1 Revision history for Perl extension Time::HiRes.
4 - add hints/irix.pl to turn off overly POSIX flags that
5 cause hide struct timespec to be hidden (and compilation
10 - add c:/temp to the list of temp directories to probe
11 so that cygwin (and win*?) builds are happy. This was
12 needed at least in my cygwin 1.3.20/w2k setup.
15 - modernize the constants code (from Nicholas Clark)
18 - At some point the ability to figure our the correct incdir
19 for EXTERN.h (either a core perl build, or an installed perl)
20 had broken (which lead into all test compiles failing with
21 a core perl build, but thanks to the robustness of Makefile.PL
22 nothing of this was visible). The brokenness seemed to be
23 caused by $ENV{PERL_CORE} not being on for core builds?
24 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
25 right, and both styles of build should work again.
28 - Nicholas Clark noticed that the my_catdir() emulation function
29 was broken (which means that we didn't really work for Perls
31 - inspired by fixing the above made the whole Makefile.PL -w
33 - tightened up the Makefile.PL output, less whitespace
36 - fix from Craig Berry for better building in VMS with PERL_CORE
39 - no functional changes
40 - move lib/Time/HiRes.pm as Hires.pm
41 - libraries scanning was slightly broken (always scanned
42 for a library even when $Config{libs} already had it)
45 - Ray Zimmerman ran into a race condition in Mac OS X.
46 A 0.01-second alarm fired before the test expected.
47 The test first slept indefinitely (blocking for signals)
48 and only after that tested for the signal having been sent.
49 Since the signal had already been sent, the test #12 never
50 completed. The solution: test first, then block.
51 - default to being silent on all probing attempts, set the
52 environment variable VERBOSE to a true value to see the
53 details (the probing command and the possible errors)
56 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
57 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
58 cannot be upgraded from CPAN to override the 5.8.0 version
60 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
61 because the include directories did not adjust themselves
62 if $ENV{PERL_CORE} (Hugo van der Sanden)
63 - add documentation about the restart of select() under alarm()
66 - small documentation tweaks
70 - better VMS operation (Craig Berry)
73 - our time machine is accelerating: now works with Perl 5.004_01
74 (tried with 5.003_07 and 5.002 but I get segmentation faults
75 from running the Makefile.PL with those in Tru64 4.0D)
78 - backward compatibility (pre-5.6.0) tweaks:
79 - no XSLoader in 5.00503, use DynaLoader instead
80 - no SvPV_nolen, either
81 - no PerlProc_pause(), either
82 - now tested with 5.00404 and 5.00503
83 - Makefile.PL requires 5.00404 (no more 5.002)
84 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
85 (this means that one can mix subsecond sleeps with alarms)
86 - because of nanosleep we probe for -lrt and -lposix4
87 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
88 is available by exportable constants Time::HiRes::d_func
89 (since older Perl do not have them in %Config, and even
90 5.8.0 does not probe for nanosleep)
93 - backward compatibility (pre-5.6.1) tweaks:
94 - define NV if no NVTYPE
95 - define IVdf if needed (note: the Devel::PPPort
96 in 5.8.0 does not try hard hard enough since
97 the IVSIZE might not be defined)
98 - define NVgf if needed
99 - grab the typemap from 5.8.0 for the NV stuff
102 - release 1.29_02 as 1.30
106 - fix a silly unclosed comment typo in HiRes.xs
107 - document and export REALTIME_REALPROF (Solaris)
111 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
112 (need to patch this also in Perl 5.[89])
113 - remove CVS revision log from HiRes.xs
117 The following numbered patches refer to the Perl 5.7 changes,
118 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
120 - 17558: Add #!./perl to the .t
121 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
122 - 16198: political correctness, from Simon Cozens
123 - 15857: doc tweaks, from Jarkko Hietaniemi
124 - 15593: optimization in .xs, from Paul Green
125 - 14892: pod fixes, from Robin Barker
126 - 14100: VOS fixes, from Paul Green
127 - 13422: XS segfault, from Marc Lehmann
128 - 13378: whether select() gets restarted on signals, depends
129 - 13354: timing constraints, again, from Andy Dougherty
130 - 13278: can't do subecond alarms with ualarm;
131 break out early if alarms do not seem to be working
132 - 13266: test relaxation (cygwin gets lower hires
133 times than lores ones)
134 - 12846: protect against high load, from Jarkko Hietaniemi
135 - 12837: HiRes.t VMS tweak, from Craig A. Berry
136 - 12797: HiRes.t VMS tweak, from Charles Lane
137 - 12769: HiRes.t VMS tweak, from Craig A. Berry
138 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
139 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
140 - 12692: alarm() ain't gonna work if ualarm() ain't,
141 from Gurusamy Sarathy
142 - 12680: minor VMS tweak, from Charles Lane
143 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
144 - 12609: croak on negative time, from Jarkko Hietaniemi
145 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
146 - 12594: MacOS Classic timeofday, from Chris Nandor
147 - 12473: allow for more than one second for sleep() and usleep()
148 - 12458: test tuning, relax timing constraints,
149 from Jarkko Hietaniemi
150 - 12449: make sleep() and usleep() to return the number
151 of seconds and microseconds actually slept (analogously
152 with the builtin sleep()), also make usleep() croak if
153 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
154 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
155 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
156 - 12196: use ftime() on Win32, from Artur Bergman
157 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
158 - 12105: use GetSystemTime() on Win32, from Artur Bergman
159 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
160 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
161 - 11797: problem in HiRes.t, from John P. Linderman
162 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
163 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
164 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
165 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
166 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
167 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
169 - 10942: MPE/IX test tweaks, from Mark Bixby
170 - 10784: unnecessary pod2man calls, from Andy Dougherty
171 - 10354: ext/ + -Wall, from Doug MacEachern
172 - 10320: fix the BOOT section to call myU2time correctly
173 - 10317: correct casting for AIX< from H. Merijn Brand
174 - 10119: document that the core time() may be rounding, not truncating
175 - 10118: test fix, from John Peacock
176 - 9988: long =item, from Robin Barker
177 - 9714: correct test output
178 - 9708: test also the scalar aspect of getitimer()
179 - 9705: Add interval timers (setitimer, getitimer)
180 - 9692: do not require at least 5.005 using XS
182 The following changes were made on top of the changes
183 made for Time::HiRes during the Perl 5.7 development
184 cycle that culminated in the release of Perl 5.8.0.
