sage-logger

#!/usr/bin/env bash
#
# sage-logger [-p] COMMAND LOGFILE
#
# Evaluate shell command COMMAND while logging stdout and stderr to
# LOGFILE. If either the command or the logging failed, return a
# non-zero exit status.
#
# If the -p argument is given, each line printed to stdout is prefixed
# with the name of the log file.
#
# AUTHOR:
#
# - Jeroen Demeyer (2015-07-26): initial version based on old pipestatus
#   script (#18953)
#
#*****************************************************************************
#       Copyright (C) 2015 Jeroen Demeyer <jdemeyer@cage.ugent.be>
#
#  Distributed under the terms of the GNU General Public License (GPL)
#  as published by the Free Software Foundation; either version 2 of
#  the License, or (at your option) any later version.
#                  http://www.gnu.org/licenses/
#*****************************************************************************

use_prefix=false

if [[ "$1" = -p ]]; then
    use_prefix=true
    shift
fi

cmd="$1"
logfile="$2"
logname="$(basename $logfile .log)"
logdir=`dirname "$logfile"`

if [[ $use_prefix = true ]]; then
    prefix="[${logname}] "
else
    prefix=""
fi

# Use sed option to reduce buffering, to make the output appear more
# smoothly. For GNU sed, this is the --unbuffered option.
# For BSD sed (which is also on OS X), this is the -l option.
if sed </dev/null 2>/dev/null --unbuffered ""; then
    SED="sed --unbuffered"
elif sed </dev/null 2>/dev/null -l ""; then
    SED="sed -l"
else
    SED="sed"
fi

mkdir -p "$logdir"

# Redirect stdout and stderr to a subprocess running tee.
# We trap SIGINT such that SIGINT interrupts the main process being
# run, not the logging.
( exec 2>&1; eval "$cmd" ) | \
    ( trap '' SIGINT; tee -a "$logfile" | $SED "s/^/$prefix/" )

pipestatus=(${PIPESTATUS[*]})

if [ ${pipestatus[1]} -ne 0 ]; then
    exit ${pipestatus[1]}
else
    exit ${pipestatus[0]}
fi