path: root/doc/bash-style.cli

// file      : doc/bash-style.cli
// license   : MIT; see accompanying LICENSE file

"\title=Bash Style Guide"

// NOTES
//
// - Maximum <pre> line is 70 characters.
//

"\h1|Table of Contents|"
"\$TOC$"

"
\h1#intro|Introduction|

Bash works best for simple tasks. Needing arrays, arithmetic, and so on, is
usually a good indication that the task at hand is too complex for Bash.

Most of the below rules can be broken if there is a good reason for it.
Besides making things consistent, rules free you from having to stop and think
every time you encounter a particular situation. But if it feels that the
prescribed way is clearly wrong, then it probably makes sense to break it.
You just need to be clear on why you are doing it.

See also \l{https://google.github.io/styleguide/shell.xml Google's Bash Style
Guide} as well as \l{https://github.com/progrium/bashstyle Let's do Bash
right!}; we agree with quite a few (but not all) items in there. In particular,
the former provides a lot more rationale compared to this guide.

\h1#style|Style|

Don't use any extensions for your scripts. That is, call it just \c{foo}
rather than \c{foo.sh} or \c{foo.bash}. Use lower-case letters and dash
to separate words, for example \c{foo-bar}.

Indentation is two spaces (not tabs). Maximum line length is 79 characters
(excluding newline). Use blank lines between logical blocks to improve
readability.

Variable and function names should use lower-case letters with underscores
separating words.

For \c{if}/\c{while} and \c{for}/\c{do} the corresponding \c{then} or \c{do}
is written on the same line after a semicolon, for example:

\
if [ ... ]; then
fi

for x in ...; do
done
\

Do use \c{elif} instead of nested \c{else} and \c{if}.

For \c{if} use \c{[ ]} for basic tests and \c{[[ ]]} if the previous form is
not sufficient or hairy. In particular, \c{[[ ]]} results in cleaner code
for complex expressions, for example:

\
if [[ \"$foo\" && (\"$bar\" || \"$baz\") ]]; then
  ...
fi
\

\h1#struct|Structure|

The overall structure of the script should be as follows:

\
#! /usr/bin/env bash

# <SUMMARY>
#
# [<FUNCTIONALITY-DESCRIPTION>]
#
# [<OPTIONS-DESCRIPTION>]
#
usage=\"usage: $0 <OPTIONS>\"

owd=\"$(pwd)\"
trap \"{ cd '$owd'; exit 1; }\" ERR
set -o errtrace   # Trap in functions and subshells.
shopt -s lastpipe # Execute last pipeline command in current shell.

function info () { echo \"$*\" 1>&2; }
function error () { info \"$*\"; exit 1; }

[<OPTIONS-ARGUMENTS-DEFAULTS>]

[<OPTIONS-ARGUMENTS-PARSING>]

[<OPTIONS-ARGUMENTS-VALIDATION>]

<FUNCTIONALITY>
\

\h#struct-summary|SUMMARY|

One-two sentences describing what the script does.

\h#struct-func-desc|FUNCTIONALITY-DESCRIPTION|

More detailed functionality description for more complex scripts.

\h#struct-opt-desc|OPTIONS-DESCRIPTION|

Description of command line options. For example:

\
# -q
#   Run quiet.
#
# -t <dir>
#   Specify the alternative toolchain installation directory.
\

\h#struct-opt|OPTIONS|

Command line options summary. For example:

\
usage=\"usage: $0 [-q] [-t <dir>] <file>\"
\

\h#struct-opt-arg-default|OPTIONS-ARGUMENTS-DEFAULTS|

Set defaults to variables that will contain option/argument values. For
example:

\
quiet=\"n\"
tools=/usr/local
file=
\

\h#struct-opt-arg-parse|OPTIONS-ARGUMENTS-PARSING|

Parse the command line options/arguments. For example:

\
while [ \"$#\" -gt 0 ]; do
  case \"$1\" in
    -q)
      quiet=\"y\"
      shift
      ;;
    -t)
      shift
      tools=\"${1%/}\"
      shift
      ;;
    *)
      if [ -n \"$file\" ]; then
        error \"$usage\"
      fi

      file=\"$1\"
      shift
      ;;
  esac
done
\

If the value you are expecting from the command line is a directory path,
then always strip the trailing slash (as shown above for the \c{-t} option).

