perl's times() support is messy, esp on Win32 #14611
Comments
From rosiejjjj@outlook.comCreated by @bulk88I looked at Win32 Perl's implementation of times() today, which is below. DllExport int
win32_times(struct tms *timebuf)
{
FILETIME user;
FILETIME kernel;
FILETIME dummy;
if (GetProcessTimes(GetCurrentProcess(), &dummy, &dummy,
&kernel,&user)) {
timebuf->tms_utime = filetime_to_clock(&user);
timebuf->tms_stime = filetime_to_clock(&kernel);
timebuf->tms_cutime = 0;
timebuf->tms_cstime = 0;
} else {
/* That failed - e.g. Win95 fallback to clock() */
timebuf->tms_utime = process_time_so_far;
timebuf->tms_stime = 0;
timebuf->tms_cutime = 0;
timebuf->tms_cstime = 0;
}
return process_time_so_far;
}Problems, 1. it looks like there is Win95 code which was forgotten to be PP(pp_tms)
{
#ifdef HAS_TIMES
dSP;
struct tms timesbuf;
EXTEND(SP, 4);
(void)PerlProc_times(×buf);
mPUSHn(((NV)timesbuf.tms_utime)/(NV)PL_clocktick);
if (GIMME_V == G_ARRAY) {
mPUSHn(((NV)timesbuf.tms_stime)/(NV)PL_clocktick);
mPUSHn(((NV)timesbuf.tms_cutime)/(NV)PL_clocktick);
mPUSHn(((NV)timesbuf.tms_cstime)/(NV)PL_clocktick);
}
RETURN;
#else
# ifdef PERL_MICRO
dSP;
mPUSHn(0.0);
EXTEND(SP, 4);
if (GIMME_V == G_ARRAY) {
mPUSHn(0.0);
mPUSHn(0.0);
mPUSHn(0.0);
}
RETURN;
# else
DIE(aTHX_ "times not implemented");
# endif
#endif /* HAS_TIMES */
}the retval is unused, and potentially, strictly according to the book, I think MS CRT clock() is too complicated in implementation for the What about making PL_clocktick a constant (if possible on the platform) related links http://code.metager.de/source/xref/gnu/octave/gnulib-hg/lib/times.c Perl Info |
From @jkeenanIs there any particular aspect of the win32 implementation of times() for which you would like to propose a patch? Thank you very much. -- |
|
The RT System itself - Status changed from 'new' to 'open' |
toddr
added
the
Closable?
Migrated from rt.perl.org#124152 (status was 'open')
Searchable as RT124152$
The text was updated successfully, but these errors were encountered: