Skip to content
Commits on Mar 30, 2012
  1. Merge branch 'dev'

    committed
  2. Added .gitignore exception for the top-level /bin folder which contai…

    committed
    …ns the built assemblies.
  3. Added .gitignore exception for the top-level /bin folder which contai…

    committed
    …ns the built assemblies.
  4. Merge pull request #75 from ahawker/dev

    committed
    Dev
  5. Added basic CalcMaxAllChannels, CalcNormMaxAllChannels and GetLibVers…

    committed
    …ion command integration tests.
  6. Fixed minor readme error.

    committed
Commits on Mar 19, 2012
  1. Fixed CalculateArrayLength<T> in LibsndfileArrayMarshaller. Unmanaged…

    committed
    …MemoryHandle.Size is already expanded by the size of the type, so the length of the array must divide and not multiply.
Commits on Mar 3, 2012
  1. Marked LibsndfileFormat with [Flags] attribute. Not necessary, but mo…

    committed
    …st of the time, the format will be used in bitwise-or computations.
  2. Implementation change to UnmangedMemoryHandle. We will now no longer …

    committed
    …attempt to free unmanaged memory unless we explicitly call Dispose, directly or indirectly via a 'using' statement.
Commits on Mar 2, 2012
  1. Refactored use of marshaller in ApiNativeWrapper. We cannot attach to…

    committed
    … the handle returned by functions as we cannot free this memory without incurring a COM exception.
  2. Changed both Open methods to return an empty handle if the open fails…

    committed
    … instead of throwing a general exception. If open returns a bad handle, users can lookup the error via an ErrorString function.
  3. Marked NativeApiWrapper and NativeCommandApiWrapper classes as intern…

    committed
    …al sealed so they are not exposed to any not-testing assemblies.
Commits on Feb 28, 2012
  1. Refactored LibsndfileArrayMarshaller to be slightly less confusing an…

    committed
    …d have slightly higher performance.
    
    ArrayMarshaller no longer returns a delegate to a function which can convert the memory to a T[], but performs the actual conversion and returns the results.
Commits on Feb 27, 2012
  1. Merge branch 'dev'

    committed
  2. Added SetInstrument command with comments and tests. Unit Tests fail …

    committed
    …in the same manner as others due to structs + moq.
  3. Updating GetBroadcastInfoTests to now account for the case of returni…

    committed
    …ng null when the command returns false (no struct to marshal back).
  4. Changed GetBroadcastInfo to act similar to GetLoopInfo. This function…

    committed
    …s are slightly different than others that return a struct, because the Command retval will return true/false where other struct-based commands will return 0 on pass, any non-zero on error.
    
    Since Command returning false for these two commands is completely valid, we need to have a third state for our structs, which is why we use the Nullable<T>. It could be a TryGetBroadcastInfo method, but that breaks the attempt at identical naming schema with Libsndfile.
  5. Added GetLoopInfo command with comments and tests. This function retu…

    committed
    …rns a nullable struct, so we arent forced to use an exception as flow control. We only throw an exception if the command returns an unexpected value (not 0,1) and return null on the false case.
  6. Added SetBroadcastInfo command with comments and tests. Multiple test…

    committed
    …s fail due to Moq and structs...same as some of the FormatX tests.
Something went wrong with that request. Please try again.