SDTS++ Pending Development Issues

This enumerates implementation issues that have to be resolved. I've broken these issues down by sub-system and into design, interface, and implementation categories. However, I've consolidated items common to both sub-systems into a separate category. Please note that I've intentionally left out the builder sub-system.

(And, yeah, I dropped '8211' from the I/O sub-class names I used because I got tired of typing it all the time. Cope.)

Common Stuff

Design

Adding robust error handling.

Interface

Stupid STL operator{<,>,==,!=}() needed by VC++.

Implementation

None pending.

Container Sub-System

Design

None pending.

Interface

sc_Subfield doesn't support floats yet

Implementation

None pending.

I/O Sub-System

Design

Do we ignore setting/getting field/unit terminators?
Need to change the sio_Reader and sio_ReaderIterator design. [MAC. Done]
There is currently NO random access iterator

Interface

Need to change the lexer/parser globals to head off any possible namespace conflicts.[MAC. Working this now.]
sio_MakeFieldFormat() XXX map<> comments may now be ok.
sio_FieldFormat uses operator{>,<}() to compare mnemonic comparisons. This is used for find() implementation. Should we remove this and go with find_if() instead? [MAC. This is ok.]
sio_FieldFormat has vector in emumerator -- ok in VC++, probably ok elsewhere [MAC. This is ok.]
sio_MakeFieldFormat() is probably fine in sio_8211FieldFormat.h[MAC. This is ok.]
makeFixedSubfield() reconciling binary length units with units used everywhere else [MAC. Fixed by adjusting lengths for binary subfields in subfield formats.]
sio_DDRLeader needs to be manually created from an sio_Leader just like sio_DDRField is created from a sio_Field. Currently this relies on RTTI.
DONE sio_Converter hierarchy needs to support remaining types

Implementation

sio_DDRField implementation needs to be more robust
in sio_FieldFormat yywrap() should be moved to FormatLexer.l[MAC. Done.]
union in sio_SubfieldFormat implementation fine [MAC. Not an issue.]
sio_Converter robustness needs enhanced/fixed
in sio_Reader, _fillScField() kludge -- it may be ok to leave for now. It works currently and may actually be the optimal method.[MAC. It's ok.]
in sio_Reader need to add appropriate commentary for iterator = insert< container.end(), thing() ) [MAC. Added comments.]
in sio_Reader need to add additional check to see if dropped leaders are in effect; currently checking for 'R' each time [MAC. Added bool check.]
in sio_Reader commentary on ios::peek() may be removed; this is ok [MAC. This is ok.]
in sio_Reader I think it's safe to say that it needs a DDR member [MAC. Well, DUH.]
in sio_Utils need to investigate better conversion implementations