This is the mail archive of the archer@sourceware.org mailing list for the Archer project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [python] Implement gdb.search_memory.


El mar, 24-02-2009 a las 08:05 +0000, Phil Muldoon escribiÃ:
> Jan Kratochvil wrote:
> > .././gdb/python/python.c:393: error: passing argument 3 of âPyObject_AsReadBufferâ from incompatible pointer type
<snip>   
> > /usr/include/python2.5/abstract.h:
<snip>
> >      PyAPI_FUNC(int) PyObject_AsReadBuffer(PyObject *obj,
> >                                           const void **buffer,
> >                                           Py_ssize_t *buffer_len);

Strange, I have the same prototype in that header file here (I use
Debian lenny's python2.5-dev version 2.5.2-15, btw) and compilation is
fine here.

The problem seems to be that our compilers have their nitpicking levels
set to different values. Perhaps someone could tell me how I can make my
compiler behave like Fedora's? I'm using:

% gcc --version
gcc (Debian 4.3.3-3) 4.3.3

I have a set of commits which I was just about to push, but I'll hold
off for now to avoid triggering your additional alarms there.

> > Could you please check-in something backward/forward compatible?

This is not the issue here, it seems. But do we want to enforce
backward/forward compatibility for each pushed commit? We'd have to ask
everybody to compile their changes against the three different python
versions that configure currently detects before they push to the
branch.

> I needed to merge some code over to the archer-tromey-python branch
> today and wanted a clean compile before I did so. I checked in these
> fixes to the above warning, and one other.

Thanks!

> Please check the latter case,
> I'm not sure if setting the buffer to NULL will lead to correct results
> in a failure condition.

Well, it doesn't make things worse than what they already were. I
couldn't find a place which says what should happen to ptrptr in case of
an error. Perhaps we shouldn't touch it. I'll make the change.
-- 
[]'s
Thiago Jung Bauermann
IBM Linux Technology Center


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]