diff options
Diffstat (limited to 'doc/html/changes.html')
-rw-r--r-- | doc/html/changes.html | 37 |
1 files changed, 37 insertions, 0 deletions
diff --git a/doc/html/changes.html b/doc/html/changes.html new file mode 100644 index 0000000..ecdfa24 --- /dev/null +++ b/doc/html/changes.html @@ -0,0 +1,37 @@ +changes +---------- +done +*1) Should change Long8 to FileHandle * which is a caddr_t +*2) Should change open() arguments for C to be "r" "rw" "r+" standard +ANSI. +*3) Get rid of the IO::Write mode with IO::Create +*4) For f77, need to have modenames that are consistent with + the standard f77 IO names (maybe define in a common block). + +*6) Convert all ' to " in the C++ file +*7) Need to add IOisValid() subroutine to C and fortran libraries. + Actually instead, have the IOopen() methods return NULL if failure + and do the delete automatically internally!!!! + Keep IOisValid() around for the possible SocketIO option since the + validity can change in midstream... +*8) As well as sizeOf() and nBytes() and nElements() +*9) Need static members nBytes(int rank,int *dims,IO::DataType datatype); + nElements(int rank,int *dims,IO::DataType datatype); +*10) Need to have Unsigned integer types uInt8 uInt16 uInt32 uInt64 + + +------------ +Put off +5) Need to have full prototypes for c++ version with the meaning of the +returnvalue. +11) Change the seeking to have "current_dataset" and "ndatainfo". where +the ndatainfo is a counter for the number of times datainfo has been +called. So it acts as a trigger for incrementing the dataset pointer. +And that a data info read is not necessary to enable reading other +things like attributes. It also makes it easier to determine the +end-of-file condition. +----------- +Might not do +12) Change datatype names to IOFLOAT32 from FLOAT32 for C. + +// need a plan for 64-bit ints on the PC. |