\h#struct-opt-arg-valid|OPTIONS-ARGUMENTS-VALIDATION|

Validate option/argument values. For example:

\
if [ -z \"$file\" ]; then
  error \"$usage\"
fi

if [ ! -d \"$file\" ]; then
  fail \"'$file' does not exist or is not a directory\"
fi
\

\h#struct-func|FUNCTIONALITY|

Implement script logic. For diagnostics use the \c{info()} and \c{error()}
functions defined above (so that it goes to stderr, not stdout). If using
functions, then define them just before use.

\h1#quote|Quoting|

We quote every variable expansion, no exceptions. For example:

\
if [ -n \"$foo\" ]; then
  ...
fi
\

This also applies to command substitution (which we always write as
\c{$(foo arg)} rather than \c{`foo arg`}), for example:

\
list=\"$(cat foo)\"
\

Note that a command substitution creates a new quoting context, for example:

\
list=\"$(basename \"$1\")\"
\

We also quote values that are \i{strings} as opposed to options/file names,
paths, enum-like values, or integers. Prefer single quotes for \c{sed}
scripts, for example:

\
url=\"https://example.org\"  # String.
quiet=y                    # Enum-like.
verbosity=1                # Integer.
dir=/etc                   # Directory path.
out=/dev/null              # File path.
file=manifest              # File name.
option=--quiet             # Option name.
seds='s%^./%%'             # sed script.
\

Note that quoting will inhibit globbing so you may end up with expansions
along these lines:

\
rm -f \"$dir/$name\".*
\

\N|One exception to this quoting rule is arithmetic expansion (\c{$((\ ))}):
Bash treats it as if it was double-quoted and, as a result, any inner quoting
is treated literally. For example:

\
z=$(($x + $y))           # Ok.
z=$((\"$x\" + \"$y\"))       # Error.
z=$(($x + $(echo \"$y\"))) # Ok.
\

|


If you have multiple values (e.g., program arguments) that may contain spaces,
don't try to handle them with quoting and use arrays instead. Here is a
typical example of a space-aware argument handling:

\
files=()

while [ \"$#\" -gt 0 ]; do
  case \"$1\" in

    ...

    *)
      shift
      files=(\"${files[@]}\" \"$1\")
      shift
      ;;
  esac
done

rm -f \"${files[@]}\"
\

In the same vein, never write:

\
cmd $*
\

Instead always write:

\
cmd \"$@\"
\

Also understand the difference between \c{@} and \c{*} expansion:

\
files=('one' '2 two' 'three')
echo \"files: ${files[@]}\"  # $1='files: one', $2='2 two', $3='three'
echo \"files: ${files[*]}\"  # $1='files: one 2 two three'
\

\h1#subshell|Subshell|

Bush executes certain constructs in \i{subshells} and some of these constructs
may not be obvious:

\ul|

\li|Explicit subshell: \c{(...)}|

\li|Pipeline: \c{...|...}|

\li|Command substitution: \c{$(...)}|

\li|Process substitution: \c{<(...)}, \c{>(...)}|

\li|Background: \c{...&}, \c{coproc ...}|

|

Naturally, a subshell cannot modify any state in the parent shell, which
sometimes leads to counter-intuitive behavior, for example:

\
lines=()

... | while read l; do
  lines+=(\"$l\")
done
\

At the end of the loop, \c{lines} will remain empty since the loop body is
executed in a subshell. One way to resolve this is to use the program
substitution instead of the pipeline:

\
lines=()

while read l; do
  lines+=(\"$l\")
done < <(...)
\

This, however, results in an unnatural, backwards-looking (compared to the
pipeline) code. Instead, we can request the last command of the pipeline to be
executed in the parent shell with the \c{lastpipe} shell option, for example:

\
shopt -s lastpipe

lines=()

... | while read l; do
  lines+=(\"$l\")
done
\

\N|The \c{lastpipe} shell option is inherited by functions and subshells.|


\h1#error-handing|Error Handling|

Our scripts use the error trap to automatically terminate the script in case
any command fails. This is also propagated to functions and subshells by
specifying the \c{errtrace} shell option.

