1 Revision history for Perl extension Time::HiRes.
4 - Window/Cygwin: if the performance counter drifts by more than
5 two seconds from the system clock (due to ntp adjustments,
6 for example), recalibrate our internal counter: from Jan Dubois,
7 based on [cpan #5933] by Jerry D. Hedden.
10 - Give a clearer message if the tests timeout (perl change #22253)
11 - Don't use /tmp or its moral equivalents (perl bug #15036,
15 - Windows: ming32 patch from Mike Pomraning (use Perl's Const64()
16 instead of VC-specific i64 suffix)
19 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
22 - Windows: higher resolution time() by using the Windows
23 performance counter API, from Jan Dubois and Anton Shcherbinin.
24 The exact new higher resolution depends on the hardware,
25 but it should be quite a bit better than using the basic
29 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
30 the subtest 18. No known analysis nor fix, but added
31 an alarm (that requires fork() and alarm()) to the test.
34 - doc tweaks from mjd (perl change #20456)
35 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
38 - add a message (for non-core builds) to Makefile.PL about
39 the LC_ALL=C workaround
40 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
41 - the nanosleep() probe was broken (perl change #20061)
42 - use existence instead of definedness for feature probes
44 - MPE/iX tweak (perl change #20042)
45 - do not use HAS_NANOSLEEP (perl change #19898)
48 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
49 - OS/2 can always mix subsecond sleeps with signals
50 (part of perl change #19789)
53 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
57 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
58 - unnecessary (nanosleep is in libc anyway)
59 - harmful (-lrt slows down execution)
60 - incompatible (with many distributions' pthreads)
63 - do not create files in blib directories under core
64 (perl change #19160, from rgs)
65 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
68 - guarantee that $xdefine in HiRes.t is always defined
69 (perl change #19109, from IlyaZ)
70 - a cleaner way to detect PERL_CORE (perl change #19111,
74 - add hints/irix.pl to turn off overly POSIX flags that
75 cause hide struct timespec to be hidden (and compilation
76 to fail) (bleadperl change #19085)
77 - documentation tweaks
80 - add c:/temp to the list of temp directories to probe
81 so that cygwin (and win*?) builds are happy. This was
82 needed at least in my cygwin 1.3.20/w2k setup.
85 - modernize the constants code (from Nicholas Clark)
88 - At some point the ability to figure our the correct incdir
89 for EXTERN.h (either a core perl build, or an installed perl)
90 had broken (which lead into all test compiles failing with
91 a core perl build, but thanks to the robustness of Makefile.PL
92 nothing of this was visible). The brokenness seemed to be
93 caused by $ENV{PERL_CORE} not being on for core builds?
94 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
95 right, and both styles of build should work again.
98 - Nicholas Clark noticed that the my_catdir() emulation function
99 was broken (which means that we didn't really work for Perls
101 - inspired by fixing the above made the whole Makefile.PL -w
103 - tightened up the Makefile.PL output, less whitespace
106 - fix from Craig Berry for better building in VMS with PERL_CORE
109 - no functional changes
110 - move lib/Time/HiRes.pm as Hires.pm
111 - libraries scanning was slightly broken (always scanned
112 for a library even when $Config{libs} already had it)
115 - Ray Zimmerman ran into a race condition in Mac OS X.
116 A 0.01-second alarm fired before the test expected.
117 The test first slept indefinitely (blocking for signals)
118 and only after that tested for the signal having been sent.
119 Since the signal had already been sent, the test #12 never
120 completed. The solution: test first, then block.
