# Copyright 1999-2003 Gentoo Technologies, Inc. # Distributed under the terms of the GNU General Public License v2 # $Header: /var/cvsroot/gentoo-x86/eclass/eutils.eclass,v 1.33 2003/06/12 18:21:23 vapier Exp $ # # Author: Martin Schlemmer # # This eclass is for general purpose functions that most ebuilds # have to implement themselves. # # NB: If you add anything, please comment it! ECLASS=eutils INHERITED="$INHERITED $ECLASS" DEPEND="$DEPEND !bootstrap? ( sys-devel/patch )" DESCRIPTION="Based on the ${ECLASS} eclass" # This function generate linker scripts in /usr/lib for dynamic # libs in /lib. This is to fix linking problems when you have # the .so in /lib, and the .a in /usr/lib. What happens is that # in some cases when linking dynamic, the .a in /usr/lib is used # instead of the .so in /lib due to gcc/libtool tweaking ld's # library search path. This cause many builds to fail. # See bug #4411 for more info. # # To use, simply call: # # gen_usr_ldscript libfoo.so # # Note that you should in general use the unversioned name of # the library, as ldconfig should usually update it correctly # to point to the latest version of the library present. # # (26 Oct 2002) # gen_usr_ldscript() { # Just make sure it exists dodir /usr/lib cat > ${D}/usr/lib/$1 <<"END_LDSCRIPT" /* GNU ld script Because Gentoo have critical dynamic libraries in /lib, and the static versions in /usr/lib, we need to have a "fake" dynamic lib in /usr/lib, otherwise we run into linking problems. See bug #4411 on http://bugs.gentoo.org/ for more info. */ GROUP ( /lib/libxxx ) END_LDSCRIPT dosed "s:libxxx:$1:" /usr/lib/$1 return 0 } # Simple function to draw a line consisting of '=' the same length as $* # # (11 Nov 2002) # draw_line() { local i=0 local str_length="" # Handle calls that do not have args, or wc not being installed ... if [ -z "$1" -o ! -x "$(which wc 2>/dev/null)" ] then echo "===============================================================" return 0 fi # Get the length of $* str_length="$(echo -n "$*" | wc -m)" while [ "$i" -lt "${str_length}" ] do echo -n "=" i=$((i + 1)) done echo return 0 } # Default directory where patches are located EPATCH_SOURCE="${WORKDIR}/patch" # Default extension for patches EPATCH_SUFFIX="patch.bz2" # Default options for patch EPATCH_OPTS="" # List of patches not to apply. Not this is only file names, # and not the full path .. EPATCH_EXCLUDE="" # Change the printed message for a single patch. EPATCH_SINGLE_MSG="" # Force applying bulk patches even if not following the style: # # ??_${ARCH}_foo.${EPATCH_SUFFIX} # EPATCH_FORCE="no" # This function is for bulk patching, or in theory for just one # or two patches. # # It should work with .bz2, .gz, .zip and plain text patches. # Currently all patches should be the same format. # # You do not have to specify '-p' option to patch, as it will # try with -p0 to -p5 until it succeed, or fail at -p5. # # Above EPATCH_* variables can be used to control various defaults, # bug they should be left as is to ensure an ebuild can rely on # them for. # # Patches are applied in current directory. # # Bulk Patches should preferibly have the form of: # # ??_${ARCH}_foo.${EPATCH_SUFFIX} # # For example: # # 01_all_misc-fix.patch.bz2 # 02_sparc_another-fix.patch.bz2 # # This ensures that there are a set order, and you can have ARCH # specific patches. # # If you however give an argument to epatch(), it will treat it as a # single patch that need to be applied if its a file. If on the other # hand its a directory, it will set EPATCH_SOURCE to this. # # (10 Nov 2002) # epatch() { local PIPE_CMD="" local STDERR_TARGET="${T}/$$.out" local PATCH_TARGET="${T}/$$.patch" local PATCH_SUFFIX="" local SINGLE_PATCH="no" local x="" if [ "$#" -gt 1 ] then eerror "Invalid arguments to epatch()" die "Invalid arguments to epatch()" fi if [ -n "$1" -a -f "$1" ] then SINGLE_PATCH="yes" local EPATCH_SOURCE="$1" local EPATCH_SUFFIX="${1##*\.}" elif [ -n "$1" -a -d "$1" ] then # Allow no extension if EPATCH_FORCE=yes ... used by vim for example ... if [ "${EPATCH_FORCE}" = "yes" ] && [ -z "${EPATCH_SUFFIX}" ] then local EPATCH_SOURCE="$1/*" else local EPATCH_SOURCE="$1/*.${EPATCH_SUFFIX}" fi else if [ ! -d ${EPATCH_SOURCE} ] then if [ -n "$1" -a "${EPATCH_SOURCE}" = "${WORKDIR}/patch" ] then EPATCH_SOURCE="$1" fi echo eerror "Cannot find \$EPATCH_SOURCE! Value for \$EPATCH_SOURCE is:" eerror eerror " ${EPATCH_SOURCE}" echo die "Cannot find \$EPATCH_SOURCE!" fi local EPATCH_SOURCE="${EPATCH_SOURCE}/*.${EPATCH_SUFFIX}" fi case ${EPATCH_SUFFIX##*\.} in bz2) PIPE_CMD="bzip2 -dc" PATCH_SUFFIX="bz2" ;; gz|Z|z) PIPE_CMD="gzip -dc" PATCH_SUFFIX="gz" ;; ZIP|zip) PIPE_CMD="unzip -p" PATCH_SUFFIX="zip" ;; *) PIPE_CMD="cat" PATCH_SUFFIX="patch" ;; esac if [ "${SINGLE_PATCH}" = "no" ] then einfo "Applying various patches (bugfixes/updates)..." fi for x in ${EPATCH_SOURCE} do # New ARCH dependant patch naming scheme... # # ???_arch_foo.patch # if [ -f ${x} ] && \ ([ "${SINGLE_PATCH}" = "yes" -o "${x/_all_}" != "${x}" -o "`eval echo \$\{x/_${ARCH}_\}`" != "${x}" ] || \ [ "${EPATCH_FORCE}" = "yes" ]) then local count=0 local popts="${EPATCH_OPTS}" if [ -n "${EPATCH_EXCLUDE}" ] then if [ "`eval echo \$\{EPATCH_EXCLUDE/${x##*/}\}`" != "${EPATCH_EXCLUDE}" ] then continue fi fi if [ "${SINGLE_PATCH}" = "yes" ] then if [ -n "${EPATCH_SINGLE_MSG}" ] then einfo "${EPATCH_SINGLE_MSG}" else einfo "Applying ${x##*/}..." fi else einfo " ${x##*/}..." fi echo "***** ${x##*/} *****" > ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/} echo >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/} # Allow for prefix to differ ... im lazy, so shoot me :/ while [ "${count}" -lt 5 ] do # Generate some useful debug info ... draw_line "***** ${x##*/} *****" >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/} echo >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/} if [ "${PATCH_SUFFIX}" != "patch" ] then echo -n "PIPE_COMMAND: " >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/} echo "${PIPE_CMD} ${x} > ${PATCH_TARGET}" >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/} else PATCH_TARGET="${x}" fi echo -n "PATCH COMMAND: " >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/} echo "patch ${popts} -p${count} < ${PATCH_TARGET}" >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/} echo >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/} draw_line "***** ${x##*/} *****" >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/} if [ "${PATCH_SUFFIX}" != "patch" ] then if ! (${PIPE_CMD} ${x} > ${PATCH_TARGET}) >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/} 2>&1 then echo eerror "Could not extract patch!" #die "Could not extract patch!" count=5 break fi fi if (cat ${PATCH_TARGET} | patch ${popts} --dry-run -f -p${count}) >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/} 2>&1 then draw_line "***** ${x##*/} *****" > ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/}.real echo >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/}.real echo "ACTUALLY APPLYING ${x##*/}..." >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/}.real echo >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/}.real draw_line "***** ${x##*/} *****" >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/}.real cat ${PATCH_TARGET} | patch ${popts} -p${count} >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/}.real 2>&1 if [ "$?" -ne 0 ] then cat ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/}.real >> ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/} echo eerror "A dry-run of patch command succeeded, but actually" eerror "applying the patch failed!" #die "Real world sux