Merge branch 'upstream' into debian/sid

1 files changed, 29 insertions, 11 deletions

diff --git a/lib/sisu/v0/shared_sem.rb b/lib/sisu/v0/shared_sem.rb
index da989999..f6303ee0 100644
--- a/lib/sisu/v0/shared_sem.rb
+++ b/lib/sisu/v0/shared_sem.rb
@@ -65,38 +65,56 @@ module SiSU_sem
       @para=para
     end
     def rgx
+      def exclude
+        /^<:code>/
+      end
       def each_csc
-        /[a-z]+[:;]\{|\}[:;][a-z]+/
+        /\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
-        /[a-z]+:\{|\}:[a-z]+/
+        /\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
-        /[a-z]+;\{|\};[a-z]+/
+        /\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)/
-        #/(([a-z]+)([:;])\{(.+?)\}\3\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
     def rm
       def sem_marker_parts
-        @para.gsub!(rgx.each_csc,'')
+        unless @para =~ rgx.exclude
+          @para.gsub!(rgx.each_csc,'')
+        end
         @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