\N|While the \c{pipefail} and \c{nounset} options may also seem like a good
idea, they have subtle, often latent pitfalls that make them more trouble than
they are worth (see \l{https://mywiki.wooledge.org/BashPitfalls#pipefail
\c{pipefail} pitfalls}, \l{https://mywiki.wooledge.org/BashPitfalls#nounset
\c{nounset} pitfalls}).

In particular, without \c{pipefail}, non-zero exist of any command in the
pipeline except the last is ignored. As a result, the pipeline needs to be
designed to work correctly in such cases, normally by relying on the input (or
lack thereof) to the last command to convey the failure. Alternatively, the
exit status of the pipeline commands can be explicitly checked using the
\c{PIPESTATUS} array.|

If you need to check the exit status of a command, use \c{if}, for example:

\
if grep \"foo\" /tmp/bar; then
  info \"found\"
fi

if ! grep \"foo\" /tmp/bar; then
  info \"not found\"
fi
\

Note that the \c{if}-condition can be combined with capturing the output, for
example:

\
if v=\"$(...)\"; then
  ...
fi
\

But keep in mind that in Bash a failure is often indistinguishable from a
true/false result. For example, in the above \c{grep} command, the result will
be the same whether there is no match or if the file does not exist.

Furthermore, in certain contexts, the above-mentioned error trap is ignored.
Quoting from the Bash manual:

\i{The \c{ERR} trap is not executed if the failed command is part of the
command list immediately following an \c{until} or \c{while} keyword, part of
the test following the \c{if} or \c{elif} reserved words, part of a command
executed in a \c{&&} or \c{||} list except the command following the final
\c{&&} or \c{||}, any command in a pipeline but the last, or if the command’s
return status is being inverted using \c{!}. These are the same conditions
obeyed by the \c{errexit} (\c{-e}) option.}

To illustrate the gravity of this point, consider the following example:

\
function cleanup()
{
  cd \"$1\"
  rm -f *
}

if ! cleanup /no/such/dir; then
  ...
fi
\

Here, the \c{cleanup()} function will continue executing (and may succeed)
even if the \c{cd} command has failed.

Note, however, that notwithstanding the above statement from the Bash manual,
the trap is executed in all the commands of a pipeline provided the
\c{errtrace} option is specified (presumably because commands of a pipeline
are said to execute in subshells). As a result, the above code can be made to
work using the pipe trick:

\
cleanup /no/such/dir | cat

if [ \"${PIPESTATUS[0]}\" -ne 0 ]; then
  ...
fi
\

\N|If the \c{pipefail} shell option is set, then the explicit \c{PIPESTATUS}
check is not necessary since the function failure will trigger the error trap
in the current shell.|

The recommendation is then to avoid calling functions in contexts where the
error trap is ignored resorting to the pipe trick where that's not possible.
And to be mindful of the potential ambiguity between the true/false result and
failure for other commands. The use of the \c{&&} and \c{||} command
expressions is best left to the interactive shell.

