[PATCH 2 of 2] osutil: implement setprocname to set process title for some platforms

[Yuya Nishihara]

On Sun, 13 Nov 2016 14:56:59 +0000, Jun Wu wrote:
> Excerpts from Yuya Nishihara's message of 2016-11-13 23:18:10 +0900:
> > On Fri, 11 Nov 2016 21:14:37 +0000, Jun Wu wrote:
> > > # HG changeset patch
> > > # User Jun Wu <quark at fb.com>
> > > # Date 1478898677 0
> > > #      Fri Nov 11 21:11:17 2016 +0000
> > > # Node ID 4ccb6bcaf25a741d3a0d8abd6c674573ef76069a
> > > # Parent  98761d64eaaf67f3bdb99f3f80a57910e2624b78
> > > # Available At https://bitbucket.org/quark-zju/hg-draft 
> > > #              hg pull https://bitbucket.org/quark-zju/hg-draft  -r 4ccb6bcaf25a
> > > osutil: implement setprocname to set process title for some platforms
> > 
> > Seems fine. We can't guarantee argv[i] is null terminated for all i, but
> > that seems less dangerous since the last argv[i] always has '\0'.
> 
> You mean if some other code modifies "argv"? Seems fine - we may just think
> the buffer is smaller.

I wanted to say there's no risk of OOB memory access even if some
other codes do strsomething(argv[i]). They might see the data in argv[i + 1],
which isn't desirable but not hazardous. So this patch looks good.

> > > +#ifdef setprocname_method
> > > +static PyObject *setprocname(PyObject *self, PyObject *args)
> > > +{
> > > +    const char *name = NULL;
> > > +    if (!PyArg_ParseTuple(args, "s", &name))
> > > +        return NULL;
> > > +
> > > +#if 'a' == setprocname_method
> > > +    setproctitle("%s", name);
> > > +#elif 'b' == setprocname_method
> > > +    static char *argvstart = NULL;
> > > +    static size_t argvsize = 0;
> > 
> > Strictly speaking, these declarations should be moved to the block head. (But
> > almost all modern compilers but for MSVC would allow mixed declarations.)
> 
> Or we can put them in a "{}" scope - seems more readable.

Yep.