get rid of null-termination and 'sort -z' since that option isn't on

new ISP's BSD sort.
This commit is contained in:
ehouse 2006-02-10 05:12:25 +00:00
parent 098adea254
commit 233479a959
2 changed files with 4 additions and 5 deletions

View file

@ -21,8 +21,7 @@ LANGFILTER_PRECLIP: tr -d '\r' |
LANGFILTER_POSTCLIP: | tr [a-z] [A-Z]
LANGFILTER_POSTCLIP: | grep '^[A-Z]*$'
LANGFILTER_POSTCLIP: | tr '\n' '\000'
LANGFILTER_POSTCLIP: | sort -z
LANGFILTER_POSTCLIP: | sort -u
# We can trust sort (above) to do the right thing since there's no
# high ascii. dict2dawg.pl is much faster if I can trust that its

View file

@ -235,10 +235,10 @@ dawg$(LANG)%.stamp: $(LANG)Main.dict.gz ../dict2dawg.pl table.bin ../Makefile.la
start=$$(echo $@ | sed -e 's/dawg$(LANG)\([0-9]*\)to[0-9]*.stamp/\1/'); \
end=$$(echo $@ | sed -e 's/dawg$(LANG)[0-9]*to\([0-9]*\).stamp/\1/'); \
echo $${start} and $$end; \
zcat $< | grep "^.\{$${start},$${end}\}$$" | tr '\n' '\0'| \
sort -z | ../dict2dawg.pl $(TABLE_ARG) table.bin -b 28000 \
zcat $< | grep "^.\{$${start},$${end}\}$$" | \
sort -u | ../dict2dawg.pl $(TABLE_ARG) table.bin -b 28000 \
-ob dawg$(LANG)$* \
-sn $(LANG)StartLoc.bin -k -term 0 -wc $(LANG)$*_wordcount.bin \
-sn $(LANG)StartLoc.bin -k -term 10 -wc $(LANG)$*_wordcount.bin \
$(FORCE_4) -ns $(LANG)$*_nodesize.bin
touch $@