Skip to content
Snippets Groups Projects
Commit d9625dd9 authored by Andreas Kusalananda Kähäri's avatar Andreas Kusalananda Kähäri
Browse files

Back to using tar instead of pax. Tru64 pax seems to cut off long path names.

Won't touch mart databases (or anything matching $avoid_re).
Will only work with revision 14 and 15 (or anything matching $require_re).
parent 0ef3f9c2
No related branches found
No related tags found
...@@ -17,15 +17,15 @@ function getdb ...@@ -17,15 +17,15 @@ function getdb
typeset dbver=${db}_${ver} typeset dbver=${db}_${ver}
typeset url='ftp://ftp.ensembl.org'${path}/${dbver}.tar path=${path#/}
typeset url='ftp://ftp.ensembl.org/'${path}/${dbver}.tar
if [[ ! -d databases ]]; then if [[ ! -d databases ]]; then
mkdir databases mkdir databases
fi fi
if [[ ! -d databases/${db}_${ver} ]]; then if [[ ! -d databases/${db}_${ver} ]]; then
trap "rm -rf databases/${db}_${ver}; exit 1" INT trap "rm -rf databases/${db}_${ver}; exit 1" INT
lynx -source $url | (cd databases; pax -r -v \ lynx -source $url | (cd databases; tar -x -v -f- -P"${path}/")
-s "/.*${dbver}\//${dbver}\//")
trap - INT trap - INT
fi fi
} }
...@@ -48,10 +48,11 @@ function do_delta ...@@ -48,10 +48,11 @@ function do_delta
typeset apply_out=deltas/${db}_${v1}_delta_${v2}_apply.out typeset apply_out=deltas/${db}_${v1}_delta_${v2}_apply.out
if [[ ! -f $build_out ]]; then if [[ ! -f $build_out ]]; then
trap "rm $build_out; exit 1" INT
getdb $path1 $db $v1 getdb $path1 $db $v1
getdb $path2 $db $v2 getdb $path2 $db $v2
trap "rm $build_out; exit 1" INT
/usr/bin/time perl -w ./build.pl -c ./xdelta.osf \ /usr/bin/time perl -w ./build.pl -c ./xdelta.osf \
-s databases -d deltas \ -s databases -d deltas \
$db $v1 $v2 2>&1 | tee $build_out $db $v1 $v2 2>&1 | tee $build_out
...@@ -85,13 +86,19 @@ typeset -ft getdb ...@@ -85,13 +86,19 @@ typeset -ft getdb
typeset -ft do_delta typeset -ft do_delta
typeset -ft cleandb typeset -ft cleandb
# A regular expression that should be avoided
avoid_re='mart'
# A regular expression that should be required
require_re='1[45]_'
version_re='[0-9][0-9]*_[0-9][0-9]*' version_re='[0-9][0-9]*_[0-9][0-9]*'
# Use ftp://ftp.ensembl.org/ls-lR.Z to figure out what files are # Use ftp://ftp.ensembl.org/ls-lR.Z to figure out what files are
# available # available
lynx -source ftp://ftp.ensembl.org/ls-lR.Z | \ lynx -source ftp://ftp.ensembl.org/ls-lR.Z | \
sed -n 's/^\.\(.*data\/mysql\)\/\(.*\)_\('"$version_re"'\):$/\1 \2 \3/p' | \ sed -n 's/^\.\(.*data\/mysql\)\/\(.*\)_\('"$version_re"'\):$/\1 \2 \3/p' | \
sort -k2 >ls-lR grep -v $avoid_re | grep $require_re | sort -k2 >ls-lR
while read path db ver; do while read path db ver; do
if [[ $db != $this_db ]]; then if [[ $db != $this_db ]]; then
......
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment