commit 8d420a7ea454a35308c4e8db917b5a5c626fcbd0
Author: Tanmai Khanna <khanna.tanmai@gmail.com>
Date:   Fri Aug 28 00:54:42 2020 +0530

    Change blank handling to use a blank queue | position for <b/> has no meaning in transfer now | Added tests

diff --git a/apertium/transfer.cc b/apertium/transfer.cc
index 8109073..ac3e24b 100644
--- a/apertium/transfer.cc
+++ b/apertium/transfer.cc
@@ -51,7 +51,6 @@ Transfer::destroy()
 
 Transfer::Transfer() :
 word(0),
-blank(0),
 lword(0),
 lblank(0),
 last_lword(0),
@@ -72,7 +71,6 @@ nwords(0)
   internal_null_flush = false;
   trace = false;
   trace_att = false;
-  emptyblank = "";
   in_lu = false;
   in_let_var = false;
 }
@@ -472,11 +470,15 @@ Transfer::evalString(xmlNode *element)
         return ti.getContent();
 
       case ti_b:
-        if(ti.getPos() >= 0 && checkIndex(element, ti.getPos(), lblank))
+        if(!blank_queue.empty())
         {
-          return !blank?"":*(blank[ti.getPos()]);
+          string retblank = blank_queue.front();
+          blank_queue.pop();
+          
+          return retblank;
         }
-        else {
+        else
+        {
           return " ";
         }
         break;
@@ -588,15 +590,7 @@ Transfer::evalString(xmlNode *element)
   }
   else if(!xmlStrcmp(element->name, (const xmlChar *) "b"))
   {
-    if(element->properties == NULL)
-    {
-      evalStringCache[element] = TransferInstr(ti_b, " ", -1);
-    }
-    else
-    {
-      int pos = atoi((const char *) element->properties->children->content) - 1;
-      evalStringCache[element] = TransferInstr(ti_b, "", pos);
-    }
+    evalStringCache[element] = TransferInstr(ti_b, " ", -1);
   }
   else if(!xmlStrcmp(element->name, (const xmlChar *) "get-case-from"))
   {
@@ -887,6 +881,12 @@ Transfer::processOut(xmlNode *localroot)
       }
     }
   }
+  
+  while(!blank_queue.empty()) //flush remaining blanks
+  {
+    fputws_unlocked(UtfConverter::fromUtf8(blank_queue.front()).c_str(), output);
+    blank_queue.pop();
+  }
 }
 
 string
@@ -1427,15 +1427,8 @@ Transfer::processCallMacro(xmlNode *localroot)
     myword = new TransferWord *[npar];
     std::fill(myword, myword+npar, (TransferWord *)(0));
   }
-  string **myblank = NULL;
-  if(npar > 0)
-  {
-    myblank = new string *[npar];
-    myblank[npar-1] = &emptyblank;
-  }
-
+  
   int idx = 0;
-  int lastpos = 0;
   for(xmlNode *i = localroot->children; npar && i != NULL; i = i->next)
   {
     if(i->type == XML_ELEMENT_NODE)
@@ -1446,23 +1439,12 @@ Transfer::processCallMacro(xmlNode *localroot)
       }
       int pos = atoi((const char *) i->properties->children->content)-1;
       myword[idx] = word[pos];
-      if(idx-1 >= 0)
-      {
-        if(lastpos + 1 > lblank) { // if a 1-pattern rule calls macro with same
-          noblank = "";            // param twice the blank array will be empty
-          myblank[idx-1] = &noblank;
-        }
-        else {
-          myblank[idx-1] = blank[lastpos];
-        }
-      }
+
       idx++;
-      lastpos = pos;
     }
   }
 
   swap(myword, word);
-  swap(myblank, blank);
   swap(npar, lword);
 
   for(xmlNode *i = macro->children; i != NULL; i = i->next)
@@ -1474,11 +1456,9 @@ Transfer::processCallMacro(xmlNode *localroot)
   }
 
   swap(myword, word);
-  swap(myblank, blank);
   swap(npar, lword);
 
   delete[] myword;
-  delete[] myblank;
 }
 
 int
@@ -2617,18 +2597,19 @@ Transfer::applyRule()
       lword = limit;
       if(limit != 1)
       {
-        blank = new string *[limit - 1];
         lblank = limit - 1;
       }
       else
       {
-        blank = NULL;
         lblank = 0;
       }
     }
     else
     {
-        blank[i-1] = new string(UtfConverter::toUtf8(*tmpblank[i-1]));
+      if(int(blank_queue.size()) + 1 < last_lword)
+      {
+        blank_queue.push(string(UtfConverter::toUtf8(*tmpblank[i-1])));
+      }
     }
 
     pair<wstring, int> tr;
@@ -2767,17 +2748,7 @@ Transfer::applyRule()
     }
     delete[] word;
   }
-  if(blank)
-  {
-    for(unsigned int i = 0; i != limit - 1; i++)
-    {
-      delete blank[i];
-      blank[i] = 0;
-    }
-    delete[] blank;
-  }
   word = NULL;
-  blank = NULL;
   tmpword.clear();
   tmpblank.clear();
   ms.init(me->getInitial());
diff --git a/apertium/transfer.h b/apertium/transfer.h
index 6a5b06a..68409a7 100644
--- a/apertium/transfer.h
+++ b/apertium/transfer.h
@@ -34,6 +34,7 @@
 #include <map>
 #include <set>
 #include <vector>
+#include <queue>
 
 using namespace std;
 
