Permalink
Commits on Nov 4, 2010
  1. merged with trunk

    paende committed Nov 4, 2010
  2. some patches to fix BC backend

    paende committed Nov 4, 2010
Commits on Oct 28, 2010
  1. library patch

    paende committed Oct 28, 2010
  2. - merged with trunk

    - some small changes for LLVM 2.8
    paende committed Oct 28, 2010
Commits on Oct 20, 2010
  1. merged with trunk

    paende committed Oct 20, 2010
Commits on Sep 29, 2010
  1. merge with trunk release 1.1.0

    paende committed Sep 29, 2010
Commits on Aug 30, 2010
  1. - merge with trunk

    paende committed Aug 30, 2010
  2. preparing for merge

    paende committed Aug 30, 2010
Commits on Jun 23, 2010
  1. Making the LLVM backend working for variant 99.

    Pipeline is working, using OldIO. (No support for exceptions)
    Hello world compiles
    paende committed Jun 23, 2010
  2. - merge with trunk

    paende committed Jun 23, 2010
Commits on Jun 17, 2010
  1. small fixed caused by a merge problem

    paende committed Jun 17, 2010
Commits on Jun 16, 2010
  1. -merge with trunk

    paende committed Jun 16, 2010
Commits on Jun 10, 2010
  1. - solved major llvm isue by using a work around that does not trigger…

    … the gcroot bug in the llvm compiler
    paende committed Jun 10, 2010
  2. replaced some hardcoded llvmtypes

    paende committed Jun 10, 2010
  3. merge with trunk

    paende committed Jun 10, 2010
  4. preparing

    paende committed Jun 10, 2010
Commits on Jun 9, 2010
  1. - solved a bug causing a segfault on parameter derefencing in special…

    … cases.
    
    # first test round with nofib on variant 8 gives the following results:
    all working tests trigger serveral garbage collections and a lot of memory is cleaned after GC
    
    Always works:
    - TAK
    
    Works:
    - digits-of-e1, works with args up to 250
    - wheel-sieve2, works with args up to 4000
    
    with larger args a malloc panic is triggered (?)
    
    LLVM-Bug:
    the other test cases (primes, queens) trigger a bug in the LLVM compiler.
    the LLVM compiler crashes on an failed assertion caused by removing the GC intrinsics.
    according to llvm devs its a bug and they will try to fix it.
    paende committed Jun 9, 2010
Commits on Jun 8, 2010
  1. cleaned up some code

    paende committed Jun 8, 2010
Commits on Jun 4, 2010
  1. - fixed final (for now) bug regarding some local pointers not marked …

    …as root.
    
    - GC is working now for my test case! but needs some more testing with nofib suite.
    paende committed Jun 4, 2010
Commits on Jun 2, 2010
Commits on May 27, 2010
  1. added some debug info to find some last bugs

    paende committed May 27, 2010
Commits on May 26, 2010
  1. small update

    paende committed May 26, 2010
  2. parameters are stored on the stack and registered as gcroot to be abl…

    …e to update pointer during a GC
    paende committed May 26, 2010
  3. some progress

    paende committed May 26, 2010
Commits on May 20, 2010
  1. begin of globals to global array

    paende committed May 20, 2010
  2. replaced RP with a position in the globals array

    paende committed May 20, 2010
Commits on May 19, 2010
  1. start adding all globals to an global array

    paende committed May 19, 2010
  2. small fix

    paende committed May 19, 2010
  3. unstable stage, preparing to change globals

    paende committed May 19, 2010
Commits on May 17, 2010
  1. Fixed an error in propagating stackInfo between Silly and C.

    Now all nofib programs survive the new garbage collector (C backend).
    The new GC is always faster than Boehm, for some programs even much faster.
    More programs now survive the largest input (exp3_8 and gen-regexps).
    Only the primes program still crashes on the largest input.
    (Also, double is still not implemented, so "integrate", "paraffins" and "fib-double" do not compile).
    JeroenFokker committed May 17, 2010
Commits on May 14, 2010
  1. some progress on GC. still not fully functional.

    paende committed May 14, 2010
  2. - merged with trunk for GC changes silly

    - updated node descr with new info
    - automatically generated node descr
    - cleaned up some unused constructs
    paende committed May 14, 2010