Comment on page
useExec
Hook that executes a command and returns the AsyncState corresponding to the execution of the command.
It follows the
stale-while-revalidate
cache invalidation strategy popularized by HTTP RFC 5861. useExec
first returns the data from cache (stale), then executes the command (revalidate), and finally comes with the up-to-date data again.The last value will be kept between command runs.
There are two ways to use the hook.
The first one should be preferred when executing a single file. The file and its arguments don't have to be escaped.
function useExec<T, U>(
file: string,
arguments: string[],
options?: {
shell?: boolean | string;
stripFinalNewline?: boolean;
cwd?: string;
env?: NodeJS.ProcessEnv;
encoding?: BufferEncoding | "buffer";
input?: string | Buffer;
timeout?: number;
parseOutput?: ParseExecOutputHandler<T>;
initialData?: U;
keepPreviousData?: boolean;
execute?: boolean;
onError?: (error: Error) => void;
onData?: (data: T) => void;
onWillExecute?: (args: string[]) -> void;
}
): AsyncState<T> & {
revalidate: () => void;
mutate: MutatePromise<T | U | undefined>;
};
The second one can be used to execute more complex commands. The file and arguments are specified in a single
command
string. For example, useExec('echo', ['Raycast'])
is the same as useExec('echo Raycast')
.If the file or an argument contains spaces, they must be escaped with backslashes. This matters especially if
command
is not a constant but a variable, for example with environment.supportPath
or process.cwd()
. Except for spaces, no escaping/quoting is needed.The
shell
option must be used if the command uses shell-specific features (for example, &&
or ||
), as opposed to being a simple file followed by its arguments.function useExec<T, U>(
command: string,
options?: {
shell?: boolean | string;
stripFinalNewline?: boolean;
cwd?: string;
env?: NodeJS.ProcessEnv;
encoding?: BufferEncoding | "buffer";
input?: string | Buffer;
timeout?: number;
parseOutput?: ParseExecOutputHandler<T>;
initialData?: U;
keepPreviousData?: boolean;
execute?: boolean;
onError?: (error: Error) => void;
onData?: (data: T) => void;
onWillExecute?: (args: string[]) -> void;
}
): AsyncState<T> & {
revalidate: () => void;
mutate: MutatePromise<T | U | undefined>;
};
file
is the path to the file to execute.arguments
is an array of strings to pass as arguments to the file.
or
command
is the string to execute.
With a few options:
options.shell
is a boolean or a string to tell whether to run the command inside of a shell or not. Iftrue
, uses/bin/sh
. A different shell can be specified as a string. The shell should understand the-c
switch.We recommend against using this option since it is:- not cross-platform, encouraging shell-specific syntax.
- slower, because of the additional shell interpretation.
- unsafe, potentially allowing command injection.
options.stripFinalNewline
is a boolean to tell the hook to strip the final newline character from the output. By default, it will.options.cwd
is a string to specify the current working directory of the child process. By default, it will beprocess.cwd()
.options.env
is a key-value pairs to set as the environment of the child process. It will extend automatically fromprocess.env
.options.encoding
is a string to specify the character encoding used to decode thestdout
andstderr
output. If set to"buffer"
, thenstdout
andstderr
will be aBuffer
instead of a string.options.input
is a string or a Buffer to write to thestdin
of the file.options.timeout
is a number. If greater than0
, the parent will send the signalSIGTERM
if the child runs longer than timeout milliseconds. By default, the execution will timeout after 10000ms (eg. 10s).options.parseOutput
is a function that accepts the output of the child process as an argument and returns the data the hooks will return - see ParseExecOutputHandler. By default, the hook will returnstdout
.
options.keepPreviousData
is a boolean to tell the hook to keep the previous results instead of returning the initial value if there aren't any in the cache for the new arguments. This is particularly useful when used for data for a List to avoid flickering. See Argument dependent on user input for more information.
options.initialData
is the initial value of the state if there aren't any in the Cache yet.
options.execute
is a boolean to indicate whether to actually execute the function or not. This is useful for cases where one of the function's arguments depends on something that might not be available right away (for example, depends on some user inputs). Because React requires every hook to be defined on the render, this flag enables you to define the hook right away but wait until you have all the arguments ready to execute the function.options.onError
is a function called when an execution fails. By default, it will log the error and show a generic failure toast with an action to retry.options.onData
is a function called when an execution succeeds.options.onWillExecute
is a function called when an execution will start.
