make build work again with older ICU

The recent addition of Hebrew line breaking rules apparently does not
work with genbrk from ICU 4.6. Maybe we should just require 4.x (where x
is the first version where it is implemented) as minimum ICU version?
IMHO it would be better than this insane (and fragile) filtering...

Change-Id: Ia2adad3621fa7a2f319a07fb569b1f1c1eb6db7c

3 files changed, 34 insertions, 11 deletions

diff --git a/config_host.mk.in b/config_host.mk.in
index da7b3b41f87f..11123316a7ff 100644
--- a/config_host.mk.in
+++ b/config_host.mk.in
@@ -219,6 +219,8 @@ export ICU_MAJOR=@ICU_MAJOR@
 export ICU_MICRO=@ICU_MICRO@
 export ICU_MINOR=@ICU_MINOR@
 export ICU_RECLASSIFIED_CLOSE_PARENTHESIS=@ICU_RECLASSIFIED_CLOSE_PARENTHESIS@
+export ICU_RECLASSIFIED_CONDITIONAL_JAPANESE_STARTER=@ICU_RECLASSIFIED_CONDITIONAL_JAPANESE_STARTER@
+export ICU_RECLASSIFIED_HEBREW_LETTER=@ICU_RECLASSIFIED_HEBREW_LETTER@
 export ICU_RECLASSIFIED_PREPEND_SET_EMPTY=@ICU_RECLASSIFIED_PREPEND_SET_EMPTY@
 export ILIB=@ILIB@
 @x_Cygwin@ export INCLUDE=
diff --git a/configure.in b/configure.in
index ddf47e8de005..3659e5c74833 100644
--- a/configure.in
+++ b/configure.in
@@ -7691,6 +7691,8 @@ ICU_MINOR=
 ICU_MICRO=
 ICU_RECLASSIFIED_CLOSE_PARENTHESIS="YES"
 ICU_RECLASSIFIED_PREPEND_SET_EMPTY="YES"
+ICU_RECLASSIFIED_CONDITIONAL_JAPANESE_STARTER="YES"
+ICU_RECLASSIFIED_HEBREW_LETTER="YES"
 AC_MSG_CHECKING([which icu to use])
 if test "$with_system_icu" = "yes"; then
     AC_MSG_RESULT([external])
@@ -7760,6 +7762,10 @@ You can use --with-system-icu-for-build=force to use it anyway.])
         else
             ICU_RECLASSIFIED_PREPEND_SET_EMPTY="NO"
         fi
+        if test "$ICU_MAJOR" -eq 4 -a "$ICU_MINOR" -le 6; then
+            ICU_RECLASSIFIED_CONDITIONAL_JAPANESE_STARTER="NO"
+            ICU_RECLASSIFIED_HEBREW_LETTER="NO"
+        fi
     fi
 
     libo_MINGW_CHECK_DLL([ICUDATA], [icudata][$ICU_MAJOR][$ICU_MINOR])
@@ -7779,6 +7785,8 @@ AC_SUBST(ICU_MAJOR)
 AC_SUBST(ICU_MINOR)
 AC_SUBST(ICU_MICRO)
 AC_SUBST(ICU_RECLASSIFIED_CLOSE_PARENTHESIS)
+AC_SUBST([ICU_RECLASSIFIED_CONDITIONAL_JAPANESE_STARTER])
+AC_SUBST([ICU_RECLASSIFIED_HEBREW_LETTER])
 AC_SUBST(ICU_RECLASSIFIED_PREPEND_SET_EMPTY)
 AC_SUBST([MINGW_ICUDATA_DLL])
 AC_SUBST([MINGW_ICUI18N_DLL])
diff --git a/i18npool/CustomTarget_breakiterator.mk b/i18npool/CustomTarget_breakiterator.mk
index 1c701847b409..3e0df9ee3b2f 100644
--- a/i18npool/CustomTarget_breakiterator.mk
+++ b/i18npool/CustomTarget_breakiterator.mk
@@ -98,18 +98,31 @@ $(i18npool_BIDIR)/%.brk : $(i18npool_BIDIR)/%.txt $(i18npool_GENBRKTARGET)
 		$(i18npool_GENBRK) -r $< -o $@ $(if $(findstring s,$(MAKEFLAGS)),> /dev/null))
 
 # fdo#31271 ")" reclassified in more recent Unicode Standards / ICU 4.4
-# Prepend set empty as of Unicode Version 6.1 / ICU 4.9, which bails out if used.
-# NOTE: strips every line with _word_ 'Prepend', including $Prepend
+# * Prepend set empty as of Unicode Version 6.1 / ICU 4.9, which bails out if used.
+#   NOTE: strips every line with _word_ 'Prepend', including $Prepend
+# * Conditional_Japanese_Starter does not exist in ICU 4.6, which bails out if used.
+# * Hebrew_Letter does not exist in ICU 4.6, which bails out if used.
+#   NOTE: I sincerely hope there is a better way to avoid problems than this abominable
+#   sed substitution...
 $(i18npool_BIDIR)/%.txt : \
 	$(SRCDIR)/i18npool/source/breakiterator/data/%.txt | $(i18npool_BIDIR)/.dir
-ifeq ($(ICU_RECLASSIFIED_CLOSE_PARENTHESIS),YES)
-ifeq ($(ICU_RECLASSIFIED_PREPEND_SET_EMPTY),YES)
-	sed "s#\[:LineBreak =  Close_Punctuation:\]#\[\[:LineBreak =  Close_Punctuation:\] \[:LineBreak = Close_Parenthesis:\]\]#" $< | sed "/Prepend/d" > $@
-else
-	sed "s#\[:LineBreak =  Close_Punctuation:\]#\[\[:LineBreak =  Close_Punctuation:\] \[:LineBreak = Close_Parenthesis:\]\]#" $< > $@
-endif
-else
-	cp $< $@
-endif
+	sed -e ': dummy' \
+		$(if $(filter YES,$(ICU_RECLASSIFIED_CLOSE_PARENTHESIS)),-e "s#\[:LineBreak =  Close_Punctuation:\]#\[& \[:LineBreak = Close_Parenthesis:\]\]#") \
+		$(if $(filter-out YES,$(ICU_RECLASSIFIED_CONDITIONAL_JAPANESE_STARTER)),\
+			-e '/\[:LineBreak =  Conditional_Japanese_Starter:\]/d' \
+			-e 's# $$CJ##' \
+		) \
+		$(if $(filter-out YES,$(ICU_RECLASSIFIED_HEBREW_LETTER)),\
+			-e '/\[:LineBreak =  Hebrew_Letter:\]/d' \
+			-e '/^$$HLcm =/d' \
+			-e '/^$$HLcm  $$NUcm;/d' \
+			-e '/^$$NUcm  $$HLcm;/d' \
+			-e '/^$$HL $$CM+;/d' \
+			-e 's# | $$HL\(cm\)\?##g' \
+			-e 's#$$HLcm ##g' \
+			-e 's# $$HL##g' \
+		) \
+		$(if $(filter YES,$(ICU_RECLASSIFIED_PREPEND_SET_EMPTY)),-e "/Prepend/d") \
+		$< > $@
 
 # vim: set noet sw=4 ts=4: