1 Revision history for Perl extension Time::HiRes.
4 - somehow 1.67 had a lot of doubled lines (a major cut-and-paste
5 error suspected), but miraculously it still worked since the
6 doubling took place below the __END__ token
7 - undef Pause() before defining it to avoid redefinition warnings
8 during compilation in case perl.h had already defined Pause()
9 (part of perl change #24271)
13 - (internal) don't ignore the return value of gettimeofday()
14 - (external) return undef or an empty if the C gettimeofday() fails
15 (affects Time::HiRes gettimeofday() and the hires time())
19 - fix the 'hierachy' typo in Makefile.PL [rt.cpan.org #8492]
20 - should now build in Solaris [rt.cpan.org #7165] (since 1.64)
21 - should now build in Cygwin [rt.cpan.org #7535] (since 1.64)
22 - close also [rt.cpan.org #5933] "Time::HiRes::time does not pick up time adjustments like ntp" since ever reproducing it in the same environment
23 has become rather unlikely
26 - one should not mix u?alarm and sleep (the tests modified
27 by 1.65, #12 and #13, hung in Solaris), now we just busy
28 loop executing an empty block
29 - in the documentation underline the unspecificity of mixing
31 - small spelling fixes
34 - regenerate ppport.h with Devel::PPPort 3.03,
35 now the MY_CXT_CLONE is defined in ppport.h,
36 we no more need to do that.
38 - the test #12 would often hang in sigsuspend() (at least that's
39 where Mac OS X' ktrace shows it hanging). With the sleep()s
40 changed to sleep(1)s, the tests still pass but no hang after
41 a few hundred repeats.
44 - Win32 and any ithread build: ppport.h didn't define
45 MY_CXT_CLONE, which seems to be a Time-HiResism.
48 - Skip testing if under PERL_CORE and Time::HiRes has not
49 been Configured (from Marcus Holland-Moritz, core change
51 - Use ppport.h generated by Devel::PPPort 3.01,
52 allowing cutting away our own portability code.
53 - Don't use $ENV{PERL_CORE} for < 5.6.0.
54 - Don't use "for my $i" for <= 5.003.
55 - Don't use Pause() for <= 5.003.
56 - Can't use newSVpvf for <= 5.003.
57 (most of the changes from Marcus)
60 - Win32: reset reading from the performance counters every
61 five minutes to better track wall clock time (thanks to
62 PC timers being often quite bad), should help long-running
66 - Win32: Patch from Steve Hay
67 [PATCH] Re: [perl #30755] [Win32] Different results from Time::HiRes::gettimeofdayunder the debugger
68 to [perl #30755] reported by Nigel Sandever
70 - Cygwin: Use the Win32 recalibration code also in Cygwin if the
71 <w32api/windows.h> APIs are available. Cygwin testing by
72 Yitzchak Scott-Thoennes.
74 - Solaris: use -lposix4 to get nanosleep for Solaris 2.6,
75 after that keep using -lrt, patch from Alan Burlison,
76 bug reported in [cpan #7165]
79 - Change the Win32 recalibration limit to 0.5 seconds and tweak
80 the documentation to blather less about the gory details of the
81 Win32 implementation and more about the complications in general
82 of meddling with the system clock.
85 - Document the 1.57 change better.
88 - Win32/Cygwin/MinGW: if the performance counter drifts by more
89 than two seconds from the system clock (due to ntp adjustments,
90 for example), recalibrate our internal counter: from Jan Dubois,
91 based on [cpan #5933] by Jerry D. Hedden.
94 - Give a clearer message if the tests timeout (perl change #22253)
95 - Don't use /tmp or its moral equivalents (perl bug #15036,
99 - Windows: mingw32 patch from Mike Pomraning (use Perl's Const64()
100 instead of VC-specific i64 suffix)
103 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
106 - Windows: higher resolution time() by using the Windows
107 performance counter API, from Jan Dubois and Anton Shcherbinin.
