sigoden/argc

A Bash CLI framework, also a Bash-based command runner.

argc's Issues

Eats all double-dashes

It's typical for programs to treat an argument of -- as signifying "end-of-arguments" and do not further parsing. Argc sort of follows this, but it removes all -- arguments from the input.

# @cmd
# @arg		command!		Command to run
test() {
	echo "${argc_command[@]}"
}

$ argc foo -- -- -- --
foo
$ echo foo -- -- -- --
foo -- -- -- --

[gdb under argc] ctrl-c leads to terminate, not interrupt

In gdb the default behavior of ctrl-c is to interrupt the process. For example:

  (gdb) c
  Continuing.
  ^C
  Thread 1 "main" received signal SIGINT, Interrupt.
  0x00002aaab428480d in nanosleep () from /lib64/libc.so.6
  (gdb)

But if run gdb from argc, ctrl-c would terminate the gdb debug process.

# @cmd debug
debug () {
  gdb -p `pgrep main` \
     -ex 'c'
  #exec gdb -p `pgrep main` \
  #   -ex 'c'
}

argc debug
 # in gdb 
 # ctrl-c would terminate the debug process.

Its exit code is 130 and the output is like:

  ^C
  
  Thread 1 "main" received signal SIGINT, Interrupt.
  [Switching to Thread 0x2aaaaab30640 (LWP 177491)]
  0x00002aaab5bfa80d in nanosleep () from /lib64/libc.so.6
  (130)

Is there any solution or workaround to not terminate the gdb?

Add completions for fish shell

I saw on README.md that argc has completions for bash/zsh/powershell. Maybe consider adding fish as well?

Dotenv file not read

Hello, thanks for this amazing project which is helping me a lot in standardizing my scripts and simplify how to configure them using flags/args etc.

I'm having some issues with the @env and @meta dotenv annotations.

Probably I'm misunderstanding the way these annotations are meant to be used, but my expectation was that if a .env file was present in the current folder, it was read and used to override the default values from the @env annotation.

For example, in my script I have these annotations

# @env MY_PROFILE=someprofile The profile used
# @meta dotenv

And I've created a .env file in the same folder filled like this

MY_PROFILE=special

When I print the actual env variable I'm getting this

echo $MY_PROFILE => "someprofile" 
# I was expecting instead "special"

Is this the expected behavior? There's a way to override the env using the .env file?

Many thank

Add command to generate boilerplate argcfile.sh

Use argc --argc-create <cmd...> to generate boilerplate file

For example, run argc --argc-create build test run will create argcfile.sh in current directory with following content

#!/usr/bin/env bash

set -e

# @cmd
build() {
    echo "build to be implemented"
}

# @cmd
test() {
    echo "test to be implemented"
}

# @cmd
run() {
    echo "run to be implemented"
}

eval $(argc --argc-eval "$0" "$@")

Multi-line documentation

Hi, I was just looking for a better task runner than my previous choice, and I came across this project. I think it's very simple and elegant and would make a great choice for my scripts. However, one of the main things that I want is to include lots of documentation in the generated help messages, and there presently isn't a way to do that. What I'd like to see is some sort of "long description" field that gets printed in the usage message, applicable to both the top command and subcommands. The precise syntax doesn't really matter, but here is a proposal to consider:

# @description my sweet command
# Extra lines after the @description, which don't start with an @, are 
# treated as the long description. A line which is not a comment ends
# the block.

# @cmd Run subcommand
# Extra comments after the @cmd line, which don't start with an @, are
# treated as the long description. A line which is not a comment ends
# the block.
subcommand() { true; }

This long description would just be put in the command help:

my sweet command

USAGE: test <COMMAND>

COMMANDS:
  subcommand  Run subcommand

Extra lines after the @description, which don't start with an @, are 
treated as the long description. A line which is not a comment ends
the block.

---

Run subcommand

USAGE: test subcommand

Extra comments after the @cmd line, which don't start with an @, are
treated as the long description. A line which is not a comment ends
the block.

Thanks for your consideration!

subcommand supports alias

tag

# @cmd
# @alias b
build() {
  :;
}

# @cmd
# @alias t,tst
test() {
  :;
}

Argc is checking for non-cmd/alias function duplicates

argc is checking all functions for duplicates, but it should only check top-level ones which are tagged with @cmd or @alias.

