search

collin80 / SavvyCAN

QT based cross platform canbus tool
MIT License
1.04k stars 288 forks source link

Error on parsing DBC signals #231

Closed cyber5tar86 closed 4 years ago

cyber5tar86 commented 4 years ago

Using your tool i've found a bug on this signal: SG_ ACT_PWR : 0|32@1- (1,0) [-3.4E+038|3.4E+038] "kW" PUMA

it's tagged as SIGNED type, instead it should be a float

collin80 commented 4 years ago

Well, I don't actually see the problem. The thing is, @1 = Little endian integer. For a float it should be either @2 or @3. Though, for a 32 bit single precision float it would be @2. So, it seems SavvyCAN is accepting the definition as presented. If you do actually want it to be a float then change that line to use @2 instead of @1

jackm commented 4 years ago

The thing is, @1 = Little endian integer. For a float it should be either @2 or @3. Though, for a 32 bit single precision float it would be @2. So, it seems SavvyCAN is accepting the definition as presented. If you do actually want it to be a float then change that line to use @2 instead of @1

Is this true about byte order values > 1? Where did you come across this information? As far as I am aware, the byte order field can only be 0 (big endian) or 1 (little endian).

Perhaps you meant the SIG_VALTYPE_ section where 1 => 32bit IEEE float and 2 => 64bit IEEE double? Both must be of signed value type (-, not +) since they are floats.