1 Revision history for Perl extension Time::HiRes.
4 - installation patch from Gisle Aas: in Perls 5.8.x and later
5 use MakeMaker INSTALLDIRS value of 'perl' instead of 'site'.
8 - [cpan #14608] Solaris 8 perl 5.005_03 File::Spec module does not have method rel2abs
9 (the workaround is not to use rel2abs, should not be necessary)
10 - [cpan #14642] U2time wrongly exported on the C API
11 (patch supplied by the reporter, SALVA@cpan.org)
12 - add release dates to Changes
15 - Time::HiRes::nanosleep support for Solaris [PATCH]
16 (POSIX::uname() not available if building with core perl,
17 from Gisle Aas, via perl5-porters, perl change #25295)
20 - going back to the 1.68 loader setup (using DynaLoader)
21 since too many weird things starting breaking
22 - fix a typo in José Auguste-Etienne's name
25 - a thinko in the nanosleep() detection
26 - move more changes stuff from the README to Changes
27 - add -w to the Makefile.PL
30 - oops in 1.69 about @ISA (not affecting anything but silly)
31 - add copyright 2005 to HiRes.pm
32 - add copyright and license to HiRes.xs
33 - add copyrights 2003, 2004, 2005 to README
36 - actually run a test for nanosleep
37 (if there is no $Config{d_nanosleep}) since e.g. in AIX 4.2
38 it seems that one can link in nanosleep() but then calling
39 it fails instantly and sets errno to ENOSYS (Not implemented).
40 This may be fixable in the AIX case by figuring out the right
41 (realtime POSIX?) libs and whatnot, but in the general case
42 running a real test case is better. (Of course, this change
43 will no doubt run into portability problems because of the
44 execution step...) Note that because of hysterical raisins
45 most Perls do NOT have $Config{d_nanosleep} (scanning for
46 it by Configure would in many platforms require linking in
47 things like -lrt, which would in many platforms be a bad idea
49 (from José Auguste-Etienne)
50 - support XSLoader also since it's much faster
52 - add SEE ALSO (BSD::Resource and Time::TAI64)
55 - somehow 1.67 had a lot of doubled lines (a major cut-and-paste
56 error suspected), but miraculously it still worked since the
57 doubling took place below the __END__ token
58 - undef Pause() before defining it to avoid redefinition warnings
59 during compilation in case perl.h had already defined Pause()
60 (part of perl change #24271)
64 - (internal) don't ignore the return value of gettimeofday()
65 - (external) return undef or an empty if the C gettimeofday() fails
66 (affects Time::HiRes gettimeofday() and the hires time())
70 - fix the 'hierachy' typo in Makefile.PL [rt.cpan.org #8492]
71 - should now build in Solaris [rt.cpan.org #7165] (since 1.64)
72 - should now build in Cygwin [rt.cpan.org #7535] (since 1.64)
73 - 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
74 has become rather unlikely
77 - one should not mix u?alarm and sleep (the tests modified
78 by 1.65, #12 and #13, hung in Solaris), now we just busy
79 loop executing an empty block
80 - in the documentation underline the unspecificity of mixing
82 - small spelling fixes
85 - regenerate ppport.h with Devel::PPPort 3.03,
86 now the MY_CXT_CLONE is defined in ppport.h,
87 we no more need to do that.
89 - the test #12 would often hang in sigsuspend() (at least that's
90 where Mac OS X' ktrace shows it hanging). With the sleep()s
91 changed to sleep(1)s, the tests still pass but no hang after
92 a few hundred repeats.
95 - Win32 and any ithread build: ppport.h didn't define
96 MY_CXT_CLONE, which seems to be a Time-HiResism.
99 - Skip testing if under PERL_CORE and Time::HiRes has not
100 been Configured (from Marcus Holland-Moritz, core change
102 - Use ppport.h generated by Devel::PPPort 3.01,
103 allowing cutting away our own portability code.
104 - Don't use $ENV{PERL_CORE} for < 5.6.0.
105 - Don't use "for my $i" for <= 5.003.
106 - Don't use Pause() for <= 5.003.
107 - Can't use newSVpvf for <= 5.003.