108 The exact new higher resolution depends on the hardware,
109 but it should be quite a bit better than using the basic
113 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
114 the subtest 18. No known analysis nor fix, but added
115 an alarm (that requires fork() and alarm()) to the test.
118 - doc tweaks from mjd (perl change #20456)
119 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
122 - add a message (for non-core builds) to Makefile.PL about
123 the LC_ALL=C workaround
124 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
125 - the nanosleep() probe was broken (perl change #20061)
126 - use existence instead of definedness for feature probes
128 - MPE/iX tweak (perl change #20042)
129 - do not use HAS_NANOSLEEP (perl change #19898)
132 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
133 - OS/2 can always mix subsecond sleeps with signals
134 (part of perl change #19789)
137 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
141 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
142 - unnecessary (nanosleep is in libc anyway)
143 - harmful (-lrt slows down execution)
144 - incompatible (with many distributions' pthreads)
147 - do not create files in blib directories under core
148 (perl change #19160, from rgs)
149 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
152 - guarantee that $xdefine in HiRes.t is always defined
153 (perl change #19109, from IlyaZ)
154 - a cleaner way to detect PERL_CORE (perl change #19111,
158 - add hints/irix.pl to turn off overly POSIX flags that
159 cause hide struct timespec to be hidden (and compilation
160 to fail) (bleadperl change #19085)
161 - documentation tweaks
164 - add c:/temp to the list of temp directories to probe
165 so that cygwin (and win*?) builds are happy. This was
166 needed at least in my cygwin 1.3.20/w2k setup.
169 - modernize the constants code (from Nicholas Clark)
172 - At some point the ability to figure our the correct incdir
173 for EXTERN.h (either a core perl build, or an installed perl)
174 had broken (which lead into all test compiles failing with
175 a core perl build, but thanks to the robustness of Makefile.PL
176 nothing of this was visible). The brokenness seemed to be
177 caused by $ENV{PERL_CORE} not being on for core builds?
178 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
179 right, and both styles of build should work again.
182 - Nicholas Clark noticed that the my_catdir() emulation function
183 was broken (which means that we didn't really work for Perls
185 - inspired by fixing the above made the whole Makefile.PL -w
187 - tightened up the Makefile.PL output, less whitespace
190 - fix from Craig Berry for better building in VMS with PERL_CORE
193 - no functional changes
194 - move lib/Time/HiRes.pm as Hires.pm
195 - libraries scanning was slightly broken (always scanned
196 for a library even when $Config{libs} already had it)
199 - Ray Zimmerman ran into a race condition in Mac OS X.
200 A 0.01-second alarm fired before the test expected.
201 The test first slept indefinitely (blocking for signals)
202 and only after that tested for the signal having been sent.
203 Since the signal had already been sent, the test #12 never
204 completed. The solution: test first, then block.
205 - default to being silent on all probing attempts, set the
206 environment variable VERBOSE to a true value to see the
207 details (the probing command and the possible errors)
210 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
211 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
212 cannot be upgraded from CPAN to override the 5.8.0 version
214 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
215 because the include directories did not adjust themselves
216 if $ENV{PERL_CORE} (Hugo van der Sanden)
217 - add documentation about the restart of select() under alarm()
220 - small documentation tweaks
224 - better VMS operation (Craig Berry)
227 - our time machine is accelerating: now works with Perl 5.004_01
228 (tried with 5.003_07 and 5.002 but I get segmentation faults
229 from running the Makefile.PL with those in Tru64 4.0D)
232 - backward compatibility (pre-5.6.0) tweaks:
233 - no XSLoader in 5.00503, use DynaLoader instead
234 - no SvPV_nolen, either
235 - no PerlProc_pause(), either
236 - now tested with 5.00404 and 5.00503
237 - Makefile.PL requires 5.00404 (no more 5.002)
238 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