121 - default to being silent on all probing attempts, set the
122 environment variable VERBOSE to a true value to see the
123 details (the probing command and the possible errors)
126 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
127 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
128 cannot be upgraded from CPAN to override the 5.8.0 version
130 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
131 because the include directories did not adjust themselves
132 if $ENV{PERL_CORE} (Hugo van der Sanden)
133 - add documentation about the restart of select() under alarm()
136 - small documentation tweaks
140 - better VMS operation (Craig Berry)
143 - our time machine is accelerating: now works with Perl 5.004_01
144 (tried with 5.003_07 and 5.002 but I get segmentation faults
145 from running the Makefile.PL with those in Tru64 4.0D)
148 - backward compatibility (pre-5.6.0) tweaks:
149 - no XSLoader in 5.00503, use DynaLoader instead
150 - no SvPV_nolen, either
151 - no PerlProc_pause(), either
152 - now tested with 5.00404 and 5.00503
153 - Makefile.PL requires 5.00404 (no more 5.002)
154 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
155 (this means that one can mix subsecond sleeps with alarms)
156 - because of nanosleep we probe for -lrt and -lposix4
157 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
158 is available by exportable constants Time::HiRes::d_func
159 (since older Perl do not have them in %Config, and even
160 5.8.0 does not probe for nanosleep)
163 - backward compatibility (pre-5.6.1) tweaks:
164 - define NV if no NVTYPE
165 - define IVdf if needed (note: the Devel::PPPort
166 in 5.8.0 does not try hard hard enough since
167 the IVSIZE might not be defined)
168 - define NVgf if needed
169 - grab the typemap from 5.8.0 for the NV stuff
172 - release 1.29_02 as 1.30
176 - fix a silly unclosed comment typo in HiRes.xs
177 - document and export REALTIME_REALPROF (Solaris)
181 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
182 (need to patch this also in Perl 5.[89])
183 - remove CVS revision log from HiRes.xs
187 The following numbered patches refer to the Perl 5.7 changes,
188 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
190 - 17558: Add #!./perl to the .t
191 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
192 - 16198: political correctness, from Simon Cozens
193 - 15857: doc tweaks, from Jarkko Hietaniemi
194 - 15593: optimization in .xs, from Paul Green
195 - 14892: pod fixes, from Robin Barker
196 - 14100: VOS fixes, from Paul Green
197 - 13422: XS segfault, from Marc Lehmann
198 - 13378: whether select() gets restarted on signals, depends
199 - 13354: timing constraints, again, from Andy Dougherty
200 - 13278: can't do subecond alarms with ualarm;
201 break out early if alarms do not seem to be working
202 - 13266: test relaxation (cygwin gets lower hires
203 times than lores ones)
204 - 12846: protect against high load, from Jarkko Hietaniemi
205 - 12837: HiRes.t VMS tweak, from Craig A. Berry
206 - 12797: HiRes.t VMS tweak, from Charles Lane
207 - 12769: HiRes.t VMS tweak, from Craig A. Berry
208 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
209 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
210 - 12692: alarm() ain't gonna work if ualarm() ain't,
211 from Gurusamy Sarathy
212 - 12680: minor VMS tweak, from Charles Lane
213 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
214 - 12609: croak on negative time, from Jarkko Hietaniemi
215 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
216 - 12594: MacOS Classic timeofday, from Chris Nandor
217 - 12473: allow for more than one second for sleep() and usleep()
218 - 12458: test tuning, relax timing constraints,
219 from Jarkko Hietaniemi
220 - 12449: make sleep() and usleep() to return the number
221 of seconds and microseconds actually slept (analogously
222 with the builtin sleep()), also make usleep() croak if
223 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
224 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
225 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
226 - 12196: use ftime() on Win32, from Artur Bergman
227 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
228 - 12105: use GetSystemTime() on Win32, from Artur Bergman
229 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
230 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
231 - 11797: problem in HiRes.t, from John P. Linderman
232 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
233 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
234 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
235 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
236 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
237 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
239 - 10942: MPE/IX test tweaks, from Mark Bixby
240 - 10784: unnecessary pod2man calls, from Andy Dougherty
241 - 10354: ext/ + -Wall, from Doug MacEachern
242 - 10320: fix the BOOT section to call myU2time correctly
243 - 10317: correct casting for AIX< from H. Merijn Brand
244 - 10119: document that the core time() may be rounding, not truncating
245 - 10118: test fix, from John Peacock
246 - 9988: long =item, from Robin Barker
247 - 9714: correct test output
248 - 9708: test also the scalar aspect of getitimer()
249 - 9705: Add interval timers (setitimer, getitimer)
250 - 9692: do not require at least 5.005 using XS
252 The following changes were made on top of the changes
253 made for Time::HiRes during the Perl 5.7 development
254 cycle that culminated in the release of Perl 5.8.0.