Reproducer:

#!/bin/sh

# unused function
util_func() {
	case $(basename ${SHELL}) in
		zsh)
			_local_func() {
				echo "do this"
			}
			;;

		bash)
			_local_func() {
				echo "do that instead"
			}
			;;
	esac
}

# @cmd Run tests
test() {
	echo "do something"
}

eval "$(argc "$0" "$@")"

If I run the script, argc exits with the following error:

_local_func(line 13) is conflicted with cmd or alias at line 7

IMHO, _local_func() should not be checked for duplicates, as well as util_func() because none of them have @cmd or @alias tags,

Also, all forms of heredoc should be ignored too. For example, using the following script:

#!/bin/sh

# just a random heredoc
cat >/dev/null <<EOF
func_test() {
	echo test
}
func_test() {
	echo test
}
EOF

# @cmd Run tests
test() {
	echo "do something"
}

eval "$(argc "$0" "$@")"

argc exits with the following error:

func_test(line 8) is conflicted with cmd or alias at line 5

This seems to be a regression as the two above example are working with v0.12.0, but failing with v0.13.0.

Output escaping is broken

At least variable values are not propperly escaped.
e.g. running the demo.sh like this:

export foo="this is broken" ; ./demo.sh download -t '$foo' bar

should result in

cmd:                      download
flag:   --force           
option: --tries           $foo
arg:    source            bar
arg:    target            

but results in

cmd:                      download
flag:   --force           
option: --tries           this is broken
arg:    source            bar
arg:    target 

