Mailman 3 January 2017 - pypy-dev

os_read & rposix.c_read
by John Zhang 18 Jan '17

18 Jan '17

Hi all, Anyone knows why pypy.module._file.readinto uses a locally defined ‘os_read’ rather than rpython.rlib.rposix.c_read? It seems that the functions are defined almost the same way, except one uses rffi.CCHARP the other uses rffi.VOIDP, and the rposix.c_read has a more complete ExternalCompilationInfo. Whilst this doesn’t pose any problem to C backend, it does cause a small trouble in my Mu VM backend, as the two external functions have deemed different even though they are the same, thus causing a name conflict. The fix is simple in either case: use c_read in _file.readinto, or I just add a small hack in my backend. But I’d like to know why this is the case in the first place, is there a special reason for it? Regards, John Zhang ------------------------------------------------------ John Zhang Research Assistant Programming Languages, Design & Implementation Division Computer Systems Group ANU College of Engineering & Computer Science 108 North Rd The Australian National University Acton ACT 2601 john.zhang(a)anu.edu.au<mailto:john.zhang@anu.edu.au>

2 1

Fwd: Re: Adding init/variables to W_Root
by Carl Friedrich Bolz 14 Jan '17

14 Jan '17

Oops, forgot to cc pypy-dev... Carl Friedrich -------- Original Message -------- From: Carl Friedrich Bolz <cfbolz(a)gmx.de> Sent: December 19, 2016 5:59:01 PM GMT+01:00 To: Frank Wang <frankw(a)mit.edu> Subject: Re: [pypy-dev] Adding init/variables to W_Root Hi Frank, The solution is to add an _attrs_ = ('__weakref__', 'rb_flags',) declaration to the body of W_Root. It's like __slots__, but only for RPython. Cheers, Carl Friedrich On December 19, 2016 5:42:12 PM GMT+01:00, Frank Wang <frankw(a)mit.edu> wrote: >Hi Armin, > >If I modify the line to __slots__ = ('__weakref__', 'rb_flags',), I get >the >following error: > >[translation:ERROR] AssertionError: ><pypy.interpreter.generator.GeneratorIterator object at >0x00000000073e87c8>: to run register_finalizer() untranslated, the >object >must not have __slots__ > >If I remove that line completely, I get the following error, > >[translation:ERROR] Exception: <class >'pypy.objspace.std.dictmultiobject.W_DictMultiObject'> has slots or >_attrs_, but not its base class >Processing block: > block@3 is a <class 'rpython.flowspace.flowcontext.SpamBlock'> > in (pypy.interpreter.mixedmodule:83)MixedModule.getdictvalue > containing the following operations: > v0 = getattr(space_0, ('finditem_str')) > v1 = getattr(self_0, ('w_dict')) > w_value_0 = simple_call(v0, v1, name_0) > v2 = getattr(self_0, ('lazy')) > v3 = bool(v2) > --end-- > >I wonder what error in your opinion is easier to get around. > >Thanks for the help! > >Frank > >On Mon, Dec 19, 2016 at 2:26 AM, Armin Rigo <armin.rigo(a)gmail.com> >wrote: > >> Hi Frank, >> >> On 19 December 2016 at 05:29, Frank Wang <frankw(a)mit.edu> wrote: >> > Attribute 'rb_flags' on <ClassDef 'pypy.interpreter. >> baseobjspace.W_Root'> >> > should be read-only. >> >> Ah, did you adapt or remove this line in class W_Root? >> >> __slots__ = ('__weakref__',) >> >> The annotator complains, I think, because 'rb_flags' is forbidden by >> this line (which is useful to avoid having random attributes >> accidentally move up to W_Root). >> >> >> A bientôt, >> >> Armin. >> > > >------------------------------------------------------------------------ > >_______________________________________________ >pypy-dev mailing list >pypy-dev(a)python.org >https://mail.python.org/mailman/listinfo/pypy-dev

4 10

RevDB
by Armin Rigo 14 Jan '17

14 Jan '17

Hi all, RevDB (https://bitbucket.org/pypy/revdb), the reverse debugger, was updated to 5.6.x. It contains a few RevDB-specific fixes as well. See the new tag RevDB-pypy2.7-v5.6.1 in the repository, or on direct download at https://bitbucket.org/pypy/pypy/downloads?tab=tags . Please contact me if you'd really like Windows support. A bientôt, Armin.

