[Buildroot] [PATCH] support/scripts: tool to create fragments

[Yann E. MORIN]

Matt, Sam, All,

On 2016-10-25 15:01 -0500, Matt Weber spake thusly:
> From: Sam Voss <samuel.voss at rockwellcollins.com>
> 
> Add script to create kconfig fragment; defaults to busybox fragment if
> only one config is specified (compares to package/busybox/busybox.config).

As stated by Thomas, this is re-inventing diffconfig.

The only difference being the output format.

Currently diffconfig can output a diff-like, but it can also output a
"merge" style (although I'm not sure what it means, given that a merge
on your example is just empty).

I think it would be much better to update diffconfig to allow it to spit
out a fragment format. Maybe something like:

    $ diffconfig -f config.base config.modified
    CONFIG_DEVTMPFS_MOUNT=y
    CONFIG_DEVTMPFS=y
    # CONFIG_X86_MPPARSE is not set

Thoughts?

Note: either way, I'd prefer both files to be mandatory arguments, with
the original one specified first and the modified one specified last,
and that there is no default comparison against busybox.

Otherwise, some review below...

> Signed-off-by: Sam Voss <samuel.voss at rockwellcollins.com>
> Signed-off-by: Matt Weber <matthew.weber at rockwellcollins.com>
> ---
>  support/scripts/gen-config-fragment.sh | 59 ++++++++++++++++++++++++++++++++++
>  1 file changed, 59 insertions(+)
>  create mode 100755 support/scripts/gen-config-fragment.sh
> 
> diff --git a/support/scripts/gen-config-fragment.sh b/support/scripts/gen-config-fragment.sh
> new file mode 100755
> index 0000000..a8b5345
> --- /dev/null
> +++ b/support/scripts/gen-config-fragment.sh
> @@ -0,0 +1,59 @@
> +#!/bin/bash
> +
> +################################################################################
> +# DESCRIPTION:
> +#     Creates a fragment by comparing two kconfigs. Default usage creates a
> +#       busybox fragment, however providing two config files will override this.
> +#
> +#     Required arguments:
> +#       $1 - config #1. This is generally the new one created
> +#
> +#     Optional arguments:
> +#       $2 - config #2: The configuration file to compare against. If none
> +#            supplied, will compare against package/busybox/busybox.config
> +################################################################################
> +usage ()
> +{
> +use="gen-config-fragment.sh:
> +    Creates a fragment by comparing two kconfigs. Default usage creates a
> +      busybox fragment, however providing two config files will override this.
> +
> +    Required arguments:
> +      $1 - config #1. This is generally the new one created
> +
> +    Optional arguments:
> +      $2 - config #2: The configuration file to compare against. If none
> +           supplied, will compare against package/busybox/busybox.config"
> +        printf "$use"
> +        exit 1

Dont use a variable; use an here-document (leading TABS, repesented by
»»»» here, are ignored):

    cat <<-_EOF_
    »»»»gen-config-fragment.sh:
    »»»»    Creates a fragment by comparing two kconfigs.

    »»»»Arguments:
    »»»»    $1 : the unmodified original configuration file
    »»»»    $2 : the modified configuration file

    »»»»Blabla...
    _EOF_

> +}
> +
> +if [ -z "$1" ]; then
> +        usage
> +else
> +        FirstConfig="$1"
> +fi
> +
> +if [ -z "$2" ]; then
> +        SecondConfig="package/busybox/busybox.config"
> +else
> +        SecondConfig="$2"
> +fi
> +
> +loadedFirstConfig=()
> +loadedSecondConfig=()

Declare the variables:

    declare -a loadedFirstConfig loadedSecondConfig

> +while read line; do
> +        if [[ ${line:0:1} != '#' ]] || [[ ${line:2:3} == 'BR2' ]]; then

Use just plain '[ ... ]' to do tests, since they are POSIX; quote
strings:

    if [ "${line:0:3}" = "BR2_" -o "${line:2:4}" = "BR2_" ]; then

> +                loadedFirstConfig+=("${line// /_space_}")

I'm not too fond of the magic placeholder _space_. But I can't find an
easy and better solution... Maybe make it even less probable, like:
    _S_P_A_C_E_

> +        fi
> +done < $FirstConfig

Quote strings:  "${FirstConfig}"

But there is a better and faster way to fill in the array:

    loadedFirstConfig=( $( sed -r -e 's/ /_S_P_A_C_E_/g' "${FirstConfig}" ) )

But it turns out you don;t need the array in the first place, see below.

> +
> +while read -r line; do
> +        if [[ ${line:0:1} != '#' ]] || [[ ${line:2:3} == 'BR2' ]]; then
> +                loadedSecondConfig+=("${line// /_space_}")
> +        fi
> +done < $SecondConfig
> +
> +
> +D=($(comm -23 <(printf '%s\n' "${loadedFirstConfig[@]}" | sort -d) <(printf '%s\n' "${loadedSecondConfig[@]}" | sort -d)))

Line too long; split it.

> +printf '%s\n' "${D[@]//_space_/ }" > "$FirstConfig.fragment"

Just output to stdout; let the user redirect to the file of his choice.

But we can do it much more simply, in a single command. Your script
would be just (without even a requirement for bash):

    #!/bin/sh

    usage() {
        cat <<-_EOF_
        Blabla
        _EOF_
    }

    [ ${#} -eq 2 ] || { usage; exit 1; }
    [ -f "${1}" ]  || { usage; exit 1; }
    [ -f "${2}" ]  || { usage; exit 1; }

    comm -23 <( sed -r -e 's/ /_S_P_A_C_E_/g' "${FirstConfig}"  |sort -d ) \
             <( sed -r -e 's/ /_S_P_A_C_E_/g' "${SecondConfig}" |sort -d ) \
    |sed -r -e 's/_S_P_A_C_E_/ /g'

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'