239 (this means that one can mix subsecond sleeps with alarms)
240 - because of nanosleep we probe for -lrt and -lposix4
241 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
242 is available by exportable constants Time::HiRes::d_func
243 (since older Perl do not have them in %Config, and even
244 5.8.0 does not probe for nanosleep)
247 - backward compatibility (pre-5.6.1) tweaks:
248 - define NV if no NVTYPE
249 - define IVdf if needed (note: the Devel::PPPort
250 in 5.8.0 does not try hard hard enough since
251 the IVSIZE might not be defined)
252 - define NVgf if needed
253 - grab the typemap from 5.8.0 for the NV stuff
256 - release 1.29_02 as 1.30
260 - fix a silly unclosed comment typo in HiRes.xs
261 - document and export REALTIME_REALPROF (Solaris)
265 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
266 (need to patch this also in Perl 5.[89])
267 - remove CVS revision log from HiRes.xs
271 The following numbered patches refer to the Perl 5.7 changes,
272 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
274 - 17558: Add #!./perl to the .t
275 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
276 - 16198: political correctness, from Simon Cozens
277 - 15857: doc tweaks, from Jarkko Hietaniemi
278 - 15593: optimization in .xs, from Paul Green
279 - 14892: pod fixes, from Robin Barker
280 - 14100: VOS fixes, from Paul Green
281 - 13422: XS segfault, from Marc Lehmann
282 - 13378: whether select() gets restarted on signals, depends
283 - 13354: timing constraints, again, from Andy Dougherty
284 - 13278: can't do subecond alarms with ualarm;
285 break out early if alarms do not seem to be working
286 - 13266: test relaxation (cygwin gets lower hires
287 times than lores ones)
288 - 12846: protect against high load, from Jarkko Hietaniemi
289 - 12837: HiRes.t VMS tweak, from Craig A. Berry
290 - 12797: HiRes.t VMS tweak, from Charles Lane
291 - 12769: HiRes.t VMS tweak, from Craig A. Berry
292 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
293 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
294 - 12692: alarm() ain't gonna work if ualarm() ain't,
295 from Gurusamy Sarathy
296 - 12680: minor VMS tweak, from Charles Lane
297 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
298 - 12609: croak on negative time, from Jarkko Hietaniemi
299 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
300 - 12594: MacOS Classic timeofday, from Chris Nandor
301 - 12473: allow for more than one second for sleep() and usleep()
302 - 12458: test tuning, relax timing constraints,
303 from Jarkko Hietaniemi
304 - 12449: make sleep() and usleep() to return the number
305 of seconds and microseconds actually slept (analogously
306 with the builtin sleep()), also make usleep() croak if
307 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
308 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
309 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
310 - 12196: use ftime() on Win32, from Artur Bergman
311 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
312 - 12105: use GetSystemTime() on Win32, from Artur Bergman
313 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
314 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
315 - 11797: problem in HiRes.t, from John P. Linderman
316 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
317 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
318 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
319 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
320 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
321 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
323 - 10942: MPE/IX test tweaks, from Mark Bixby
324 - 10784: unnecessary pod2man calls, from Andy Dougherty
325 - 10354: ext/ + -Wall, from Doug MacEachern
326 - 10320: fix the BOOT section to call myU2time correctly
327 - 10317: correct casting for AIX< from H. Merijn Brand
328 - 10119: document that the core time() may be rounding, not truncating
329 - 10118: test fix, from John Peacock
330 - 9988: long =item, from Robin Barker
331 - 9714: correct test output
332 - 9708: test also the scalar aspect of getitimer()
333 - 9705: Add interval timers (setitimer, getitimer)
334 - 9692: do not require at least 5.005 using XS
336 The following changes were made on top of the changes
337 made for Time::HiRes during the Perl 5.7 development
338 cycle that culminated in the release of Perl 5.8.0.
