The London Perl and Raku Workshop takes place on 26th Oct 2024. If your company depends on Perl, please consider sponsoring and/or attending.
Revision history for Perl extension Proc::ProcessTable.

0.01  Mon Jun  8 11:37:47 1998
	- original version; created by h2xs 1.18

0.03  Sun Jul 26 09:07:50 1998
	- added argument to va_end to make linux-ppc work

0.04  Fri Aug 14 22:51:31 1998
        - preliminary pre-alpha solaris support

0.05  Wed Dec  9 18:47:39 1998
	- fixed minor compilation problems for solaris. Changed rss to
	bytes on linux for consistency with solaris; also added "time"
	and "ctime" fields to linux for consistency. Updated
	documentation in Process.pm. Folded in AIX port by David
	Paquet <david-paquet@usa.net>.

0.06  Wed Dec  9 23:42:05 1998
	- added HPUX port by Mike Romberg <romberg@fsl.noaa.gov>.

0.07  Wed Jan  6 14:57:04 1999
	- added FreeBSD port by Slaven Rezic <eserte@cs.tu-berlin.de>.

0.08  Thu Feb  4 00:35:53 1999 
	- Fixed bug in solaris hints file that was looking for a
	version of solaris returned by "uname -r" as > 2.5, when
	"uname -r" returns the "SunOS" version of 5.5. Fixed bug in
	linux module in which only the first field of the command line
	was getting passed in cmndline attribute.

0.09   Fri Feb 19 23:23:53 1999 
	- Fixed bug in Linux.c, fopen() called without corresponding
	fclose. Added empty DESTROY subs to ProcessTable.pm and
	Process.pm to fix occasional problems with mod_perl. Changed
	all instances of sv_undef in ProcessTable.xs to PL_sv_undef to
	comply with new standard.

0.10	Sat Feb 20 11:29:35 1999
	- Added code to ProcessTable.xs to check perl patchlevel
	version and use either PL_sv_undef or sv_undef since lots of
	people are still using older versions of perl.

0.11	Sat Mar  6 19:04:39 1999
	- Fixed code that checked perl patchlevel version to check
	directly for definition of PL_sv_undef and define it as
	sv_undef if it's not there, since it's not clear what
	patchlevel we need to check for. Also fixed embarrassing and
	unnecessary call to close() a string in Solaris.c

0.12	Sun Mar 21 10:53:56 1999
	- Hopefully the final fix for the sv_undef buglet.

0.14	Fri May 14 06:30:15 1999 
	- Added port for IRIX from W. Phillip Moore <wpm@ms.com> in
	0.13, but forgot to add files to MANIFEST and edit Changes
	(Duh). W. Phillip Moore's port also included a minor patch for
	Solaris to get the page size from the system. 

0.15	Tue Jun  1 21:33:16 1999
	- Added patch from Peter ? <hooft@natlab.research.philips.com>
	for IRIX version differences. There seems to be a lot of IRIX
	out there.

0.16	Mon Aug  9 22:59:46 1999
	- Added nanosecond Solaris patch from Rolf Petter Halle
	<rph@nextel.no>. Added port for dec_osf from Bernhard
	Schmalhofer <Bernhard.Schmalhofer@fazi.de>, which included new 
	version of test code as well. Added bsdi port from Sean Eskins
	<sean@gilasoft.com>. 


0.17	Thu Aug 19 10:57:16 1999
	- Lost connection to ISP during upload to PAUSE, and PAUSE
	won't let me reload it so I have to increment the version
	number. Kind of annoying.

0.18	Thu Aug 19 10:57:16 1999
	- Fixed bug (at least for linux) in Makefile.PL that was
	causing ProcessTable.xs not to be converted to ProcessTable.c; 
	this was carelessly introduced by me with integration of the
	port to dec_osf. Added port for netbsd from Peter Reich
	<pr@alles.prima.de>.

0.19	Mon Aug 30 00:22:02 1999
	- Reinstated 'OBJECT' argument in WriteMakefile in Makefile.PL 
	since it was breaking the build on dec-osf, but this time with 
	explicit object files to look for which seems to work, at
	least under Linux. Why this should be necessary is
	unclear. Added "fname" field to HPUX (was "cmd") for
	consistency with other ports and for new test. Added very
	basic "README.hpux". Added check in ProcessTable.pm to prune
	search tree for unreadable directories when finding TTY list;
	this was causing spurious errors during build on dec_osf
	since some directories under /dev are readable only by root.

0.20	Wed Sep  8 22:27:28 1999
	- Added "fields" method to module to return list of fields
	supported for the current architecture. This is implemented in 
	a slightly kludgy way, since it needs to call the table method 
	if the table method hasn't been called before, since only the
	table method knows the names of the fields. Also changed the
	test code to dump all the fields/values available for the
	current process table, instead of just dumping pids and
	"fname" which isn't supported on all architectures (though it
	should be). Added "example.pl" file which contains the same
	code. 

0.21	Fri Sep 10 12:32:13 1999
	- Bug fix; new "fields" method was crapping out on everything
	except Linux-Intel.

0.22	Tue Oct  5 07:47:54 1999 
	- Bug fix for Solaris; nanosecond process times are only
	available on 2.6 (and above?); this was causing the module to
	dump core on earlier systems. Also, size and rss appear to be
	in kilobytes, not pages in 2.6 and above (they are in pages in 
	the struct prpsinfo in < 2.6, but also available directly as
	bytes).