this is since calling argc -e demo.sh download -t '$foo' bar results in

argc_tries="$foo"
argc_source="bar"
download
``` riight there is an unescaped `$foo`

windows 下执行 ./bin/argc shell:setup powershell 没有效果

已经设置 ARGC_SHELL_PATH 为 scoop 下的 git-bash.exe 路径

Consider exec'ing Argcfile

A small change which would be a bit more hygienic is to exec the Argcfile.sh when it's used.

Consider an Argcfile.sh like this one:

#!/usr/bin/env bash

main() {
    exec bash
}

eval "$(argc --argc-eval "$0" "$@")"

When we run this, it results in a process tree like this one:

$ ps uf
USER         PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
ubuntu      2034  0.1  0.0 247380 17112 pts/1    Ss   10:40   0:06 -fish
ubuntu     36428  0.0  0.0   3304     4 pts/1    Sl   11:53   0:00  \_ argc
ubuntu     36430  1.0  0.0   8664  5436 pts/1    S+   11:53   0:00      \_ bash

The feature request is for argc (PID 36428) to use exec when it runs the Argcfile.sh, instead of starting it as a child process. This is minor but would be slightly cleaner.

Add ability to force command string

As command strings are based on the shell function name, it's not possible to use dashes.

For example, using the following script:

#!/bin/sh

# @describe This is a test

# @cmd      Start the service
start_service() {
    echo do stuff
}

argc $0 "$@"

The help function is:

This is a test

USAGE: argc-test <COMMAND>

COMMANDS:
  start_service  Start the service

I would like to be able to use dashes instead of underscores as word-delimiters in command names:

This is a test

USAGE: argc-test <COMMAND>

COMMANDS:
  start-service  Start the service

I see two potential solutions:

• replace all underscores (_) with dashes (-) in command names (maybe not by default, for example by adding a flag for that)
• or (probably better) add a @name flag to be able to force the command name:

# @cmd      Start the service
# @name     start-service
start_service() {
    echo do stuff
}

I know it's possible to use @alias to make it work, but I would prefer to hide the service_start function name in this output:

This is a test

USAGE: argc-test <COMMAND>

COMMANDS:
  start_service  Start service [aliases: start-service]

Let me know what you think.

And thanks for this great tool ! Can't wait to see the upcoming v1.0.0 released.

EDIT: looks like v1.0.0 was released 30 minutes ago 😄

Maintain default option when internally calling bash function

#!/bin/bash

# @cmd A wrapper for multiple subcommands
wrapper() {
  subcommandA
  # subcommandB
  # subcommandC
}

# @cmd A subcommand
# @option --sub_option=defaultValue
subcommandA() {
  echo "subcommandA ${argc_sub_option}"
}
eval "$(argc --argc-eval "$0" "$@")"

$ ./test subcommandA
subcommandA defaultValue

$ ./test wrapper
subcommandA

If a subcommand "wrapper" consists of multiple other subcommands, the subcommands option default values are not beeing present.
Is this expected behavior?

Better support for "external subcommands"

I do a lot of wrapping commands with argc. Generally, I include an escape hatch to pass arguments directly to the wrapped program, and this works well (post #232 🙏) using argc my-command -- --foo-bar.

A useful addition to argc might be an "external subcommand" feature. Basically, when there is an external subcommand in play, as soon as an unsupported argument is found, all remaining arguments get treated as positional, even if argc would otherwise recognize them. There's a few ways this would work:

• Instead of @arg args+ we use @remaining_args args+ or something (@rest, @remainder, ...). Must be after all @arg, and after this starts to match, it consumes the rest of the input, ignoring further flags and options.
• Instead of @arg args+ we use @arg args... or some other sigil. Same idea.
• We add some @meta external_subcommand or something which applies this rule to the last @arg.

Here's a quick example:

#!/usr/bin/env bash
# @describe Example Argcfile

set -eu

# @cmd Wrapper that works today (wraps terraform)
#
# Example: run with auto-approval
#   argc terraform apply -- -auto-approve
# @arg    subcommand                   Subcommand for terraform
# @arg    args*                        Additional arguments for terraform
# @option -e --environment=production  Name of the environment to operate in
terraform() {
	echo terraform -chdir=terraform "${argc_subcommand:?}" -var-file="${argc_environment:?}.tfvars" ${args_args+"${argc_args[@]}"}
}

# @cmd Wrapper that would be nice (1st idea)
#
# Example: run with auto-approval
#   argc terraform2 apply -auto-approve
# @arg    subcommand                   Subcommand for terraform
# @remaining_args args                 Additional arguments for terraform
# @option -e --environment=production  Name of the environment to operate in
terraform2() {
	echo terraform -chdir=terraform "${argc_subcommand:?}" -var-file="${argc_environment:?}.tfvars" ${args_args+"${argc_args[@]}"}
}

