/// let arg_vec = vec!["my_prog", "some", "args", "to", "parse"];
///
/// let mut app = App::new("myprog");
/// // Args and options go here...
/// let matches = app.get_matches_from_safe_borrow(arg_vec)
/// .unwrap_or_else( |e| { panic!("An error occurs: {}", e) });
/// ```
pub fn get_matches_from_safe_borrow<I, T>(&mut self, itr: I) -> ClapResult<ArgMatches<'a>>
where I: IntoIterator<Item = T>,
T: Into<OsString>
{
// Verify all positional assertions pass
self.0.verify_positionals();
// If there are global arguments, we need to propgate them down to subcommands
// before parsing incase we run into a subcommand
self.0.propogate_globals();
let mut matcher = ArgMatcher::new();
let mut it = itr.into_iter();
// Get the name of the program (argument 1 of env::args()) and determine the
// actual file
// that was used to execute the program. This is because a program called
// ./target/release/my_prog -a
// will have two arguments, './target/release/my_prog', '-a' but we don't want
// to display
// the full path when displaying help messages and such
if !self.0.is_set(AppSettings::NoBinaryName) {
if let Some(name) = it.next() {
let bn_os = name.into();
let p = Path::new(&*bn_os);
if let Some(f) = p.file_name() {
if let Some(s) = f.to_os_string().to_str() {
if let None = self.0.meta.bin_name {
self.0.meta.bin_name = Some(s.to_owned());
}
}
}
}
}
// do the real parsing
if let Err(e) = self.0.get_matches_with(&mut matcher, &mut it) {
return Err(e);
}
Ok(matcher.into())
}
// Re-implements ClapError::exit except it checks if we should wait for input before exiting
// since ClapError doesn't have that info and the error message must be printed before exiting
fn maybe_wait_for_exit(&self, e: Error) -> ! {
if e.use_stderr() {
wlnerr!("{}", e.message);
if self.0.is_set(AppSettings::WaitOnError) {
wlnerr!("\nPress [ENTER] / [RETURN] to continue...");
let mut s = String::new();
let i = io::stdin();
i.lock().read_line(&mut s).unwrap();
}
process::exit(1);
}
e.exit()
}
}