Fixed integer x = ***_toS(x) issue. Added postfix *_ro to non simple signals.

astand · astand · commit b5a687b23e25 · 2021-03-04T13:08:31.000+03:00
diff --git a/src/codegen/c-main-generator.cpp b/src/codegen/c-main-generator.cpp
@@ -456,19 +456,26 @@ void CiMainGenerator::WriteSigStructField(const SignalDescriptor_t& sig, bool bi
 
   fwriter->AppendLine("", 2);
 
-  if (sig.IsDoubleSig)
+  if (!sig.IsSimpleSig)
   {
     // this code only required be d-signals (floating point values based)
     // it placed additional signals to struct for conversion
     // to/from physical values. For non-simple and non-double signal
     // there is no necessity to create addition fields
     // @sigfloat_t must be typedefed by user (e.g. double / float)
+
+    // UPD: from this commit, all non-Simple signals has it's
+    // own 'shadow' (_phys) copies, the problem with intermediate type (not simpe and
+    // not double) is that the x = ***_toS(x) takes place in each Pack_* call
+    // the signals which are not changing from Pack_* to Pack_* will change its values (!)
     fwriter->AppendLine(StrPrint("#ifdef %s", fdesc->usesigfloat_def.c_str()));
 
-    fwriter->AppendLine(StrPrint("  sigfloat_t %s_phys;", sig.Name.c_str()));
+    if (sig.IsDoubleSig)
+      fwriter->AppendLine(StrPrint("  sigfloat_t %s;", sig.NameFloat.c_str()));
+    else
+      fwriter->AppendLine(StrPrint("  %s %s;", PrintType((int)sig.Type).c_str(), sig.NameFloat.c_str()));
 
     fwriter->AppendLine(StrPrint("#endif // %s", fdesc->usesigfloat_def.c_str()), 2);
-
   }
 }
 
@@ -484,23 +491,14 @@ void CiMainGenerator::WriteUnpackBody(const CiExpr_t* sgs)
     fwriter->AppendLine(StrPrint("  _m->%s = %s;", sname, expr.c_str()));
 
     // print sigfloat conversion
-    if (sgs->msg.Signals[num].IsDoubleSig)
+    if (!sgs->msg.Signals[num].IsSimpleSig)
     {
       fwriter->AppendLine(StrPrint("#ifdef %s", fdesc->usesigfloat_def.c_str()));
-      fwriter->AppendLine(StrPrint("  _m->%s_phys = (sigfloat_t)(%s_%s_fromS(_m->%s));", sname, fdesc->DRVNAME.c_str(), sname,
-          sname));
+      fwriter->AppendLine(StrPrint("  _m->%s = (sigfloat_t)(%s_%s_fromS(_m->%s));",
+          sgs->msg.Signals[num].NameFloat.c_str(), fdesc->DRVNAME.c_str(), sname, sname));
       fwriter->AppendLine(StrPrint("#endif // %s", fdesc->usesigfloat_def.c_str()), 2);
     }
 
-    else if (!sgs->msg.Signals[num].IsSimpleSig)
-    {
-      // print unpack conversion for non-simple and non-double signals
-      // for this case conversion fromS is performed to signal itself
-      // without (sigfloat_t) type casting
-      fwriter->AppendLine(StrPrint("#ifdef %s", fdesc->usesigfloat_def.c_str()));
-      fwriter->AppendLine(StrPrint("  _m->%s = (%s_%s_fromS(_m->%s));", sname, fdesc->DRVNAME.c_str(), sname, sname));
-      fwriter->AppendLine(StrPrint("#endif // %s", fdesc->usesigfloat_def.c_str()), 2);
-    }
     else if (num + 1 == sgs->to_signals.size())
     {
       // last signal without phys part, put \n manually
@@ -604,9 +602,9 @@ void CiMainGenerator::PrintPackCommonText(const std::string& arrtxt, const CiExp
         if (sgs->msg.Signals[n].IsDoubleSig)
         {
           // print toS from *_phys to original named sigint (integer duplicate of signal)
-          fwriter->AppendLine(StrPrint("  _m->%s = %s_%s_toS(_m->%s_phys);",
+          fwriter->AppendLine(StrPrint("  _m->%s = %s_%s_toS(_m->%s);",
               sgs->msg.Signals[n].Name.c_str(), fdesc->DRVNAME.c_str(),
-              sgs->msg.Signals[n].Name.c_str(), sgs->msg.Signals[n].Name.c_str()));
+              sgs->msg.Signals[n].Name.c_str(), sgs->msg.Signals[n].NameFloat.c_str()));
         }
         else
         {
diff --git a/src/parser/dbclineparser.cpp b/src/parser/dbclineparser.cpp
@@ -231,6 +231,15 @@ bool DbcLineParser::ParseSignalLine(SignalDescriptor_t* sig, const std::string&
     {
       sig->IsSimpleSig = true;
     }
+
+    if (!sig->IsSimpleSig)
+    {
+      // For this case the name of signal must be marked specially
+      // to pay attention that if SIGFLOAT is enabled, this signal
+      // must behave as ReadOnly (_ro)
+      sig->NameFloat = sig->Name + "_phys";
+      sig->Name += "_ro";
+    }
   }
 
   if (tailpart.size() == 3)
diff --git a/src/types/message.h b/src/types/message.h
@@ -37,7 +37,8 @@ typedef struct
 {
   // Signal name
   std::string Name;
-
+  // Signal float name
+  std::string NameFloat;
   // Unit
   std::string Unit;
 

Original file line number	Diff line number	Diff line change
`@@ -231,6 +231,15 @@ bool DbcLineParser::ParseSignalLine(SignalDescriptor_t* sig, const std::string&`
`231`	`231`	`{`
`232`	`232`	`sig->IsSimpleSig = true;`
`233`	`233`	`}`
	`234`	`+`
	`235`	`+ if (!sig->IsSimpleSig)`
	`236`	`+ {`
	`237`	`+ // For this case the name of signal must be marked specially`
	`238`	`+ // to pay attention that if SIGFLOAT is enabled, this signal`
	`239`	`+ // must behave as ReadOnly (_ro)`
	`240`	`+ sig->NameFloat = sig->Name + "_phys";`
	`241`	`+ sig->Name += "_ro";`
	`242`	`+ }`
`234`	`243`	`}`
`235`	`244`
`236`	`245`	`if (tailpart.size() == 3)`
Original file line number	Diff line number	Diff line change
`@@ -37,7 +37,8 @@ typedef struct`
`37`	`37`	`{`
`38`	`38`	`// Signal name`
`39`	`39`	`std::string Name;`
`40`		`-`
	`40`	`+ // Signal float name`
	`41`	`+ std::string NameFloat;`
`41`	`42`	`// Unit`
`42`	`43`	`std::string Unit;`
`43`	`44`