1 Revision history for Perl extension Time::HiRes.
4 - no functional changes
5 - move lib/Time/HiRes.pm as Hires.pm
6 - libraries scanning was slightly broken (always scanned
7 for a library even when $Config{libs} already had it)
10 - Ray Zimmerman ran into a race condition in Mac OS X.
11 A 0.01-second alarm fired before the test expected.
12 The test first slept indefinitely (blocking for signals)
13 and only after that tested for the signal having been sent.
14 Since the signal had already been sent, the test #12 never
15 completed. The solution: test first, then block.
16 - default to being silent on all probing attempts, set the
17 environment variable VERBOSE to a true value to see the
18 details (the probing command and the possible errors)
21 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
22 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
23 cannot be upgraded from CPAN to override the 5.8.0 version
25 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
26 because the include directories did not adjust themselves
27 if $ENV{PERL_CORE} (Hugo van der Sanden)
28 - add documentation about the restart of select() under alarm()
31 - small documentation tweaks
35 - better VMS operation (Craig Berry)
38 - our time machine is accelerating: now works with Perl 5.004_01
39 (tried with 5.003_07 and 5.002 but I get segmentation faults
40 from running the Makefile.PL with those in Tru64 4.0D)
43 - backward compatibility (pre-5.6.0) tweaks:
44 - no XSLoader in 5.00503, use DynaLoader instead
45 - no SvPV_nolen, either
46 - no PerlProc_pause(), either
47 - now tested with 5.00404 and 5.00503
48 - Makefile.PL requires 5.00404 (no more 5.002)
49 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
50 (this means that one can mix subsecond sleeps with alarms)
51 - because of nanosleep we probe for -lrt and -lposix4
52 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
53 is available by exportable constants Time::HiRes::d_func
54 (since older Perl do not have them in %Config, and even
55 5.8.0 does not probe for nanosleep)
58 - backward compatibility (pre-5.6.1) tweaks:
59 - define NV if no NVTYPE
60 - define IVdf if needed (note: the Devel::PPPort
61 in 5.8.0 does not try hard hard enough since
62 the IVSIZE might not be defined)
63 - define NVgf if needed
64 - grab the typemap from 5.8.0 for the NV stuff
67 - release 1.29_02 as 1.30
71 - fix a silly unclosed comment typo in HiRes.xs
72 - document and export REALTIME_REALPROF (Solaris)
76 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
77 (need to patch this also in Perl 5.[89])
78 - remove CVS revision log from HiRes.xs
82 The following numbered patches refer to the Perl 5.7 changes,
83 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
85 - 17558: Add #!./perl to the .t
86 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
87 - 16198: political correctness, from Simon Cozens
88 - 15857: doc tweaks, from Jarkko Hietaniemi
89 - 15593: optimization in .xs, from Paul Green
90 - 14892: pod fixes, from Robin Barker
91 - 14100: VOS fixes, from Paul Green
92 - 13422: XS segfault, from Marc Lehmann
93 - 13378: whether select() gets restarted on signals, depends
94 - 13354: timing constraints, again, from Andy Dougherty
95 - 13278: can't do subecond alarms with ualarm;
96 break out early if alarms do not seem to be working
97 - 13266: test relaxation (cygwin gets lower hires
98 times than lores ones)
99 - 12846: protect against high load, from Jarkko Hietaniemi
100 - 12837: HiRes.t VMS tweak, from Craig A. Berry
101 - 12797: HiRes.t VMS tweak, from Charles Lane
102 - 12769: HiRes.t VMS tweak, from Craig A. Berry
103 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
104 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
105 - 12692: alarm() ain't gonna work if ualarm() ain't,
106 from Gurusamy Sarathy
107 - 12680: minor VMS tweak, from Charles Lane
108 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
109 - 12609: croak on negative time, from Jarkko Hietaniemi
110 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
111 - 12594: MacOS Classic timeofday, from Chris Nandor
112 - 12473: allow for more than one second for sleep() and usleep()
113 - 12458: test tuning, relax timing constraints,
114 from Jarkko Hietaniemi
115 - 12449: make sleep() and usleep() to return the number
116 of seconds and microseconds actually slept (analogously
117 with the builtin sleep()), also make usleep() croak if
118 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
119 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
120 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
121 - 12196: use ftime() on Win32, from Artur Bergman
122 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
123 - 12105: use GetSystemTime() on Win32, from Artur Bergman
124 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
125 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
126 - 11797: problem in HiRes.t, from John P. Linderman
127 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
128 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
129 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
130 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
131 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
132 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
134 - 10942: MPE/IX test tweaks, from Mark Bixby
135 - 10784: unnecessary pod2man calls, from Andy Dougherty
136 - 10354: ext/ + -Wall, from Doug MacEachern
137 - 10320: fix the BOOT section to call myU2time correctly
138 - 10317: correct casting for AIX< from H. Merijn Brand
139 - 10119: document that the core time() may be rounding, not truncating
140 - 10118: test fix, from John Peacock
141 - 9988: long =item, from Robin Barker
142 - 9714: correct test output
143 - 9708: test also the scalar aspect of getitimer()
144 - 9705: Add interval timers (setitimer, getitimer)
145 - 9692: do not require at least 5.005 using XS
147 The following changes were made on top of the changes
148 made for Time::HiRes during the Perl 5.7 development
149 cycle that culminated in the release of Perl 5.8.0.
151 - add "require 5.005" to the Makefile.PL
152 - remove the REVISION section (CVS log) from HiRes.pm
153 - add jhi's copyright alongside Douglas'
154 - move HiRes.pm to lib/Time/
156 - modify HiRes.t to use $ENV{PERL_CORE}
157 - modify the original Time::HiRes version 1.20 Makefile.PL
158 to work both with Perl 5.8.0 and the new code with pre-5.8.0
159 Perls (tried with 5.6.1)
160 - tiny tweaks and updates in README and TODO
161 - bump the VERSION to 1.29
163 1.20 Wed Feb 24 21:30 1999
164 - make our usleep and ualarm substitutes into hrt_usleep
165 and hrt_ualarm. This helps static links of Perl with other
166 packages that also have usleep, etc. From
167 Ilya Zakharevich <ilya@math.ohio-state.edu>
168 - add C API stuff. From Joshua Pritikin
169 <joshua.pritikin@db.com>
170 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
171 - hopefully correct "-lc" fix for SCO.
174 1.19 Tue Sep 29 22:30 1998
175 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
176 <seb@stian.demon.co.uk>
177 - change GIMME_V to GIMME to help people with older versions of
179 - fix Win32 version of gettimeofday(). It didn't affect anything,
180 but it confuses people reading the code when the return value
181 is backwards (0 is success).
182 - fix Makefile.PL (more) so that detection of gettimeofday is
185 1.18 Mon Jul 6 22:40 1998
186 - add usleep() for Win32.
187 - fix Makefile.PL to fix reported HP/UX feature where unresolved
188 externals still cause an executable to be generated (though no
189 x bit set). Thanks to David Kozinn for report and explanation.
190 Problems with the fix are mine :)
192 1.17 Wed Jul 1 20:10 1998
193 - fix setitimer calls so microseconds is not more than 1000000.
194 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
195 - make Win32. We only get gettimeofday (the select hack doesn't
196 seem to work on my Win95 system).
197 - fix test 4 on 01test.t. add test to see if time() and
198 Time::HiRes::time() are close.
200 1.16 Wed Nov 12 21:05 1997
201 - add missing EXTEND in new gettimeofday scalar code.
203 1.15 Mon Nov 10 21:30 1997
204 - HiRes.pm: update pod. Provided by Gisle Aas.
205 - HiRes.xs: if gettimeofday() called in scalar context, do
206 something more useful than before. Provided by Gisle Aas.
207 - README: tell of xsubpp '-nolinenumber' woes. thanks to
208 Edward Henigin <ed@texas.net> for pointing out the problem.
210 1.14 Wed Nov 5 9:40 1997
211 - Makefile.PL: look for setitimer
212 - HiRes.xs: if missing ualarm, but we have setitimer, make up
213 our own setitimer. These were provided by Gisle Aas.
215 1.13 Tue Nov 4 23:30 1997
216 - Makefile.PL: fix autodetect mechanism to do try linking in addition
217 to just compiling; should fix Linux build problem. Fix was provided
220 1.12 Sun Oct 12 12:00:00 1997
221 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
222 you may need to comment this back out if you have an older xsubpp.
223 - HiRes.xs: set PROTOTYPES: DISABLE
225 1.11 Fri Sep 05 16:00:00 1997
227 Had some line commented out that shouldn't have been (testing
230 Previous version was corrupted.
232 1.10 Thu May 22 20:20:00 1997
233 - HiRes.xs, HiRes.pm, t/*:
234 - only compile what we have OS support for (or can
236 - only test what we compiled
237 - gross improvement to the test suite
239 This work was all done by Roderick Schertler
240 <roderick@argon.org>. If you run Linux or
241 one of the other ualarm-less platoforms, and you like this
242 module, let Roderick know; without him, it still wouldn't
243 be working on those boxes...
244 - Makefile.PL: figure out what routines the OS has and
245 only build what we need. These bits were written by Jarkko
246 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
248 1.02 Mon Dec 30 08:00:00 1996
249 - HiRes.pm: update documentation to say what to do when missing
250 ualarm() and friends.
251 - README: update to warn that ualarm() and friends need to exist
253 1.01 Fri Oct 17 08:00:00 1996
254 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
255 - HiRes.pm: put blank line between __END__ and =head1 so that
258 1.00 Tue Sep 03 13:00:00 1996
259 - original version; created by h2xs 1.16