t.mettler/giesCANne
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fixes and addition of code to be able to load old graph definitions. It's kind of broken when loading old standard graph

Browse Source 
definitions though. Otherwise graphing seems pretty good.
This commit is contained in:
Collin Kidder
2016-05-11 21:46:14 -04:00
parent b0902c8f50
commit 4fd437a44a
2 changed files with 116 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
graphingwindow.cpp
+115 -2
View File
@@ -748,7 +748,7 @@ void GraphingWindow::loadDefinitions()
QList<QByteArray> tokens = line.split(',');
if (tokens[0] == "X")
if (tokens[0] == "X") //newest format based around signals
{
gp.ID = tokens[1].toInt(NULL, 16);
gp.mask = tokens[2].toULongLong(NULL, 16);
@@ -769,6 +769,119 @@ void GraphingWindow::loadDefinitions()
gp.graphName = QString();
createGraph(gp, true);
}
else //one of the two older formats then
{
gp.ID = tokens[0].toInt(NULL, 16);
if (tokens[1] == "S") //old signal based graph definition
{
//tokens[2] is the signal name. Need to use the message ID and this name to look it up
DBC_MESSAGE *msg = dbcHandler->getFileByIdx(0)->messageHandler->findMsgByID(gp.ID);
if (msg != NULL)
{
DBC_SIGNAL *sig = msg->sigHandler->findSignalByName(tokens[2]);
if (sig)
{
gp.mask = 0xFFFFFFFF;
gp.bias = sig->bias;
gp.color.setRed(tokens[3].toInt());
gp.color.setGreen(tokens[4].toInt());
gp.color.setBlue(tokens[5].toInt());
gp.graphName = sig->name;
gp.intelFormat = sig->intelByteOrder;
if (sig->valType == SIGNED_INT) gp.isSigned = true;
else gp.isSigned = false;
gp.numBits = sig->signalSize;
gp.scale = sig->factor;
gp.startBit = sig->startBit;
gp.stride = 1;
createGraph(gp, true);
}
}
}
else //old standard graph definition
{
//hard part - this all changed drastically
//the difference between intel and motorola format is whether
//start is larger than end byte or not.
uint64_t oldMask = tokens[1].toULongLong(NULL, 16);
int oldStart = tokens[2].toInt();
int oldEnd = tokens[3].toInt();
if (oldEnd > oldStart) //motorola / big endian - hell...
{
gp.intelFormat = false;
//for now just naively use the entire bytes called for.
gp.startBit = 8 * oldStart + 7;
gp.numBits = (oldEnd - oldStart + 1) * 8;
}
else if (oldStart > oldEnd) //intel / little endian - easiest of multi-byte types
{
//have to find both ends. start bit is somewhere in oldEnd and last bit is somewhere in
//oldStart.
gp.intelFormat = true;
//start by setting a safe default if nothing else pans out.
gp.startBit = 8 * oldEnd;
int numBytes = oldStart - oldEnd + 1;
gp.numBits = numBytes * 8;
for (int b = 0; b < 8; b++)
{
if (oldMask & (1 << b))
{
gp.startBit = (8 * oldEnd) + b;
break;
}
}
for (int c = 7; c >= 0; c--)
{
if ( oldMask & (1<<(((numBytes - 1) * 8) + c)) )
{
gp.numBits -= (7-c);
break;
}
}
}
else //within a single byte - easier than the above two by a bit - always use intel format for this
{
gp.intelFormat = true;
oldMask = oldMask & 0xFF; //only this part matters
//for intel format we give startbit as the lowest bit number in the signal
//we can find that by going backward from bit 0 to 7 and picking the first bit that is 1.
//that's our start bit (+ 8*oldStart)
//set default first in case the rest falls through
gp.startBit = 8 * oldStart;
gp.numBits = 8;
for (int b = 0; b < 8; b++)
{
if (oldMask & (1 << b))
{
gp.startBit = 8 * oldStart + b;
gp.numBits = 8 - b;
break;
}
}
}
//the rest is easy stuff
if (tokens[4] == "Y") gp.isSigned = true;
else gp.isSigned = false;
gp.bias = tokens[5].toFloat();
gp.scale = tokens[6].toFloat();
gp.stride = tokens[7].toInt();
gp.color.setRed(tokens[8].toInt());
gp.color.setGreen(tokens[9].toInt());
gp.color.setBlue(tokens[10].toInt());
if (tokens.length() > 11)
gp.graphName = tokens[11];
else
gp.graphName = QString();
createGraph(gp, true);
}
}
}
}
inFile->close();
@@ -808,7 +921,7 @@ void GraphingWindow::addNewGraph()
void GraphingWindow::appendToGraph(GraphParams &params, CANFrame &frame)
{
int64_t tempVal; //64 bit temp value.
tempVal = Utility::processIntegerSignal(frame.data, params.startBit, params.numBits, params.intelFormat, params.isSigned) & params.mask;
tempVal = Utility::processIntegerSignal(frame.data, params.startBit, params.numBits, params.intelFormat, params.isSigned); //& params.mask;
if (secondsMode)
{
params.x.append((double)(frame.timestamp) / 1000000.0 - params.xbias);