1 0

Somebody has already provided code objects in assemble_loop of x86/ assembler.py !
by Shubha Ramani 08 Jan '17

08 Jan '17

I just downloaded the pypy code. Please inspect the following log and search for 'code object'. This is the exact data which vtune needs ! How did this happen ? Was it there all along ? Did someone justadd this (perhaps Armin himself) ? The signature for 'assemble_loop' has not changed between the vtune branch (last April) and now. I'm confused. Also we need the same functionality in 'assemble_bridge'. Would it be hard to add loopname to'assemble_bridge' ? The reason I am assuming assemble_bridge needs the same info is becausein the vtune_branch the vtune API was called from assemble_loop and assemble_bridge. def assemble_loop(self, jd_id, unique_id, logger, loopname, inputargs, operations, looptoken, log): def assemble_bridge(self, faildescr, inputargs, operations, original_loop_token, log, logger):

2 2

jitlog.MARK_TRACE_OPT in do_compile_bridge and do_compile_loop of compile.py
by Shubha Ramani 06 Jan '17

06 Jan '17

I'm thinking that jitlog.MARK_TRACE_OPT is the only marker I need to keep track of as I'm dumping pycode (overloading execute_frame as vmprof does). When I see a MARK_TRACE_OPT, I just walk backward to find the closest pycode dump. This will give me the actual source code lines that was JITTED. Would this approach work ? Thanks, SHubha

1 0

how do we jump from pypy/interpreter to rpython ?
by Shubha Ramani 04 Jan '17

04 Jan '17

I see space.call_function() but how exactly do we end up in the rpython directory of code ?

1 0

who calls handle_builtin_call of jtransform.py ?
by Shubha Ramani 04 Jan '17

04 Jan '17

2 2

cannot use "id" in Rpython
by Shubha Ramani 04 Jan '17

04 Jan '17

What can I use then ? I need the address of the function ? [translation:ERROR] Exception: cannot use id() in RPython; see objectmodel.compute_xxx()Processing block: block@19 is a <class 'rpython.flowspace.flowcontext.SpamBlock'> in (pypy.module._vtune.interp_vtune:28)_get_full_name containing the following operations: buf_0 = simple_call((type RStringIO)) v943 = getattr(buf_0, ('write')) v944 = getattr(pycode_0, ('co_code')) v945 = simple_call(v943, v944) v946 = getattr(buf_0, ('getsize')) size_code_0 = simple_call(v946) v947 = getattr(pycode_0, ('co_name')) id_func_0 = id(v947) v948 = getattr(pycode_0, ('co_name')) v949 = simple_call((function _safe), v948) v950 = getattr(pycode_0, ('co_firstlineno')) v951 = getattr(pycode_0, ('co_filename')) v952 = simple_call((function _safe), v951) v953 = newtuple(v949, v950, v952, size_code_0, id_func_0) v954 = mod(('vtune:%s:%d:%s:%d:%d'), v953) --end--[translation] start debugger...> /opt/shubha_vtune_pypy/rpython/annotator/unaryop.py(188)id()-> raise Exception("cannot use id() in RPython; "(Pdb+)

3 8

How do you pass in data from pypy interpreter to backend ?
by Shubha Ramani 04 Jan '17

04 Jan '17

I've really been struggling with this. For instance pycode from pyframe. I would like to pass along a data structure containing pycode tidbits all the way to the background but cannot figure out a way. Any clues would be appreciated ! Shubha

1 2