Returns an object with the AsyncState corresponding to the execution of the command as well as a couple of methods to manipulate it.
revalidate
is a method to manually call the function with the same arguments again.mutate
is a method to wrap an asynchronous update and gives some control over how theuseFetch
's data should be updated while the update is going through. By default, the data will be revalidated (eg. the function will be called again) after the update is done. See Mutation and Optimistic Updates for more information.
import { List } from "@raycast/api";
import { useExec } from "@raycast/utils";
import { cpus } from "os";
import { useMemo } from "react";
const brewPath = cpus()[0].model.includes("Apple") ? "/opt/homebrew/bin/brew" : "/usr/local/bin/brew";
export default function Command() {
const { isLoading, data } = useExec(brewPath, ["info", "--json=v2", "--installed"]);
const results = useMemo<{ id: string; name: string }[]>(() => JSON.parse(data || "{}").formulae || [], [data]);
return (
<List isLoading={isLoading}>
{results.map((item) => (
<List.Item key={item.id} title={item.name} />
))}
</List>
);
}
By default, when an argument passed to the hook changes, the function will be executed again and the cache's value for those arguments will be returned immediately. This means that in the case of new arguments that haven't been used yet, the initial data will be returned.
This behaviour can cause some flickering (initial data -> fetched data -> arguments change -> initial data -> fetched data, etc.). To avoid that, we can set
keepPreviousData
to true
and the hook will keep the latest fetched data if the cache is empty for the new arguments (initial data -> fetched data -> arguments change -> fetched data).import { useState } from "react";
import { Detail, ActionPanel, Action } from "@raycast/api";
import { useFetch } from "@raycast/utils";
export default function Command() {
const [searchText, setSearchText] = useState("");
const { isLoading, data } = useExec("brew", ["info", searchText]);
return <Detail isLoading={isLoading} markdown={data} />;
}
When passing a user input to a command, be very careful about using the
shell
option as it could be potentially dangerous.In an optimistic update, the UI behaves as though a change was successfully completed before receiving confirmation from the server that it was - it is being optimistic that it will eventually get the confirmation rather than an error. This allows for a more responsive user experience.
You can specify an
optimisticUpdate
function to mutate the data in order to reflect the change introduced by the asynchronous update.When doing so, you can specify a
rollbackOnError
function to mutate back the data if the asynchronous update fails. If not specified, the data will be automatically rolled back to its previous value (before the optimistic update).import { Detail, ActionPanel, Action, showToast, Toast } from "@raycast/api";
import { useFetch } from "@raycast/utils";
export default function Command() {
const { isLoading, data, revalidate } = useExec("brew", ["info", "--json=v2", "--installed"]);
const results = useMemo<{}[]>(() => JSON.parse(data || "[]"), [data]);
const installFoo = async () => {
const toast = await showToast({ style: Toast.Style.Animated, title: "Installing Foo" });
try {
await mutate(
// we are calling an API to do something
installBrewCask("foo"),
{
// but we are going to do it on our local data right away,
// without waiting for the call to return
optimisticUpdate(data) {
return data?.concat({ name: "foo", id: "foo" });
},
}
);
// yay, the API call worked!
toast.style = Toast.Style.Success;
toast.title = "Foo installed";
} catch (err) {
// oh, the API call didn't work :(
// the data will automatically be rolled back to its previous value
toast.style = Toast.Style.Failure;
toast.title = "Could not install Foo";
toast.message = err.message;
}
};
return (
<List isLoading={isLoading}>
{(data || []).map((item) => (
<List.Item
key={item.id}
title={item.name}
actions={
<ActionPanel>
<Action title="Install Foo" onAction={() => installFoo()} />
</ActionPanel>
}
/>
))}
</List>
);
}
An object corresponding to the execution state of the function.
// Initial State
{
isLoading: true, // or `false` if `options.execute` is `false`
data: undefined,
error: undefined
}
// Success State
{
isLoading: false,
data: T,
error: undefined
}
// Error State
{
isLoading: false,
data: undefined,
error: Error
}
// Reloading State
{
isLoading: true,
data: T | undefined,
error: Error | undefined
}
A method to wrap an asynchronous update and gives some control about how the
useFetch
's data should be updated while the update is going through.export type MutatePromise<T> = (
asyncUpdate?: Promise<any>,
options?: {
optimisticUpdate?: (data: T) => T;
rollbackOnError?: boolean | ((data: T) => T);
shouldRevalidateAfter?: boolean;
}
) => Promise<any>;
A function that accepts the output of the child process as an argument and returns the data the hooks will return.
export type ParseExecOutputHandler<T> = (args: {
/** The output of the process on stdout. */
stdout: string | Buffer; // depends on the encoding option
/** The output of the process on stderr. */
stderr: string | Buffer; // depends on the encoding option
error?: Error | undefined;
/** The numeric exit code of the process that was run. */
exitCode: number | null;
/** The name of the signal that was used to terminate the process. For example, SIGFPE. */
signal: NodeJS.Signals | null;
/** Whether the process timed out. */
timedOut: boolean;
/** The command that was run, for logging purposes. */
command: string;
/** The options passed to the child process, for logging purposes. */
options?: ExecOptions | undefined;
}) => T;
Last modified 29d ago