From daf2b7fe5f1479cecbe0c7283b34e47324d10e1b Mon Sep 17 00:00:00 2001
From: Ralph Amissah <ralph@amissah.com>
Date: Sat, 16 Feb 2008 00:33:34 +0000
Subject: markup tag match refinement

---
 CHANGELOG                 | 15 ++++++++-------
 lib/sisu/v0/shared_sem.rb | 31 +++++++++++++++++++++++--------
 2 files changed, 31 insertions(+), 15 deletions(-)

diff --git a/CHANGELOG b/CHANGELOG
index c35925bb..d6676e53 100644
--- a/CHANGELOG
+++ b/CHANGELOG
@@ -9,13 +9,14 @@ Reverse Chronological:
 
 %% STABLE MANIFEST
 
-%% sisu_0.64.5.orig.tar.gz (2008-02-15:06/5)
-http://www.jus.uio.no/sisu/pkg/src/sisu_0.64.5.orig.tar.gz
-  sisu_0.64.5.orig.tar.gz
-  sisu_0.64.5-1.dsc
-  sisu_0.64.5-1.diff.gz
-
-  * semantic tag, refinement to regex matching
+%% sisu_0.65.0.orig.tar.gz (2008-02-16:06/6)
+http://www.jus.uio.no/sisu/pkg/src/sisu_0.65.0.orig.tar.gz
+  sisu_0.65.0.orig.tar.gz
+  sisu_0.65.0-1.dsc
+  sisu_0.65.0-1.diff.gz
+
+  * semantic tag, match refinement related, (consider a possible extension)
+    [version bump for tags]
 
 %% sisu_0.64.4.orig.tar.gz (2008-02-14:06/4)
 http://www.jus.uio.no/sisu/pkg/src/sisu_0.64.4.orig.tar.gz
diff --git a/lib/sisu/v0/shared_sem.rb b/lib/sisu/v0/shared_sem.rb
index 0c9e485c..f6303ee0 100644
--- a/lib/sisu/v0/shared_sem.rb
+++ b/lib/sisu/v0/shared_sem.rb
@@ -69,25 +69,36 @@ module SiSU_sem
         /^<:code>/
       end
       def each_csc
-        /\b[a-z]+[:;]\{|\}[:;][a-z]+\b/
+        /\b[a-z]+[:;]\{|\}[:;][a-z]+\b/m
+        #/\b(?:[a-z]+(?:[_.][a-z]+)+|[a-z]+)[:;]\{|\}[:;](?:[a-z]+(?:[_.][a-z]+)+|[a-z]+)\b/m
       end
       def each_c
-        /\b[a-z]+:\{|\}:[a-z]+\b/
+        /\b[a-z]+:\{|\}:[a-z]+\b/m
+        #/\b(?:[a-z]+(?:[_.][a-z]+)+|[a-z]+):\{|\}:(?:[a-z]+(?:[_.][a-z]+)+|[a-z]+)\b/m
       end
       def each_sc
-        /\b[a-z]+;\{|\};[a-z]+\b/
+        /\b[a-z]+;\{|\};[a-z]+\b/m
+        #/\b(?:[a-z]+(?:[_.][a-z]+)+|[a-z]+);\{|\};(?:[a-z]+(?:[_.][a-z]+)+|[a-z]+)\b/m
       end
       def pair_csc
-        /(([a-z]+)(?::\{(.+?)\}:\2|;\{(.+?)\};\2))/
+        /(([a-z]+)(?::\{(.+?)\}:\2|;\{(.+?)\};\2))/m
+        #/(([a-z]+(?:[_.][a-z]+)+|[a-z]+)(?::\{(.+?)\}:\2|;\{(.+?)\};\2))/m
       end
       def pair_c
-        /(([a-z]+):\{(.+?)\}:\2)/
+        /(([a-z]+):\{(.+?)\}:\2)/m
+        #/(([a-z]+(?:[_.][a-z]+)+|[a-z]+):\{(.+?)\}:\2)/m
       end
       def pair_sc
-        /(([a-z]+);\{.+?\};\2)/
+        /(([a-z]+);\{.+?\};\2)/m
+        #/(([a-z]+(?:[_.][a-z]+)+|[a-z]+);\{.+?\};\2)/m
       end
       def whole_csc_ae
-        /(([a-z]+)(?::\[(.+?)\]:\2|;\[(.+?)\];\2))/
+        /(([a-z]+)(?::\[(.+?)\]:\2|;\[(.+?)\];\2))/m
+        #/(([a-z]+(?:[_.][a-z]+)+|[a-z]+)(?::\[(.+?)\]:\2|;\[(.+?)\];\2))/m
+      end
+      def each_csc_ae
+        /\b[a-z]+[:;]\[|\][:;][a-z]+\b/m
+        #/\b(?:[a-z]+(?:[_.][a-z]+)+|[a-z]+)[:;]\[|\][:;](?:[a-z]+(?:[_.][a-z]+)+|[a-z]+)\b/m
       end
       self
     end
@@ -99,7 +110,11 @@ module SiSU_sem
         @para
       end
       def sem_marker_added_extra_parts
-        @para.gsub!(rgx.whole_csc_ae,'')
+        unless @para =~ rgx.exclude
+          @para.gsub!(rgx.whole_csc_ae,'')
+          if para =~rgx.each_csc_ae; puts "WARNING semantic taggin error: #@para"
+          end
+        end
         @para
       end
       def all
-- 
cgit v1.2.3