compared to the dreamworld!" count=5 fi rm -f ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/}.real break fi count=$((count + 1)) done if [ "${PATCH_SUFFIX}" != "patch" ] then rm -f ${PATCH_TARGET} fi if [ "${count}" -eq 5 ] then echo eerror "Failed Patch: ${x##*/}!" eerror eerror "Include in your bugreport the contents of:" eerror eerror " ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/}" echo die "Failed Patch: ${x##*/}!" fi rm -f ${STDERR_TARGET%/*}/${x##*/}-${STDERR_TARGET##*/} eend 0 fi done if [ "${SINGLE_PATCH}" = "no" ] then einfo "Done with patching" fi } # This function return true if we are using the NPTL pthreads # implementation. # # (06 March 2003) # have_NPTL() { cat > ${T}/test-nptl.c <<-"END" #define _XOPEN_SOURCE #include #include int main() { char buf[255]; char *str = buf; confstr(_CS_GNU_LIBPTHREAD_VERSION, str, 255); if (NULL != str) { printf("%s\n", str); if (NULL != strstr(str, "NPTL")) return 0; } return 1; } END einfon "Checking for _CS_GNU_LIBPTHREAD_VERSION support in glibc ... " if gcc -o ${T}/nptl ${T}/test-nptl.c &> /dev/null then echo "yes" einfon "Checking what PTHREADS implementation we have ... " if ${T}/nptl then return 0 else return 1 fi else echo "no" fi return 1 } # This function check how many cpu's are present, and then set # -j in MAKEOPTS accordingly. # # Thanks to nall for this. # get_number_of_jobs() { local jobs=0 if [ ! -r /proc/cpuinfo ] then return 1 fi # This bit is from H?kan Wessberg , bug #13565. if [ "`egrep "^[[:space:]]*MAKEOPTS=" /etc/make.conf | wc -l`" -gt 0 ] then ADMINOPTS="`egrep "^[[:space:]]*MAKEOPTS=" /etc/make.conf | cut -d= -f2 | sed 's/\"//g'`" ADMINPARAM="`echo ${ADMINOPTS} | gawk '{match($0, /-j *[0-9]*/, opt); print opt[0]}'`" ADMINPARAM="${ADMINPARAM/-j}" fi export MAKEOPTS="`echo ${MAKEOPTS} | sed -e 's:-j *[0-9]*::g'`" if [ "${ARCH}" = "x86" -o "${ARCH}" = "hppa" -o \ "${ARCH}" = "arm" -o "${ARCH}" = "mips" ] then # these archs will always have "[Pp]rocessor" jobs="$((`grep -c ^[Pp]rocessor /proc/cpuinfo` * 2))" elif [ "${ARCH}" = "sparc" -o "${ARCH}" = "sparc64" ] then # sparc always has "ncpus active" jobs="$((`grep "^ncpus active" /proc/cpuinfo | sed -e "s/^.*: //"` * 2))" elif [ "${ARCH}" = "alpha" ] then # alpha has "cpus active", but only when compiled with SMP if [ "`grep -c "^cpus active" /proc/cpuinfo`" -eq 1 ] then jobs="$((`grep "^cpus active" /proc/cpuinfo | sed -e "s/^.*: //"` * 2))" else jobs=2 fi elif [ "${ARCH}" = "ppc" ] then # ppc has "processor", but only when compiled with SMP if [ "`grep -c "^processor" /proc/cpuinfo`" -eq 1 ] then jobs="$((`grep -c ^processor /proc/cpuinfo` * 2))" else jobs=2 fi else jobs="$((`grep -c ^cpu /proc/cpuinfo` * 2))" die "Unknown ARCH -- ${ARCH}!" fi # Make sure the number is valid ... if [ "${jobs}" -lt 1 ] then jobs=1 fi if [ -n "${ADMINPARAM}" ] then if [ "${jobs}" -gt "${ADMINPARAM}" ] then einfo "Setting make jobs to \"-j${ADMINPARAM}\" to ensure successful merge..." export MAKEOPTS="${MAKEOPTS} -j${ADMINPARAM}" else einfo "Setting make jobs to \"-j${jobs}\" to ensure successful merge..." export MAKEOPTS="${MAKEOPTS} -j${jobs}" fi fi } # Simplify/standardize adding users to the system # vapier@gentoo.org # # enewuser(username, uid, shell, homedir, groups, extra options) # # Default values if you do not specify any: # username: REQUIRED ! # uid: next available (see useradd(8)) # note: pass -1 to get default behavior # shell: /bin/false # homedir: /dev/null # groups: none # extra: comment of 'added by portage for ${PN}' enewuser() { # get the username local euser="$1"; shift if [ -z "${euser}" ] ; then eerror "No username specified !" die "Cannot call enewuser without a username" fi einfo "Adding user '${euser}' to your system ..." # setup a file for testing usernames/groups local tmpfile="`mktemp -p ${T}`" touch ${tmpfile} chown ${euser} ${tmpfile} >& /dev/null local realuser="`ls -l ${tmpfile} | awk '{print $3}'`" # see if user already exists if [ "${euser}" == "${realuser}" ] ; then einfo "${euser} already exists on your system :)" return 0 fi # options to pass to useradd local opts="" # handle uid local euid="$1"; shift if [ ! -z "${euid}" ] && [ "${euid}" != "-1" ] ; then if [ ${euid} -gt 0 ] ; then opts="${opts} -u ${euid}" else eerror "Userid given but is not greater than 0 !" die "${euid} is not a valid UID" fi else euid="next available" fi einfo " - Userid: ${euid}" # handle shell local eshell="$1"; shift if [ ! -z "${eshell}" ] ; then if [ ! -e ${eshell} ] ; then eerror "A shell was specified but it does not exist !" die "${eshell} does not exist" fi else eshell=/bin/false fi einfo " - Shell: ${eshell}" opts="${opts} -s ${eshell}" # handle homedir local ehome="$1"; shift if [ -z "${ehome}" ] ; then ehome=/dev/null fi einfo " - Home: ${ehome}" opts="${opts} -d ${ehome}" # handle groups local egroups="$1"; shift if [ ! -z "${egroups}" ] ; then local realgroup local oldifs="${IFS}" export IFS="," for g in ${egroups} ; do chgrp ${g} ${tmpfile} >& /dev/null realgroup="`ls -l ${tmpfile} | awk '{print $4}'`" if [ "${g}" != "${realgroup}" ] ; then eerror "You must add ${g} to the system first" die "${g} is not a valid GID" fi done export IFS="${oldifs}" opts="${opts} -g ${egroups}" else egroups="(none)" fi einfo " - Groups: ${egroups}" # handle extra and add the user local eextra="$@" local oldsandbox="${oldsandbox}" export SANDBOX_ON="0" if [ -z "${eextra}" ] ; then useradd ${opts} ${euser} \ -c "added by portage for ${PN}" \ || die "enewuser failed" else einfo " - Extra: ${eextra}" useradd ${opts} ${euser} ${eextra} \ || die "enewuser failed" fi export SANDBOX_ON="${oldsandbox}" if [ ! -e ${ehome} ] && [ ! -e ${D}/${ehome} ] ; then einfo " - Creating ${ehome} in ${D}" dodir ${ehome} fperms ${euser} ${ehome} fi } # Simplify/standardize adding groups to the system # vapier@gentoo.org # # enewgroup(group, gid) # # Default values if you do not specify any: # groupname: REQUIRED ! # gid: next available (see groupadd(8)) # extra: none enewgroup() { # get the group local egroup="$1"; shift if [ -z "${egroup}" ] ; then eerror "No group specified !" die "Cannot call enewgroup without a group" fi einfo "Adding group '${egroup}' to your system ..." # setup a file for testing groupname local tmpfile="`mktemp -p ${T}`" touch ${tmpfile} chgrp ${egroup} ${tmpfile} >& /dev/null local realgroup="`ls -l ${tmpfile} | awk '{print $4}'`" # see if group already exists if [ "${egroup}" == "${realgroup}" ] ; then einfo "${egroup} already exists on your system :)" return 0 fi # options to pass to useradd local opts="" # handle gid local egid="$1"; shift if [ ! -z "${egid}" ] ; then if [ ${egid} -gt 0 ] ; then opts="${opts} -g ${egid}" else eerror "Groupid given but is not greater than 0 !" die "${egid} is not a valid GID" fi else egid="next available" fi einfo " - Groupid: ${egid}" # handle extra local eextra="$@" opts="${opts} ${eextra}" # add the group local oldsandbox="${oldsandbox}" export SANDBOX_ON="0" groupadd ${opts} ${egroup} || die "enewgroup failed" export SANDBOX_ON="${oldsandbox}" } # Simple script to replace 'dos2unix' binaries # vapier@gentoo.org # # edos2unix(file, ...) edos2unix() { for f in $@ ; do cp ${f} ${T}/edos2unix sed 's/\r$//' ${T}/edos2unix > ${f} done } # new convinience patch wapper function to eventually replace epatch(), $PATCHES, $PATCHES1, src_unpack:patch, src_unpack:autopatch and /usr/bin/patch # Features: # - bulk patch handling similar to epatch()'s # - automatic patch level detection like epatch()'s # - semiautomatic patch uncompression like epatch()'s (may switch to using /usr/bin/file for extra power, instead of just looking at the filename) # - doesn't have the --dry-run overhead of epatch() - inspects patchfiles manually instead # - is called from base_src_unpack to handle $PATCHES to avoid defining src_unpack(-) just to use xpatch # accepts zero or more parameters specifying patchfiles and/or patchdirs # known issues: # - only supports unified style patches (does anyone _really_ use anything else?) # - first file addressed in a patch can't have spaces in its name or in the path mentioned in the patchfile # (can be easily fixed to be: at least one file addressed in the patch must have no spaces...) xpatch() { debug-print-function $FUNCNAME $* local list="" local list2="" declare -i plevel # parse patch sources for x in $*; do debug-print "$FUNCNAME: parsing parameter $x" if [ -f "$x" ]; then list="$list $x" elif [ -d "$x" ]; then # handles patchdirs like epatch() for now: no recursion. # patches are sorted by filename, so with an xy_foo naming scheme you'll get the right order. # only patches with _$ARCH_ or _all_ in their filenames are applied. for file in `ls -A $x`; do debug-print "$FUNCNAME: parsing in subdir: file $file" if [ -f "$x/$file" ] && [ "${file}" != "${file/_all_}" -o "${file}" != "${file/_$ARCH_}" ]; then list2="$list2 $x/$file" fi done list="`echo $list2 | sort` $list" else die "Couldn't find $x" fi done debug-print "$FUNCNAME: final list of patches: $list" for x in $list; do debug-print "$FUNCNAME: processing $x" # deal with compressed files. /usr/bin/file is in the system profile, or should be. case "`/usr/bin/file -b $x`" in *gzip*) patchfile="${T}/current.patch"; ungzip -c "$x" > "${patchfile}";; *bzip2*) patchfile="${T}/current.patch"; bunzip2 -c "$x" > "${patchfile}";; *text*) patchfile="$x";; *) die "Could not determine filetype of patch $x";; esac debug-print "$FUNCNAME: patchfile=$patchfile" # determine patchlevel. supports p0 and higher with either $S or $WORKDIR as base. target="`/bin/grep '+++' $patchfile | /usr/bin/tail -1`" debug-print "$FUNCNAME: raw target=$target" # strip target down to the path/filename. NOTE doesn't support filenames/paths with spaces in them :-( # remove leading +++ target="${target/+++ }" # ugly, yes. i dunno why doesn't this work instead: target=${target%% *} for foo in $target; do target="$foo"; break; done # duplicate slashes are discarded by patch wrt the patchlevel. therefore we need to discard them as well # to calculate the correct patchlevel. while [ "$target" != "${target/\/\/}" ]; do target="${target/\/\//\/}" done debug-print "$FUNCNAME: stripped target=$target" # look for target for basedir in "$S" "$WORKDIR" "`pwd`"; do debug-print "$FUNCNAME: looking in basedir=$basedir" cd "$basedir" # try stripping leading directories target2="$target" plevel=0 debug-print "$FUNCNAME: trying target2=$target2, plevel=$plevel" while [ ! -f "$target2" ]; do target2="${target2#*/}" # removes piece of target2 upto the first occurence of / plevel=plevel+1 debug-print "$FUNCNAME: trying target2=$target2, plevel=$plevel" [ "$target2" == "${target2/\/}" ] && break done test -f "$target2" && break # try stripping filename - needed to support patches creating new files target2="${target%/*}" plevel=0 debug-print "$FUNCNAME: trying target2=$target2, plevel=$plevel" while [ ! -d "$target2" ]; do target2="${target2#*/}" # removes piece of target2 upto the first occurence of / plevel=plevel+1 debug-print "$FUNCNAME: trying target2=$target2, plevel=$plevel" [ "$target2" == "${target2/\/}" ] && break done test -d "$target2" && break done test -f "${basedir}/${target2}" || test -d "${basedir}/${target2}" || die "Could not determine patchlevel for $x" debug-print "$FUNCNAME: determined plevel=$plevel" # do the patching ebegin "Applying patch ${x##*/}..." /usr/bin/patch -p$plevel < "$patchfile" > /dev/null || die "Failed to apply patch $x" eend $? done }