108 (most of the changes from Marcus)
111 - Win32: reset reading from the performance counters every
112 five minutes to better track wall clock time (thanks to
113 PC timers being often quite bad), should help long-running
117 - Win32: Patch from Steve Hay
118 [PATCH] Re: [perl #30755] [Win32] Different results from Time::HiRes::gettimeofdayunder the debugger
119 to [perl #30755] reported by Nigel Sandever
121 - Cygwin: Use the Win32 recalibration code also in Cygwin if the
122 <w32api/windows.h> APIs are available. Cygwin testing by
123 Yitzchak Scott-Thoennes.
125 - Solaris: use -lposix4 to get nanosleep for Solaris 2.6,
126 after that keep using -lrt, patch from Alan Burlison,
127 bug reported in [cpan #7165]
130 - Change the Win32 recalibration limit to 0.5 seconds and tweak
131 the documentation to blather less about the gory details of the
132 Win32 implementation and more about the complications in general
133 of meddling with the system clock.
136 - Document the 1.57 change better.
139 - Win32/Cygwin/MinGW: if the performance counter drifts by more
140 than two seconds from the system clock (due to ntp adjustments,
141 for example), recalibrate our internal counter: from Jan Dubois,
142 based on [cpan #5933] by Jerry D. Hedden.
145 - Give a clearer message if the tests timeout (perl change #22253)
146 - Don't use /tmp or its moral equivalents (perl bug #15036,
150 - Windows: mingw32 patch from Mike Pomraning (use Perl's Const64()
151 instead of VC-specific i64 suffix)
154 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
157 - Windows: higher resolution time() by using the Windows
158 performance counter API, from Jan Dubois and Anton Shcherbinin.
159 The exact new higher resolution depends on the hardware,
160 but it should be quite a bit better than using the basic
164 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
165 the subtest 18. No known analysis nor fix, but added
166 an alarm (that requires fork() and alarm()) to the test.
169 - doc tweaks from mjd (perl change #20456)
170 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
173 - add a message (for non-core builds) to Makefile.PL about
174 the LC_ALL=C workaround
175 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
176 - the nanosleep() probe was broken (perl change #20061)
177 - use existence instead of definedness for feature probes
179 - MPE/iX tweak (perl change #20042)
180 - do not use HAS_NANOSLEEP (perl change #19898)
183 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
184 - OS/2 can always mix subsecond sleeps with signals
185 (part of perl change #19789)
188 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
192 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
193 - unnecessary (nanosleep is in libc anyway)
194 - harmful (-lrt slows down execution)
195 - incompatible (with many distributions' pthreads)
198 - do not create files in blib directories under core
199 (perl change #19160, from rgs)
200 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
203 - guarantee that $xdefine in HiRes.t is always defined
204 (perl change #19109, from IlyaZ)
205 - a cleaner way to detect PERL_CORE (perl change #19111,
209 - add hints/irix.pl to turn off overly POSIX flags that
210 cause hide struct timespec to be hidden (and compilation
211 to fail) (bleadperl change #19085)
212 - documentation tweaks
215 - add c:/temp to the list of temp directories to probe
216 so that cygwin (and win*?) builds are happy. This was
217 needed at least in my cygwin 1.3.20/w2k setup.
220 - modernize the constants code (from Nicholas Clark)
223 - At some point the ability to figure our the correct incdir
224 for EXTERN.h (either a core perl build, or an installed perl)
225 had broken (which lead into all test compiles failing with
226 a core perl build, but thanks to the robustness of Makefile.PL
227 nothing of this was visible). The brokenness seemed to be
228 caused by $ENV{PERL_CORE} not being on for core builds?
229 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
230 right, and both styles of build should work again.
233 - Nicholas Clark noticed that the my_catdir() emulation function
234 was broken (which means that we didn't really work for Perls
236 - inspired by fixing the above made the whole Makefile.PL -w
238 - tightened up the Makefile.PL output, less whitespace
241 - fix from Craig Berry for better building in VMS with PERL_CORE
244 - no functional changes
245 - move lib/Time/HiRes.pm as Hires.pm
246 - libraries scanning was slightly broken (always scanned
247 for a library even when $Config{libs} already had it)
250 - Ray Zimmerman ran into a race condition in Mac OS X.
