Source

extradoc / sprintinfo / hildesheim2-planning.txt

Carl Friedrich B… c4e297c 
holger krekel f77e295 




























Armin Rigo 62d0e0f 
Samuele Pedroni 9a45fb4 
holger krekel b6d1436 
holger krekel 9e5b8f5 
Carl Friedrich B… c4e297c 
holger krekel b6d1436 

holger krekel 9e5b8f5 




Carl Friedrich B… c4e297c 
holger krekel 9e5b8f5 
Carl Friedrich B… c4e297c 
holger krekel 9e5b8f5 



holger krekel b6d1436 
holger krekel 9e5b8f5 
holger krekel b6d1436 

holger krekel 9e5b8f5 







holger krekel b6d1436 
holger krekel 9e5b8f5 

holger krekel b6d1436 



Carl Friedrich B… c4e297c 
holger krekel b6d1436 
Carl Friedrich B… c4e297c 


holger krekel b6d1436 




holger krekel 9e5b8f5 

Carl Friedrich B… c4e297c 


PyPy Hildesheim2 sprint planning (25th-31st July) 
-------------------------------------------------------

overall time planning:: 
    25th full sprint day 
    26th full sprint day 
    27th afternoon break
    28th morning break 
    29th full sprint day   # (compliance test decision)
    30th full sprint day 
    31th more or less full sprint day 

    each day:  
        9-10 AM breakfast
        10:30 planning session 
        evening: dinner 

    Carl Friedrich Bolz   
    Richard Emslie        
    Armin Rigo            
    Holger Krekel         whole-time 
    Christian Tismer      whole-time 
    Samuele Pedroni       whole-time  


Current Main task: RTYPER issues
-----------------------------------

Left::

    - (Holger, Armin) add a compiling option to the C backend that allows to 
      generate a standalone binary (as opposed to a Python extension module) 

    - (Samuele, Carl) WRITE A TEST or TRACKER or TRY TO FIX ->
      likely we have a possibly deep exception handling problem:
      implicitely exceptions are attached to the last (high level) 
      operation in a block but when the RTyper converts this operation
      into possibly many low-level operations then it's unlikely 
      that we can preserve the implicit assumption that only 
      the last (low-level) operation possibly raises an exception. 

      solution: unknown at this point. 

      maybe describe a few scenarios regarding exception handling 
      describing the interactions/processing of exceptions in 
      flowobjspace/annotation/rtyping 

    - (Richard) TRACKER -> finish and polish and test our own posix module

    - (Nik with help of one of us)
      integrate Niklaus's applevel _sre module as a mixed module  
      (so that you can switch it in via usemodules=_sre). This may
      also ease moving gradually to an interplevel implementation. 
      
        pypy/module/_sre/
            full_sre_project/ # with manifest and so on 
            conftest.py # don't recurse into full_sre_project (unless the unittests
            are converted to py.test) 
            __init__.py: (appleveldefs -> full_sre_projec/...) 
            DONE Armin is going to ask niklaus about his plans and 
            suggest something along these lines. 

      OUTCOME: should go as a mixed module, but we don't have to keep the full
      subproject structure but just keep _sre and the tests in PyPy.  Nik will
      keep the stand-alone version only if there is interest.  (The sre-pypy
      branch should be ignored and removed.)

    - (all) investigate the possibility of translating the interactive part of PyPy
      or alternatively (probably preferably) running it completely at applevel


    - (Richard, Eric) more llvm stuff: overflow, missing basic
      operations, and exceptions.  Think about unifying the tests between genc 
      and llvm2.

    - GOOD PROGRESS WE GUESS rewriting the marshal module on interplevel (mostly
      done but there are difficult dependencies on the applevel _codecs.py 
      and some builtins (like frozenset) that live at applevel)