Permalink
Commits on Apr 1, 2010
  1. Removed IRLists.

    Michael Edgar committed Apr 1, 2010
  2. Switched the silly IRList class over to a case of normal sexps. Silly…

    … design decision.
    Michael Edgar committed Apr 1, 2010
  3. Added metaclass spec.

    Michael Edgar committed Apr 1, 2010
  4. Fixed easy creation of IRLists.

    Michael Edgar committed Apr 1, 2010
  5. Added easier way to initialize literal IRLists.

    Michael Edgar committed Apr 1, 2010
  6. Speccing up method definition IR

    Michael Edgar committed Apr 1, 2010
  7. Finished up specs for method send IR.

    Michael Edgar committed Apr 1, 2010
  8. Adding method call specs (for the IR. Obviously full decompilation is…

    … already specced for method calls).
    
    Exact method call IR specified in the ir.rb file.
    Michael Edgar committed Apr 1, 2010
  9. added a spec for blocks with arguments

    Michael Edgar committed Apr 1, 2010
  10. Added blocks to the specifications. Caught a whitespace difference, s…

    …o block output is a little cleaner now!
    Michael Edgar committed Apr 1, 2010
Commits on Mar 31, 2010
Commits on Mar 30, 2010
  1. No ISeqs being passed around for blocks on method sends. So, all IR f…

    …orms with iseqs have been eliminated.
    Michael Edgar committed Mar 30, 2010
  2. :block sexps don't take an entire ISeq now, either. Just the stringif…

    …ied argstring.
    Michael Edgar committed Mar 30, 2010
  3. :defmethod sexps no longer take an entire ISeq as an argument! Right …

    …now, it just preprocesses the argument string, which is all we were using it for during the stringification process anyway.
    Michael Edgar committed Mar 30, 2010
  4. Trying to remove inclusion of iseq objects in the internal representa…

    …tion.
    Michael Edgar committed Mar 30, 2010
  5. Added specifications for the IRList class.

    Michael Edgar committed Mar 30, 2010
  6. Moved iseq methods to general superclass.

    Michael Edgar committed Mar 30, 2010
  7. Renamed the decompile_x methods to to_ir_x, to better reflect their m…

    …eaning.
    Michael Edgar committed Mar 30, 2010
  8. For the first time in a long, long time: all specs pass again. And al…

    …most everything is switched to a stable internal representation.
    Michael Edgar committed Mar 30, 2010
  9. Method decompilation moved to being entirely internal representation,…

    … instead of the weird string-builder way of doing stuff.
    Michael Edgar committed Mar 30, 2010
Commits on Mar 29, 2010
  1. Switching over to a more appropriate way of decompiling - removing im…

    …proper use of strings during IR construction.
    Michael Edgar committed Mar 29, 2010
  2. Got a few failures. That's because I'm finally, *FINALLY* taking stri…

    …ngs out of the to_ir method's submethods. Man that was a horrible design choice. Oh well. I'm lazy at heart and it bit me in the ass.
    Michael Edgar committed Mar 29, 2010
  3. More stuff under spec.

    Michael Edgar committed Mar 29, 2010
  4. updates to infix mean smarter output, but I had to update a string sp…

    …ec to match the new, smarter behavior.
    Michael Edgar committed Mar 29, 2010
  5. more IR speccing.

    Michael Edgar committed Mar 29, 2010
  6. Speccing the IR code.

    Michael Edgar committed Mar 29, 2010
  7. Deep indentation for IRLists.

    Michael Edgar committed Mar 29, 2010
  8. Made sure to flatten the incoming array for IRLists.

    Michael Edgar committed Mar 29, 2010
Commits on Mar 21, 2010
  1. Ever-so-slowly moving away from string-based representation. Turns ou…

    …t it was embedded deeper than I thought. But by re-doing how classes are represented, two more specs pass than before, and it fits the mental model of an AST far better.
    Michael Edgar committed Mar 21, 2010
  2. Made all specs that required exact matches also include a check for i…

    …ndentation-independent matches. That way I can tell if it's failing only on indentation or on content as well.
    Michael Edgar committed Mar 21, 2010