# @cmd Wrapper that would be nice (2nd idea)
#
# Example: run with auto-approval
#   argc terraform3 apply -auto-approve
# @arg    subcommand                   Subcommand for terraform
# @arg    args...                      Additional arguments for terraform
# @option -e --environment=production  Name of the environment to operate in
terraform3() {
	echo terraform -chdir=terraform "${argc_subcommand:?}" -var-file="${argc_environment:?}.tfvars" ${args_args+"${argc_args[@]}"}
}

# @cmd Wrapper that would be nice (3rd idea)
#
# Example: run with auto-approval
#   argc terraform4 apply -auto-approve
# @arg    subcommand                   Subcommand for terraform
# @arg    args*                      Additional arguments for terraform
# @option -e --environment=production  Name of the environment to operate in
# @meta external_subcommand
terraform4() {
	echo terraform -chdir=terraform "${argc_subcommand:?}" -var-file="${argc_environment:?}.tfvars" ${args_args+"${argc_args[@]}"}
}

if ! command -v argc >/dev/null; then
	echo "This command requires argc. Install from https://github.com/sigoden/argc" >&2
	exit 100
fi
eval "$(argc --argc-eval "$0" "$@")"

Nushell Error: Automatically spreading lists is deprecated

With Nushell 0.89.0, the completion script generated by argc --argc-completions nushell causes error:

Error:   × Automatically spreading lists is deprecated
   ╭─[<pathToArgc>\argc-completions\tmp\argc-completions.nu:1:1]
 1 │ def _argc_completer [args: list<string>] {
 2 │     argc --argc-compgen nushell "" $args
   ·                                    ──┬──
   ·                                      ╰── Spreading lists automatically when calling external commands is deprecated and will be removed in 0.91.
 3 │         | split row "\n" | range 0..-2
   ╰────
  help: Use the spread operator (put a '...' before the argument)

Possible solution (breaking change)

Use spread operator (added in Nushell 0.89.0, won't work in previous versions) in ...$args.

argc/src/bin/argc/completions/argc.nu

Lines 1 to 5 in afa5302

	def _argc_completer [args: list<string>] {
	argc --argc-compgen nushell "" $args
	| split row "\n" | range 0..-2
	| each { |line| $line | split column "\t" value description } | flatten
	}

def _argc_completer [args: list<string>] {
    argc --argc-compgen nushell "" ...$args
        | split row "\n" | range 0..-2 
        | each { |line| $line | split column "\t" value description } | flatten 
}

Completions issue with extended documentation

Hello! Thank you for adding extended documentation in #97. I've begun migrating my dotfiles to them and noticed a minor problem with the generated shell completions.

If you run argc --argc-compgen fish on the linked file, you get this output:

create-docker-workspace	Copy the docker workspace root template to the named directory.
traefik	Start an automatic local vhost for Docker containers.
The traefik container will run permanently. The dashboard will be available
at the root domain. All docker containers will be made available at URLs
according to their compose configuration, by default:

    http://SERVICE-PROJECT.lvh.me/

- Docker label reference: https://doc.traefik.io/traefik/providers/docker/
create-argcfile	Create a sample Argcfile.

As you can see, the extended documentation is included in the completion output, which causes each line to appear as a separate (invalid) completion. This also happens for zsh.

Feature request: default subcommand

Hi @sigoden, continuing to love using argc for all of my shell scripts. One pattern that I see coming up sometimes is a desire for a "default subcommand". Basically, I have a script called "backup", and I want it to... perform a backup. But I also want to add subcommands, like "backup verify" or "backup restore". With the current system, I can't do this, because the main command is completely masked by the presence of subcommands.

So the feature request is for the ability to designate a subcommand as a "default" one. Something like this:

#!/usr/bin/env bash

# @cmd
# @default_subcommand
backup() { :; }

# @cmd
restore() { :; }

eval "$(argc --argc-eval "$0" "$@")"

This script could be invoked as argc; argc backup; argc restore. Now, this behavior would cause ambiguity with positional arguments. For example, if the backup subcommand accepted arguments, should argc $PARAM be an error, or invoke argc backup $PARAM? What if $PARAM was restore? So, in order for a subcommand to be eligible as a default subcommand, it cannot accept any positional arguments.

Ideally this feature would also work with nested subcommands in the same way, so e.g. with 3 subcommands backup; backup::incremental; backup::full I could designate backup::incremental as the default for the backup command (which again could be the default subcommand for the overall script).

It would be nice if the shell completion also understood when a command was the default and showed arguments for that subcommand when there wasn't one given.

Thanks again for the consideration. Loving the activity on this project!

How to define sub command specific args, flags and options

How to define separate flags, args, options for each subcommand ?
Not able to figureout that from docs

Thanks

Inconsistent behavior when Argcfile.sh is a symlink