\N|The pipe trick cannot be used if the function needs to modify the global
state. Such a function, however, can return the exit status also as part of
the global state. The pipe trick can also be used to to ignore the exit status
of a command (provided \c{pipefail} is not set).|

The pipe trick can also be used to distinguish between different exit codes,
for example:

\
function foo()
{
  bar  # If this command fails, the function returns 1.

  if ... ; then
    return 2
  fi
}

foo | cat

case \"${PIPESTATUS[0]}\" in
  0)
    ;;
  1)
    exit 1
    ;;
  2)
    ...
    ;;
esac
\

\N|In such functions it makes sense to keep exit code 1 to mean failure so
that the inherited error trap can be re-used.|

This technique can be further extended to implement functions that both
return multiple exit codes and produce output, for example:

\
function foo()
{
  bar  # If this command fails, the function returns 1.

  if ... ; then
    return 2
  fi

  echo result
}

foo | readarray -t r

case \"${PIPESTATUS[0]}\" in
  0)
    echo \"${r[0]}\"
    ;;
  1)
    exit 1
    ;;
  2)
    ...
    ;;
esac
\

\N|We use \c{readarray} instead of \c{read} since the latter fails if the left
hand side of the pipeline does not produce anything.|


\h1#bool|Boolean|

For boolean values use empty for false and \c{true} for true. This way you
can have terse and natural looking conditions, for example:

\
first=true
while ...; do

  if [ ! \"$first\" ]; then
     ...
  fi

  if [ \"$first\" ]; then
     first=
  fi

done
\

\h1#function|Functions|

If a function takes arguments, provide a brief usage after the function
header, for example:

\
function dist() # <pkg> <dir>
{
  ...
}
\

For non-trivial/obvious functions also provide a short description of its
functionality/purpose, for example:

\
# Prepare a distribution of the specified packages and place it
# into the specified directory.
#
function dist() # <pkg> <dir>
{
  ...
}
\

Inside functions use local variables, for example:

\
function dist()
{
  local x=\"foo\"
}
\

If the evaluation of the value may fail (e.g., it contains a program
substitution), then place the assignment on a separate line since \c{local}
will cause the error to be ignored. For example:

\
function dist()
{
  local b
  b=\"$(basename \"$2\")\"
}
\

A function can return data in two primary ways: exit code and stdout.
Normally, exit code 0 means success and exit code 1 means failure though
additional codes can be used to distinguish between different kinds of
failures, signify special conditions, etc., see \l{#error-handing Error
Handling} for details.

A function can also write to stdout with the result available to the caller in
the same way as from programs (command substitution, pipeline, etc). If a
function needs to return multiple values, then it can print them separated
with newlines with the caller using the \c{readarray} builtin to read them
into an indexed array, for example:

\
function foo ()
{
  echo one
  echo two
  echo three
}

foo | readarray -t r
\

In this case, if the function can fail, then the failure should be explicitly
checked for (either by examining \c{PIPESTATUS} or via the lack of the
result), since the error trap will not be triggered (unless the \c{pipefail}
shell option is set; see \l{#error-handing Error Handling} for details). For
example:

\
foo | readarray -t r

if [ \"${PIPESTATUS[0]}\" -ne 0 ]; then
  exit 1
fi
\

\N|The use of the newline as a separator means that values may not contain
newlines. While \c{readarray} supports specifying a custom separator with the
\c{-d} option, including a \c{NUL} separator, this support is only available
since Bash 4.4.|

This technique can also be extended to return an associative array by
returning the values as an indexed array and then converting them to
an associative array with \c{eval}, for example:

\
function foo ()
{
  echo \"[a]=one\"
  echo \"[b]=two\"
  echo \"[c]=three\"
}

foo | readarray -t ia

if [ \"${PIPESTATUS[0]}\" -ne 0 ]; then
  exit 1
fi

eval declare -A aa=(\"${ia[@]}\")
\

Note that if a key or a value contains whitespaces, then it must be quoted.
The recommendation is to always quote both, for example:

\
function foo ()
{
  echo \"['a']='one ONE'\"
  echo \"['b']='two'\"
  echo \"['c']='three'\"
}
\

Or, if returning a local array:

\
function foo ()
{
  declare -A a=([a]='one ONE' [b]=two [c]=three)

  for k in \"${!a[@]}\"; do
    echo \"['$k']='${a[$k]}'\"
  done
}
\

For more information on returning data from functions, see
\l{https://mywiki.wooledge.org/BashFAQ/084 BashFAQ#084}.

"