186 - add "require 5.005" to the Makefile.PL
187 - remove the REVISION section (CVS log) from HiRes.pm
188 - add jhi's copyright alongside Douglas'
189 - move HiRes.pm to lib/Time/
191 - modify HiRes.t to use $ENV{PERL_CORE}
192 - modify the original Time::HiRes version 1.20 Makefile.PL
193 to work both with Perl 5.8.0 and the new code with pre-5.8.0
194 Perls (tried with 5.6.1)
195 - tiny tweaks and updates in README and TODO
196 - bump the VERSION to 1.29
198 1.20 Wed Feb 24 21:30 1999
199 - make our usleep and ualarm substitutes into hrt_usleep
200 and hrt_ualarm. This helps static links of Perl with other
201 packages that also have usleep, etc. From
202 Ilya Zakharevich <ilya@math.ohio-state.edu>
203 - add C API stuff. From Joshua Pritikin
204 <joshua.pritikin@db.com>
205 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
206 - hopefully correct "-lc" fix for SCO.
209 1.19 Tue Sep 29 22:30 1998
210 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
211 <seb@stian.demon.co.uk>
212 - change GIMME_V to GIMME to help people with older versions of
214 - fix Win32 version of gettimeofday(). It didn't affect anything,
215 but it confuses people reading the code when the return value
216 is backwards (0 is success).
217 - fix Makefile.PL (more) so that detection of gettimeofday is
220 1.18 Mon Jul 6 22:40 1998
221 - add usleep() for Win32.
222 - fix Makefile.PL to fix reported HP/UX feature where unresolved
223 externals still cause an executable to be generated (though no
224 x bit set). Thanks to David Kozinn for report and explanation.
225 Problems with the fix are mine :)
227 1.17 Wed Jul 1 20:10 1998
228 - fix setitimer calls so microseconds is not more than 1000000.
229 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
230 - make Win32. We only get gettimeofday (the select hack doesn't
231 seem to work on my Win95 system).
232 - fix test 4 on 01test.t. add test to see if time() and
233 Time::HiRes::time() are close.
235 1.16 Wed Nov 12 21:05 1997
236 - add missing EXTEND in new gettimeofday scalar code.
238 1.15 Mon Nov 10 21:30 1997
239 - HiRes.pm: update pod. Provided by Gisle Aas.
240 - HiRes.xs: if gettimeofday() called in scalar context, do
241 something more useful than before. Provided by Gisle Aas.
242 - README: tell of xsubpp '-nolinenumber' woes. thanks to
243 Edward Henigin <ed@texas.net> for pointing out the problem.
245 1.14 Wed Nov 5 9:40 1997
246 - Makefile.PL: look for setitimer
247 - HiRes.xs: if missing ualarm, but we have setitimer, make up
248 our own setitimer. These were provided by Gisle Aas.
250 1.13 Tue Nov 4 23:30 1997
251 - Makefile.PL: fix autodetect mechanism to do try linking in addition
252 to just compiling; should fix Linux build problem. Fix was provided
255 1.12 Sun Oct 12 12:00:00 1997
256 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
257 you may need to comment this back out if you have an older xsubpp.
258 - HiRes.xs: set PROTOTYPES: DISABLE
260 1.11 Fri Sep 05 16:00:00 1997
262 Had some line commented out that shouldn't have been (testing
265 Previous version was corrupted.
267 1.10 Thu May 22 20:20:00 1997
268 - HiRes.xs, HiRes.pm, t/*:
269 - only compile what we have OS support for (or can
271 - only test what we compiled
272 - gross improvement to the test suite
274 This work was all done by Roderick Schertler
275 <roderick@argon.org>. If you run Linux or
276 one of the other ualarm-less platoforms, and you like this
277 module, let Roderick know; without him, it still wouldn't
278 be working on those boxes...
279 - Makefile.PL: figure out what routines the OS has and
280 only build what we need. These bits were written by Jarkko
281 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
283 1.02 Mon Dec 30 08:00:00 1996
284 - HiRes.pm: update documentation to say what to do when missing
285 ualarm() and friends.
286 - README: update to warn that ualarm() and friends need to exist
288 1.01 Fri Oct 17 08:00:00 1996
289 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
290 - HiRes.pm: put blank line between __END__ and =head1 so that
293 1.00 Tue Sep 03 13:00:00 1996
294 - original version; created by h2xs 1.16