Hi sigoden, I was recently setting up a repository where I needed to symlink the Argcfile into a different location. I wanted to do this specifically because I wanted argc's directory handling (so that the script would always be run from the directory with the symlink).

The behavior I observe:

• argc --argc-script-path prints the path to the symlink
• argc will cd into the directory containing the target of the symlink.

I feel like one of these exhibits the wrong behavior. For me, I'd prefer that argc did not resolve the symlink before changing directory (since I can't imagine a time when that behavior would be useful), but if this is desired behavior, then --argc-script-path should be consistent with it.

Thanks as always for the great software package.

Arg with choices

tag

# @arg mode[dev|prod|pre]    Arg with choices
# @arg mode[=dev|prod|pre]   Arg with choices and default value

Support separate files for subcommand implementations

Hello, love the project! What a clean API and implementation!

Just curious what you'd think about supporting subcommands being implemented in other files. I'm not currently seeing a way to break large CLI projects down into smaller files for organization purposes.

I was thinking something like this could be useful:

# main.sh

# @describe A demo cli

# @cmd Upload a file
# @alias    u
# @arg target!                      File to upload
upload() {
    echo "cmd                       upload"
    echo "arg:  target              $argc_target"
}

# @cmd Download a file
# @cmd-file ./subcommands/download.sh

eval "$(argc --argc-eval "$0" "$@")"

# subcommands/download.sh

# @alias    d
# @flag     -f --force              Override existing file
# @option   -t --tries <NUM>        Set number of retries to NUM
# @arg      source!                 Url to download from
# @arg      target                  Save file to
download() {
    echo "cmd:                      download"
    echo "flag:   --force           $argc_force"
    echo "option: --tries           $argc_tries"
    echo "arg:    source            $argc_source"
    echo "arg:    target            $argc_target"
}

This might just be opening a big can of worms, but I'd love to hear your thoughts!

Filename arguments invalidated after implicit cd

Hey, loving the activity on this project. Thanks for resolving #118 so quickly. I've now ported most of my projects to using argc, and I'm seeing how far I can push it.

It's great that argc automatically cds into the directory of the Argcfile.sh it finds in the parent hierarchy; this is generally the thing that a script author wants. However, if the command accepts a filename argument, the filename is invalid when the script gets run. See this Argcfile for an example:

#!/usr/bin/env bash
# @arg file <FILE>
set -e
eval "$(argc --argc-eval "$0" "$@")"
ls $argc_file

Run this script:

mkdir directory
cd directory
touch test.txt
argc test.txt

The script fails, because the file does not exist, because the directory was changed.

I can think of a few options here:

• Export ARGC_ORIGINAL_PATH or some other variable, which contains the path of the original pwd.
• For <FILE/DIR/PATH> arguments, they are automatically prefixed with the appropriate path.

The first option seems like the best overall to me. The second requires that the argument values are tagged properly, and there could also be use cases where we don't need a specific argument, but we still want to know the original directory of the command.

Thanks for your consideration!

[FR] Colored help

Can we have colored helped, please? I haven't seen anything about color in README or issues.

clap's default color choice is auto, i.e. when terminal supports it:

https://docs.rs/clap/3.2.17/clap/enum.ColorChoice.html#variant.Auto

Test failure for 1.1.0

Hey, AUR maintainer of argc here 👋🏼

I'm getting the following test failure while trying to build the latest version:

Logs

running 10 tests
test compgen::tests::test_get_last_word ... ok
test parser::tests::test_parse_no_long_flag_arg ... ok
test parser::tests::test_parse_no_long_option_arg ... ok
test parser::tests::test_parse_line ... ok
test utils::tests::test_cobol ... ok
test parser::tests::test_parse_with_long_flag_arg ... ok
test parser::tests::test_parse_with_long_flag_arg_single_dash ... ok
test parser::tests::test_parse_positional_arg ... ok
test parser::tests::test_parse_with_long_option_arg ... ok
test parser::tests::test_parse_with_long_option_arg_single_dash ... ok

test result: ok. 10 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s

     Running unittests src/bin/argc/main.rs (target/debug/deps/argc-05cd09e0693dbda7)

running 0 tests

test result: ok. 0 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s

     Running tests/tests.rs (target/debug/deps/integration-343b0242d08b2e78)

running 79 tests
test cli::help ... ok
test cli::version ... ok
test argcfile::argcfile_path ... ok
test fail::arg_miss_default_fn ... ok
test fail::cmd_miss_fn ... ok
test fail::arg_miss_choice_fn ... ok
test fail::conflict_alias ... ok
test fail::conflict_short_flag ... ok