@@ -55,10 +56,9 @@ private:
   xmlDoc *doc;
   xmlNode *root_element;
   TransferWord **word;
-  string **blank;
+  queue <string> blank_queue;
   int lword, lblank;
   int last_lword;
-  string noblank = "";
   Buffer<TransferToken> input_buffer;
   vector<wstring *> tmpword;
   vector<wstring *> tmpblank;
diff --git a/tests/data/apertium-nno-nob.nob-nno.t1x b/tests/data/apertium-nno-nob.nob-nno.t1x
index 1de6011..e84d857 100644
--- a/tests/data/apertium-nno-nob.nob-nno.t1x
+++ b/tests/data/apertium-nno-nob.nob-nno.t1x
@@ -12,6 +12,12 @@
     <def-cat n="nacr">
       <cat-item tags="n.acr"/>
     </def-cat>
+    <def-cat n="det">
+      <cat-item tags="det"/>
+    </def-cat>
+    <def-cat n="adj">
+      <cat-item tags="adj"/>
+    </def-cat>
   </section-def-cats>
   <section-def-attrs>
     <def-attr n="a_nom">
@@ -99,13 +105,12 @@
           <chunk name="prn" case="caseFirstWord">
             <tags><tag><clip pos="3" side="tl" part="tags"/></tag></tags>
             <lu><clip pos="3" side="tl" part="whole"/></lu>
-            <b/>
             <lu>
               <clip pos="2" side="tl" part="lem"/>
               <lit-tag v="nacr.sg.m"/>
             </lu>
           </chunk>
-          <b pos="1"/>
+          <b/>
           <chunk name="det" case="caseFirstWord">
             <tags><tag><clip pos="1" side="tl" part="tags"/></tag></tags>
             <mlu>
@@ -113,7 +118,7 @@
               <lu><clip pos="3" side="tl" part="whole"/></lu>
             </mlu>
           </chunk>
-          <b pos="2"/>
+          <b/>
         </out>
       </action>
     </rule>
@@ -144,5 +149,62 @@
       </action>
     </rule>
 
+    <rule comment="superblankrule1">
+      <pattern>
+        <pattern-item n="det"/>
+        <pattern-item n="adj"/>
+        <pattern-item n="nacr"/>
+      </pattern>
+      <action>
+        <out>
+          <chunk name ="test1" case="caseFirstWord">
+            <tags><tag><clip pos="2" side="tl" part="tags"/></tag></tags>
+            <lu><clip pos="1" side="tl" part="whole"/></lu>
+            <lu><clip pos="2" side="tl" part="whole"/></lu>
+            <lu><clip pos="3" side="tl" part="whole"/></lu>
+          </chunk>
+        </out>
+      </action>
+    </rule>
+
+    <rule comment="superblankrule2">
+      <pattern>
+        <pattern-item n="adj"/>
+        <pattern-item n="det"/>
+        <pattern-item n="nacr"/>
+      </pattern>
+      <action>
+        <out>
+          <chunk name ="test1" case="caseFirstWord">
+            <tags><tag><clip pos="2" side="tl" part="tags"/></tag></tags>
+            <lu><clip pos="2" side="tl" part="whole"/></lu>
+            <b/>
+            <lu><clip pos="3" side="tl" part="whole"/></lu>
+          </chunk>
+        </out>
+      </action>
+    </rule>
+
+    <rule comment="superblankrule3">
+      <pattern>
+        <pattern-item n="nacr"/>
+        <pattern-item n="det"/>
+      </pattern>
+      <action>
+        <out>
+          <chunk name ="test1" case="caseFirstWord">
+            <tags><tag><clip pos="2" side="tl" part="tags"/></tag></tags>
+            <lu><clip pos="1" side="tl" part="whole"/></lu>
+            <b/>
+            <lu><clip pos="2" side="tl" part="whole"/></lu>
+            <b/>
+            <lu><clip pos="1" side="tl" part="whole"/></lu>
+            <b/>
+            <lu><clip pos="2" side="tl" part="whole"/></lu>
+          </chunk>
+        </out>
+      </action>
+    </rule>
+
   </section-rules>
 </transfer>
diff --git a/tests/transfer/__init__.py b/tests/transfer/__init__.py
index 805b0f1..d92b58d 100644
--- a/tests/transfer/__init__.py
+++ b/tests/transfer/__init__.py
@@ -97,12 +97,23 @@ class SlLemqTest(TransferTest):
 
 class WordboundBlankTest(TransferTest):
     inputs =          ["[blank1] [[t:s:123456]]^worda<n><acr>/wordta<n><acr>$ ;[blank2]; [[t:b:xyz123; t:l:xyz347]]^wordb<n><acr>/wordtb<n><acr>$ [blank3];  [[t:i:abc123; t:s:abc123]]^hun<prn><f>/ho<prn><f>$"]
-    expectedOutputs = ["[blank1] ^prn<prn><f>{[[t:i:abc123; t:s:abc123]]^ho<prn><f>$ [[t:b:xyz123; t:l:xyz347]]^wordtb<nacr><sg><m>$}$ ;[blank2]; ^det<n><acr>{[[t:s:123456; t:i:abc123; t:s:abc123]]^wordta<n><acr>+ho<prn><f>$}$ [blank3];  "]
+    expectedOutputs = ["[blank1] ^prn<prn><f>{[[t:i:abc123; t:s:abc123]]^ho<prn><f>$[[t:b:xyz123; t:l:xyz347]]^wordtb<nacr><sg><m>$}$ ;[blank2]; ^det<n><acr>{[[t:s:123456; t:i:abc123; t:s:abc123]]^wordta<n><acr>+ho<prn><f>$}$ [blank3];  "]
 
 class SingleLUWordboundBlankTest(TransferTest):
     inputs =          ["[blank1] [[t:s:123456]]^worda<n><acr>/wordta<n><acr>$ ;[blank2]; [[t:b:xyz123; t:l:xyz347]]^wordb<n><acr>/wordtb<n><acr>$ [blank3]; "]
     expectedOutputs = ["[blank1] ^nacr<n><acr>{[[t:s:123456]]^test<n><sg>$ [[t:s:123456]]^wordta<n><acr>$}$ ^nacr2<n><acr>{[[t:s:123456]]^testlem<n><pl>$ [[t:s:123456]]^wordta<n><acr>+postp<adj><sg>$}$ ;[blank2]; ^nacr<n><acr>{[[t:b:xyz123; t:l:xyz347]]^test<n><sg>$ [[t:b:xyz123; t:l:xyz347]]^wordtb<n><acr>$}$ ^nacr2<n><acr>{[[t:b:xyz123; t:l:xyz347]]^testlem<n><pl>$ [[t:b:xyz123; t:l:xyz347]]^wordtb<n><acr>+postp<adj><sg>$}$ [blank3]; "]
 
+class SuperblankTest(TransferTest):
+    inputs =          [ "[blank1] ^worda<det>/wordta<det>$ ;[blank2]; ^wordb<adj>/wordtb<adj>$ [blank3];  ^hun<n><acr>/ho<n><acr>$ [blank4]  ",          #Rule: superblankrule1 -> No <b/> in rule output, should flush all blanks after rule output
+                        "[blank1] ^worda<detn>/wordta<detn>$ ;[blank2]; ^wordb<adjn>/wordtb<adjn>$ [blank3];  ^hun<nn><acr>/ho<nn><acr>$ [blank4]  ",   #No rule matches, should print all blanks as is
+                        "[blank1] ^wordb<adj>/wordtb<adj>$ ;[blank2]; ^worda<det>/wordta<det>$ [blank3];  ^hun<n><acr>/ho<n><acr>$ [blank4]  ",         #Rule: superblankrule2 -> One <b/> in rule output, should print one and flush the rest
+                        "[blank1] ^hun<n><acr>/ho<n><acr>$ ;[blank2]; ^worda<det>/wordta<det>$ [blank3];   [blank4]  ",                                 #Rule: superblankrule3 -> Input rule has 1 blank, output has 3, should print input blank for the first <b/> and just spaces for the rest
+                        "[blank1] ^hun<n><acr>/ho<n><acr>$ ;[blank2]; ^worda<det>/wordta<det>$ [blank3]; ^wordb<adj>/wordtb<adj>$ ;[blank4]; ^worda<det>/wordta<det>$ [blank5];  ^hun<n><acr>/ho<n><acr>$ [blank6]  "] #Multiple rule matches
+    expectedOutputs = [ "[blank1] ^test1<adj>{^wordta<det>$^wordtb<adj>$^ho<n><acr>$}$ ;[blank2];  [blank3];   [blank4]  ",
+                        "[blank1] ^default<default>{^wordta<detn>$}$ ;[blank2]; ^default<default>{^wordtb<adjn>$}$ [blank3];  ^default<default>{^ho<nn><acr>$}$ [blank4]  ",
+                        "[blank1] ^test1<det>{^wordta<det>$ ;[blank2]; ^ho<n><acr>$}$ [blank3];   [blank4]  ",
+                        "[blank1] ^test1<det>{^ho<n><acr>$ ;[blank2]; ^wordta<det>$ ^ho<n><acr>$ ^wordta<det>$}$ [blank3];   [blank4]  ",
+                        "[blank1] ^test1<det>{^ho<n><acr>$ ;[blank2]; ^wordta<det>$ ^ho<n><acr>$ ^wordta<det>$}$ [blank3]; ^test1<det>{^wordta<det>$ ;[blank4]; ^ho<n><acr>$}$ [blank5];   [blank6]  "]
 
 class BincompatTest(BasicTransferTest):
     bindata = "data/bincompat.t1x.bin"