256 - add "require 5.005" to the Makefile.PL
257 - remove the REVISION section (CVS log) from HiRes.pm
258 - add jhi's copyright alongside Douglas'
259 - move HiRes.pm to lib/Time/
261 - modify HiRes.t to use $ENV{PERL_CORE}
262 - modify the original Time::HiRes version 1.20 Makefile.PL
263 to work both with Perl 5.8.0 and the new code with pre-5.8.0
264 Perls (tried with 5.6.1)
265 - tiny tweaks and updates in README and TODO
266 - bump the VERSION to 1.29
268 1.20 Wed Feb 24 21:30 1999
269 - make our usleep and ualarm substitutes into hrt_usleep
270 and hrt_ualarm. This helps static links of Perl with other
271 packages that also have usleep, etc. From
272 Ilya Zakharevich <ilya@math.ohio-state.edu>
273 - add C API stuff. From Joshua Pritikin
274 <joshua.pritikin@db.com>
275 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
276 - hopefully correct "-lc" fix for SCO.
279 1.19 Tue Sep 29 22:30 1998
280 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
281 <seb@stian.demon.co.uk>
282 - change GIMME_V to GIMME to help people with older versions of
284 - fix Win32 version of gettimeofday(). It didn't affect anything,
285 but it confuses people reading the code when the return value
286 is backwards (0 is success).
287 - fix Makefile.PL (more) so that detection of gettimeofday is
290 1.18 Mon Jul 6 22:40 1998
291 - add usleep() for Win32.
292 - fix Makefile.PL to fix reported HP/UX feature where unresolved
293 externals still cause an executable to be generated (though no
294 x bit set). Thanks to David Kozinn for report and explanation.
295 Problems with the fix are mine :)
297 1.17 Wed Jul 1 20:10 1998
298 - fix setitimer calls so microseconds is not more than 1000000.
299 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
300 - make Win32. We only get gettimeofday (the select hack doesn't
301 seem to work on my Win95 system).
302 - fix test 4 on 01test.t. add test to see if time() and
303 Time::HiRes::time() are close.
305 1.16 Wed Nov 12 21:05 1997
306 - add missing EXTEND in new gettimeofday scalar code.
308 1.15 Mon Nov 10 21:30 1997
309 - HiRes.pm: update pod. Provided by Gisle Aas.
310 - HiRes.xs: if gettimeofday() called in scalar context, do
311 something more useful than before. Provided by Gisle Aas.
312 - README: tell of xsubpp '-nolinenumber' woes. thanks to
313 Edward Henigin <ed@texas.net> for pointing out the problem.
315 1.14 Wed Nov 5 9:40 1997
316 - Makefile.PL: look for setitimer
317 - HiRes.xs: if missing ualarm, but we have setitimer, make up
318 our own setitimer. These were provided by Gisle Aas.
320 1.13 Tue Nov 4 23:30 1997
321 - Makefile.PL: fix autodetect mechanism to do try linking in addition
322 to just compiling; should fix Linux build problem. Fix was provided
325 1.12 Sun Oct 12 12:00:00 1997
326 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
327 you may need to comment this back out if you have an older xsubpp.
328 - HiRes.xs: set PROTOTYPES: DISABLE
330 1.11 Fri Sep 05 16:00:00 1997
332 Had some line commented out that shouldn't have been (testing
335 Previous version was corrupted.
337 1.10 Thu May 22 20:20:00 1997
338 - HiRes.xs, HiRes.pm, t/*:
339 - only compile what we have OS support for (or can
341 - only test what we compiled
342 - gross improvement to the test suite
344 This work was all done by Roderick Schertler
345 <roderick@argon.org>. If you run Linux or
346 one of the other ualarm-less platoforms, and you like this
347 module, let Roderick know; without him, it still wouldn't
348 be working on those boxes...
349 - Makefile.PL: figure out what routines the OS has and
350 only build what we need. These bits were written by Jarkko
351 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
353 1.02 Mon Dec 30 08:00:00 1996
354 - HiRes.pm: update documentation to say what to do when missing
355 ualarm() and friends.
356 - README: update to warn that ualarm() and friends need to exist
358 1.01 Fri Oct 17 08:00:00 1996
359 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
360 - HiRes.pm: put blank line between __END__ and =head1 so that
363 1.00 Tue Sep 03 13:00:00 1996
364 - original version; created by h2xs 1.16