test fail::conflict_cmd ... ok
test fail::conflict_long_flag ... ok
test fail::conflict_positional ... ok
test fail::conflict_long_option ... ok
test fail::conflict_short_option ... ok
test fail::option_miss_choice_fn ... ok
test fail::unsupported_tag ... ok
test fail::option_miss_default_fn ... ok
test fail::unexpected_arg ... ok
test cli::compgen ... ok
test create::create ... ok
test create::create_with_tasks ... ok
test argcfile::argcfile ... ok
test compgen::nested_subcmds ... ok
test compgen::subcmds ... ok
test main_fn::nested_subcmd_no_main ... ok
test main_fn::subcmd_no_main ... ok
test export::case1 ... ok
test spec::undefine_positionals ... ok
test misc::escape ... ok
test param_fn::case1 ... ok
test main_fn::nested_subcmd_main ... ok
test validate::arg_choice ... ok
test validate::arg_choice_fn_pass ... ok
test compgen::multiple ... ok
test spec::arg_no_param ... ok
test validate::arg_missing ... ok
test validate::arg_help_subcmd ... ok
test validate::arg_unknown ... ok
test spec::arg_no_option ... ok
test main_fn::no_main ... ok
test spec::option_help ... ok
test validate::arg_choice_fn_skip ... ok
test validate::option_choice_fn_pass ... ok
test validate::flag_not_multiple ... ok
test param_fn::case3 ... ok
test spec::option_shorts ... ok
test spec::same_option_positional ... ok
test compgen::shorts ... ok
test validate::option_not_multiple ... ok
test validate::option_mismatch_values ... ok
test main_fn::with_main ... ok
test validate::param_missing ... ok
test validate::flag_with_value ... ok
test validate::option_choice ... ok
test param_fn::case2 ... ok
test validate::help_version_shadow ... ok
test compgen::positionals ... ok
test spec::arg_two_multi ... ok
test wrap_help::wrap ... ok
test validate::option_missing ... ok
test validate::option_unknown ... ok
test validate::param_missing_parent ... ok
test validate::option_choice_fn_skip ... ok
test validate::empty_choices ... ok
test spec::dashdash_split ... ok
test main_fn::subcmd_main ... ok
test wrap_help::wrap2 ... ok
test wrap_help::nowrap ... ok
test validate::arg_invalid_subcmd ... ok
test validate::help_version ... ok
test spec::arg_subcmd_help ... ok
stored new snapshot /build/argc/src/argc-1.1.0/tests/snapshots/integration__compgen__choice.snap.new
stored new snapshot /build/argc/src/argc-1.1.0/tests/snapshots/integration__validate__arg_choice_fn.snap.new
stored new snapshot /build/argc/src/argc-1.1.0/tests/snapshots/integration__validate__arg_choice_multi.snap.new
test validate::arg_choice_fn ... FAILED
test compgen::choice ... FAILED
test validate::arg_choice_multi ... FAILED
stored new snapshot /build/argc/src/argc-1.1.0/tests/snapshots/integration__validate__option_choice_fn.snap.new
test validate::option_choice_fn ... FAILED
stored new snapshot /build/argc/src/argc-1.1.0/tests/snapshots/integration__spec__option_eval.snap.new
test spec::option_eval ... FAILED
stored new snapshot /build/argc/src/argc-1.1.0/tests/snapshots/integration__validate__choice_access_vars.snap.new
test validate::choice_access_vars ... FAILED
stored new snapshot /build/argc/src/argc-1.1.0/tests/snapshots/integration__validate__option_choice_multi.snap.new
stored new snapshot /build/argc/src/argc-1.1.0/tests/snapshots/integration__spec__arg_eval.snap.new
test validate::option_choice_multi ... FAILED
test spec::arg_eval ... FAILED
test interrupt::interrupt ... ok

failures:

---- validate::arg_choice_fn stdout ----
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Snapshot Summary ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Snapshot file: tests/snapshots/integration__validate__arg_choice_fn.snap
Snapshot: arg_choice_fn
Source: tests/validate.rs:64
────────────────────────────────────────────────────────────────────────────────
Expression: data
────────────────────────────────────────────────────────────────────────────────
-old snapshot
+new results
────────────┬───────────────────────────────────────────────────────────────────
    2     2 │
    3     3 │ OUTPUT
    4     4 │ cat >&2 <<-'EOF'
    5     5 │ error: invalid value 'val' for '[VAL]'
    6       │-  [possible values: abc, def, ghi]
          6 │+  [possible values: ]
    7     7 │
    8     8 │ For more information, try '--help'.
    9     9 │
   10    10 │ EOF
────────────┴───────────────────────────────────────────────────────────────────
To update snapshots run `cargo insta review`
Stopped on the first failure. Run `cargo insta test` to run all snapshots.
thread 'validate::arg_choice_fn' panicked at 'snapshot assertion for 'arg_choice_fn' failed in line 64', /build/.cargo/registry/src/github.com-1ecc6299db9ec823/insta-1.29.0/src/runtime.rs:569:9

---- compgen::choice stdout ----
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Snapshot Summary ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Snapshot file: tests/snapshots/integration__compgen__choice.snap
Snapshot: choice
Source: tests/compgen.rs:164
────────────────────────────────────────────────────────────────────────────────
Expression: data
────────────────────────────────────────────────────────────────────────────────
-old snapshot
+new results
────────────┬───────────────────────────────────────────────────────────────────
    0     0 │ ************ COMPGEN `prog  --oa ` ************
    1       │-abc
    2       │-def
    3       │-ghi
    4     1 │
          2 │+
    5     3 │ ************ COMPGEN `prog  --ob ` ************
    6     4 │ x
    7     5 │ y
    8     6 │ z
┈┈┈┈┈┈┈┈┈┈┈┈┼┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈
   16    14 │
   17    15 │ ************ COMPGEN `prog  v1 ` ************
   18    16 │ --oa
   19    17 │ --ob
   20       │-abc
   21       │-def
   22       │-ghi
   23    18 │
────────────┴───────────────────────────────────────────────────────────────────
To update snapshots run `cargo insta review`
Stopped on the first failure. Run `cargo insta test` to run all snapshots.
thread 'compgen::choice' panicked at 'snapshot assertion for 'choice' failed in line 164', /build/.cargo/registry/src/github.com-1ecc6299db9ec823/insta-1.29.0/src/runtime.rs:569:9
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

---- validate::arg_choice_multi stdout ----
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Snapshot Summary ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Snapshot file: tests/snapshots/integration__validate__arg_choice_multi.snap
Snapshot: arg_choice_multi
Source: tests/validate.rs:87
────────────────────────────────────────────────────────────────────────────────
Expression: data
────────────────────────────────────────────────────────────────────────────────
-old snapshot
+new results
────────────┬───────────────────────────────────────────────────────────────────
    1     1 │ prog cmdl abc val
    2     2 │
    3     3 │ OUTPUT
    4     4 │ cat >&2 <<-'EOF'
    5       │-error: invalid value 'val' for '[VAL]...'
    6       │-  [possible values: abc, def, ghi]
          5 │+error: invalid value 'abc' for '[VAL]...'
          6 │+  [possible values: ]
    7     7 │
    8     8 │ For more information, try '--help'.
    9     9 │
   10    10 │ EOF
────────────┴───────────────────────────────────────────────────────────────────
To update snapshots run `cargo insta review`
Stopped on the first failure. Run `cargo insta test` to run all snapshots.
thread 'validate::arg_choice_multi' panicked at 'snapshot assertion for 'arg_choice_multi' failed in line 87', /build/.cargo/registry/src/github.com-1ecc6299db9ec823/insta-1.29.0/src/runtime.rs:569:9

---- validate::option_choice_fn stdout ----
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Snapshot Summary ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Snapshot file: tests/snapshots/integration__validate__option_choice_fn.snap
Snapshot: option_choice_fn
Source: tests/validate.rs:131
────────────────────────────────────────────────────────────────────────────────
Expression: data
────────────────────────────────────────────────────────────────────────────────
-old snapshot
+new results
────────────┬───────────────────────────────────────────────────────────────────
    2     2 │
    3     3 │ OUTPUT
    4     4 │ cat >&2 <<-'EOF'
    5     5 │ error: invalid value 'val' for '<CC>'
    6       │-  [possible values: abc, def, ghi]
          6 │+  [possible values: ]
    7     7 │
    8     8 │ For more information, try '--help'.
    9     9 │
   10    10 │ EOF
────────────┴───────────────────────────────────────────────────────────────────
To update snapshots run `cargo insta review`
Stopped on the first failure. Run `cargo insta test` to run all snapshots.
thread 'validate::option_choice_fn' panicked at 'snapshot assertion for 'option_choice_fn' failed in line 131', /build/.cargo/registry/src/github.com-1ecc6299db9ec823/insta-1.29.0/src/runtime.rs:569:9

---- spec::option_eval stdout ----
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Snapshot Summary ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Snapshot file: tests/snapshots/integration__spec__option_eval.snap
Snapshot: option_eval
Source: tests/spec.rs:21
────────────────────────────────────────────────────────────────────────────────
Expression: data
────────────────────────────────────────────────────────────────────────────────
-old snapshot
+new results
────────────┬───────────────────────────────────────────────────────────────────
   71    71 │ ************ RUN ************
   72    72 │ prog cmda --cc abc
   73    73 │
   74    74 │ OUTPUT
   75       │-argc_cc=abc
   76       │-argc__args=(  )
   77       │-argc__fn=cmda
   78       │-cmda
         75 │+cat >&2 <<-'EOF'
         76 │+error: invalid value 'abc' for '<CC>'
         77 │+  [possible values: ]
         78 │+
         79 │+For more information, try '--help'.
         80 │+
         81 │+EOF
         82 │+exit 1
   79    83 │
   80    84 │ ************ RUN ************
   81    85 │ prog cmda -soa soa
   82    86 │
────────────┴───────────────────────────────────────────────────────────────────
To update snapshots run `cargo insta review`
Stopped on the first failure. Run `cargo insta test` to run all snapshots.