340 - add "require 5.005" to the Makefile.PL
341 - remove the REVISION section (CVS log) from HiRes.pm
342 - add jhi's copyright alongside Douglas'
343 - move HiRes.pm to lib/Time/
345 - modify HiRes.t to use $ENV{PERL_CORE}
346 - modify the original Time::HiRes version 1.20 Makefile.PL
347 to work both with Perl 5.8.0 and the new code with pre-5.8.0
348 Perls (tried with 5.6.1)
349 - tiny tweaks and updates in README and TODO
350 - bump the VERSION to 1.29
352 1.20 Wed Feb 24 21:30 1999
353 - make our usleep and ualarm substitutes into hrt_usleep
354 and hrt_ualarm. This helps static links of Perl with other
355 packages that also have usleep, etc. From
356 Ilya Zakharevich <ilya@math.ohio-state.edu>
357 - add C API stuff. From Joshua Pritikin
358 <joshua.pritikin@db.com>
359 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
360 - hopefully correct "-lc" fix for SCO.
363 1.19 Tue Sep 29 22:30 1998
364 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
365 <seb@stian.demon.co.uk>
366 - change GIMME_V to GIMME to help people with older versions of
368 - fix Win32 version of gettimeofday(). It didn't affect anything,
369 but it confuses people reading the code when the return value
370 is backwards (0 is success).
371 - fix Makefile.PL (more) so that detection of gettimeofday is
374 1.18 Mon Jul 6 22:40 1998
375 - add usleep() for Win32.
376 - fix Makefile.PL to fix reported HP/UX feature where unresolved
377 externals still cause an executable to be generated (though no
378 x bit set). Thanks to David Kozinn for report and explanation.
379 Problems with the fix are mine :)
381 1.17 Wed Jul 1 20:10 1998
382 - fix setitimer calls so microseconds is not more than 1000000.
383 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
384 - make Win32. We only get gettimeofday (the select hack doesn't
385 seem to work on my Win95 system).
386 - fix test 4 on 01test.t. add test to see if time() and
387 Time::HiRes::time() are close.
389 1.16 Wed Nov 12 21:05 1997
390 - add missing EXTEND in new gettimeofday scalar code.
392 1.15 Mon Nov 10 21:30 1997
393 - HiRes.pm: update pod. Provided by Gisle Aas.
394 - HiRes.xs: if gettimeofday() called in scalar context, do
395 something more useful than before. Provided by Gisle Aas.
396 - README: tell of xsubpp '-nolinenumber' woes. thanks to
397 Edward Henigin <ed@texas.net> for pointing out the problem.
399 1.14 Wed Nov 5 9:40 1997
400 - Makefile.PL: look for setitimer
401 - HiRes.xs: if missing ualarm, but we have setitimer, make up
402 our own setitimer. These were provided by Gisle Aas.
404 1.13 Tue Nov 4 23:30 1997
405 - Makefile.PL: fix autodetect mechanism to do try linking in addition
406 to just compiling; should fix Linux build problem. Fix was provided
409 1.12 Sun Oct 12 12:00:00 1997
410 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
411 you may need to comment this back out if you have an older xsubpp.
412 - HiRes.xs: set PROTOTYPES: DISABLE
414 1.11 Fri Sep 05 16:00:00 1997
416 Had some line commented out that shouldn't have been (testing
419 Previous version was corrupted.
421 1.10 Thu May 22 20:20:00 1997
422 - HiRes.xs, HiRes.pm, t/*:
423 - only compile what we have OS support for (or can
425 - only test what we compiled
426 - gross improvement to the test suite
428 This work was all done by Roderick Schertler
429 <roderick@argon.org>. If you run Linux or
430 one of the other ualarm-less platoforms, and you like this
431 module, let Roderick know; without him, it still wouldn't
432 be working on those boxes...
433 - Makefile.PL: figure out what routines the OS has and
434 only build what we need. These bits were written by Jarkko
435 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
437 1.02 Mon Dec 30 08:00:00 1996
438 - HiRes.pm: update documentation to say what to do when missing
439 ualarm() and friends.
440 - README: update to warn that ualarm() and friends need to exist
442 1.01 Fri Oct 17 08:00:00 1996
443 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
444 - HiRes.pm: put blank line between __END__ and =head1 so that
447 1.00 Tue Sep 03 13:00:00 1996
448 - original version; created by h2xs 1.16