Loïc Minier <l...@dooz.org> writes:
> On Tue, Jan 19, 2010, Loïc Minier wrote:
>> Following the thread on the sdl-config patch, please find attached a
>> patch to add a couple of portable shell functions which allow testing
>> whehter a command/builtin is available and to find the full pathname of
>> an executable in the PATH. This also replaces all uses of "which" in
>> ./configure. (This should be applied on top of the sdl-config patch.)
>
> Please find attached a new version of the patch with a simpler version
> of path_of() which uses IFS instead of the ${foo#bar} and ${foo%%bar}
> constructs. It also removes the special casing of an empty PATH.
>
> --
> Loïc Minier
>
> From 5fc05ec61d87049ea0f29b2dd51c16e260698ef8 Mon Sep 17 00:00:00 2001
> From: =?UTF-8?q?Lo=C3=AFc=20Minier?= <l...@dooz.org>
> Date: Tue, 19 Jan 2010 11:05:00 +0100
> Subject: [PATCH] Add and use has() and path_of() funcs
>
> Add has() and path_of() funcs and use them across configure; has()
> will test whether a command or builtin is available; path_of() will
> search the PATH for executables and return the full pathname if found.
> ---
> configure | 53 ++++++++++++++++++++++++++++++++++++++++++++---------
> 1 files changed, 44 insertions(+), 9 deletions(-)
>
> diff --git a/configure b/configure
> index baa2800..711e335 100755
> --- a/configure
> +++ b/configure
> @@ -27,6 +27,43 @@ compile_prog() {
> $cc $QEMU_CFLAGS $local_cflags -o $TMPE $TMPC $LDFLAGS $local_ldflags >
> /dev/null 2> /dev/null
> }
>
> +# check whether a command is available to this shell (may be either an
> +# executable or a builtin)
> +has() {
> + local_command="$1"
> + type "$local_command" >/dev/null
> +}
Why the extra variable? Using $1 directly seems just as obvious to me.
> +# search for an executable in PATH
> +path_of() {
> + local_command="$1"
> + local_ifs="$IFS"
> + local_dir=""
> +
> + # pathname has a dir component?
> + if [ "${local_command#*/}" != "$local_command" ]; then
> + if [ -x "$local_command" ] && [ ! -d "$local_command" ]; then
> + echo "$local_command"
> + return 0
> + fi
> + fi
> + if [ -z "$local_command" ]; then
> + return 1
> + fi
> +
> + IFS=:
> + for local_dir in $PATH; do
> + if [ -x "$local_dir/$local_command" ] && [ ! -d
> "$local_dir/$local_command" ]; then
> + echo "$local_dir/$local_command"
> + IFS="$local_ifs"
> + return 0
> + fi
> + done
> + # not found
> + IFS="$local_ifs"
> + return 1
> +}
> +
> # default parameters
> cpu=""
> prefix=""
> @@ -763,7 +800,7 @@ fi
> # Solaris specific configure tool chain decisions
> #
> if test "$solaris" = "yes" ; then
> - solinst=`which $install 2> /dev/null | /usr/bin/grep -v "no $install in"`
> + solinst=`path_of $install`
> if test -z "$solinst" ; then
> echo "Solaris install program not found. Use --install=/usr/ucb/install
> or"
> echo "install fileutils from www.blastwave.org using pkg-get -i
> fileutils"
> @@ -776,7 +813,7 @@ if test "$solaris" = "yes" ; then
> echo "using pkg-get -i fileutils, or use --install=/usr/ucb/install"
> exit 1
> fi
> - sol_ar=`which ar 2> /dev/null | /usr/bin/grep -v "no ar in"`
> + sol_ar=`path_of ar`
> if test -z "$sol_ar" ; then
> echo "Error: No path includes ar"
> if test -f /usr/ccs/bin/ar ; then
Is the full path of these tools really important? Doesn't look like
it to me.
--
Måns Rullgård
m...@mansr.com