thread 'spec::option_eval' panicked at 'snapshot assertion for 'option_eval' failed in line 21', /build/.cargo/registry/src/github.com-1ecc6299db9ec823/insta-1.29.0/src/runtime.rs:569:9

---- validate::choice_access_vars stdout ----
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Snapshot Summary ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Snapshot file: tests/snapshots/integration__validate__choice_access_vars.snap
Snapshot: choice_access_vars
Source: tests/validate.rs:209
────────────────────────────────────────────────────────────────────────────────
Expression: data
────────────────────────────────────────────────────────────────────────────────
-old snapshot
+new results
────────────┬───────────────────────────────────────────────────────────────────
    2     2 │
    3     3 │ OUTPUT
    4     4 │ cat >&2 <<-'EOF'
    5     5 │ error: invalid value 'foo' for '[VAL]'
    6       │-  [possible values: abc]
          6 │+  [possible values: ]
    7     7 │
    8     8 │ For more information, try '--help'.
    9     9 │
   10    10 │ EOF
┈┈┈┈┈┈┈┈┈┈┈┈┼┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈
   15    15 │
   16    16 │ OUTPUT
   17    17 │ cat >&2 <<-'EOF'
   18    18 │ error: invalid value 'foo' for '[VAL]'
   19       │-  [possible values: def]
         19 │+  [possible values: ]
   20    20 │
   21    21 │ For more information, try '--help'.
   22    22 │
   23    23 │ EOF
────────────┴───────────────────────────────────────────────────────────────────
To update snapshots run `cargo insta review`
Stopped on the first failure. Run `cargo insta test` to run all snapshots.
thread 'validate::choice_access_vars' panicked at 'snapshot assertion for 'choice_access_vars' failed in line 209', /build/.cargo/registry/src/github.com-1ecc6299db9ec823/insta-1.29.0/src/runtime.rs:569:9

---- validate::option_choice_multi stdout ----
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Snapshot Summary ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Snapshot file: tests/snapshots/integration__validate__option_choice_multi.snap
Snapshot: option_choice_multi
Source: tests/validate.rs:154
────────────────────────────────────────────────────────────────────────────────
Expression: data
────────────────────────────────────────────────────────────────────────────────
-old snapshot
+new results
────────────┬───────────────────────────────────────────────────────────────────
    1     1 │ prog cmda --ce abc val
    2     2 │
    3     3 │ OUTPUT
    4     4 │ cat >&2 <<-'EOF'
    5       │-error: invalid value 'val' for '<CE>'
    6       │-  [possible values: abc, def, ghi]
          5 │+error: invalid value 'abc' for '<CE>'
          6 │+  [possible values: ]
    7     7 │
    8     8 │ For more information, try '--help'.
    9     9 │
   10    10 │ EOF
────────────┴───────────────────────────────────────────────────────────────────
To update snapshots run `cargo insta review`
Stopped on the first failure. Run `cargo insta test` to run all snapshots.
thread 'validate::option_choice_multi' panicked at 'snapshot assertion for 'option_choice_multi' failed in line 154', /build/.cargo/registry/src/github.com-1ecc6299db9ec823/insta-1.29.0/src/runtime.rs:569:9

---- spec::arg_eval stdout ----
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Snapshot Summary ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Snapshot file: tests/snapshots/integration__spec__arg_eval.snap
Snapshot: arg_eval
Source: tests/spec.rs:58
────────────────────────────────────────────────────────────────────────────────
Expression: data
────────────────────────────────────────────────────────────────────────────────
-old snapshot
+new results
────────────┬───────────────────────────────────────────────────────────────────
   54    54 │ ************ RUN ************
   55    55 │ prog cmdj abc
   56    56 │
   57    57 │ OUTPUT
   58       │-argc_val=abc
   59       │-argc__args=( abc )
   60       │-argc__fn=cmdj
   61       │-cmdj abc
         58 │+cat >&2 <<-'EOF'
         59 │+error: invalid value 'abc' for '[VAL]'
         60 │+  [possible values: ]
         61 │+
         62 │+For more information, try '--help'.
         63 │+
         64 │+EOF
         65 │+exit 1
   62    66 │
   63    67 │ ************ RUN ************
   64    68 │ prog cmdp v1 v2
   65    69 │
────────────┴───────────────────────────────────────────────────────────────────
To update snapshots run `cargo insta review`
Stopped on the first failure. Run `cargo insta test` to run all snapshots.
thread 'spec::arg_eval' panicked at 'snapshot assertion for 'arg_eval' failed in line 58', /build/.cargo/registry/src/github.com-1ecc6299db9ec823/insta-1.29.0/src/runtime.rs:569:9


failures:
    compgen::choice
    spec::arg_eval
    spec::option_eval
    validate::arg_choice_fn
    validate::arg_choice_multi
    validate::choice_access_vars
    validate::option_choice_fn
    validate::option_choice_multi

test result: FAILED. 71 passed; 8 failed; 0 ignored; 0 measured; 0 filtered out; finished in 2.02s

error: test failed, to rerun pass `--test integration`

I'm wondering if there is a setup procedure for test or maybe a recommended way to run them?