No commit message

Better EH verbose messages for zipr, stack stamping+EH support, a new cfar config (NOG+SS vrs NogOF+SS), as well as an IR refinement technique to do recursive-decent disassembly when a landing pad is not part of a function.

Added emt support for prefork+bilr+p1 (sketchy still), K6+NOG (seems stable with all xforms enabled).  I also did some peasoup/cfar fixes to support determunistic ID assignment for IRDB objects -- this was needed for marshaling because the EMT plugin uses ID matching to find identical objects (stack, scoop, instruction, etc.)  Next, I fixed bugs in p1transform where the p1.map wasn't getting proper IDs. Lastly, I added an option to zipr/elfwrite to disable bss-optimization, this was necessary for marshaling as that bss segments stay named/mapped to the binary and cannot be combined with other segments.

No commit message

Updates to eh-write for optimizations, named callbacks.exe step to resolve_callbacks, stratafier bug fix, nos attributes and new ada tests.

performance improvements in elfwriter and log size reduction in zipr.cpp.  note:  please implement a 'very verbose' option to zipr if you need these full logs, because they are way too big.

These comments and additional output give clarity 
to the case where two free ranges abut and need to 
be merged together.

No commit message

had elfwriter use .text instead of .zipr_text for first .text segment.  gdb hates not having a .text segment.

Updates to elfwrite to handle more segments, new cfar configs typed out, and scons added to get-packages.

many updates esp. regarding interfacing NOL/NOH with ps-analyze, set-interpreter assertions being hit, writing an elf header that's bigger than 4k, etc.

No commit message

Remove DetermineWorstCaseInsnSize as public interface and force callers to specify account_for_fallthrough when calling other WC*S functions.