251 A 0.01-second alarm fired before the test expected.
252 The test first slept indefinitely (blocking for signals)
253 and only after that tested for the signal having been sent.
254 Since the signal had already been sent, the test #12 never
255 completed. The solution: test first, then block.
256 - default to being silent on all probing attempts, set the
257 environment variable VERBOSE to a true value to see the
258 details (the probing command and the possible errors)
261 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
262 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
263 cannot be upgraded from CPAN to override the 5.8.0 version
265 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
266 because the include directories did not adjust themselves
267 if $ENV{PERL_CORE} (Hugo van der Sanden)
268 - add documentation about the restart of select() under alarm()
271 - small documentation tweaks
275 - better VMS operation (Craig Berry)
278 - our time machine is accelerating: now works with Perl 5.004_01
279 (tried with 5.003_07 and 5.002 but I get segmentation faults
280 from running the Makefile.PL with those in Tru64 4.0D)
283 - backward compatibility (pre-5.6.0) tweaks:
284 - no XSLoader in 5.00503, use DynaLoader instead
285 - no SvPV_nolen, either
286 - no PerlProc_pause(), either
287 - now tested with 5.00404 and 5.00503
288 - Makefile.PL requires 5.00404 (no more 5.002)
289 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
290 (this means that one can mix subsecond sleeps with alarms)
291 - because of nanosleep we probe for -lrt and -lposix4
292 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
293 is available by exportable constants Time::HiRes::d_func
294 (since older Perl do not have them in %Config, and even
295 5.8.0 does not probe for nanosleep)
298 - backward compatibility (pre-5.6.1) tweaks:
299 - define NV if no NVTYPE
300 - define IVdf if needed (note: the Devel::PPPort
301 in 5.8.0 does not try hard hard enough since
302 the IVSIZE might not be defined)
303 - define NVgf if needed
304 - grab the typemap from 5.8.0 for the NV stuff
306 1.31 and 1.32 add more backward compatibility (now all the way
307 back to Perl 5.00404), and using nanosleep() (if available) for
312 - release 1.29_02 as 1.30
314 1.30 adds all the changes made during the Perl 5.6->5.7->5.8
315 development cycle. Most notably portability across platforms has been
316 enhanced, and the interval timers (setitimer, getitimer) have been
317 added. Note that the version of Time::HiRes that is included in Perl
318 5.8.0 calls itself 1.20_00, but it is equivalent to this Time::HiRes
319 version. Note also that in 1.30 Wegscheid turns over the maintenance
320 to Jarkko Hietaniemi.
324 - fix a silly unclosed comment typo in HiRes.xs
325 - document and export REALTIME_REALPROF (Solaris)
329 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
330 (need to patch this also in Perl 5.[89])
331 - remove CVS revision log from HiRes.xs
335 The following numbered patches refer to the Perl 5.7 changes,
336 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
338 - 17558: Add #!./perl to the .t
339 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
340 - 16198: political correctness, from Simon Cozens
341 - 15857: doc tweaks, from Jarkko Hietaniemi
342 - 15593: optimization in .xs, from Paul Green
343 - 14892: pod fixes, from Robin Barker
344 - 14100: VOS fixes, from Paul Green
345 - 13422: XS segfault, from Marc Lehmann
346 - 13378: whether select() gets restarted on signals, depends
347 - 13354: timing constraints, again, from Andy Dougherty
348 - 13278: can't do subecond alarms with ualarm;
349 break out early if alarms do not seem to be working
350 - 13266: test relaxation (cygwin gets lower hires
351 times than lores ones)
352 - 12846: protect against high load, from Jarkko Hietaniemi
353 - 12837: HiRes.t VMS tweak, from Craig A. Berry
354 - 12797: HiRes.t VMS tweak, from Charles Lane
355 - 12769: HiRes.t VMS tweak, from Craig A. Berry
356 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
357 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
358 - 12692: alarm() ain't gonna work if ualarm() ain't,
359 from Gurusamy Sarathy
360 - 12680: minor VMS tweak, from Charles Lane
361 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
362 - 12609: croak on negative time, from Jarkko Hietaniemi
363 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
364 - 12594: MacOS Classic timeofday, from Chris Nandor
365 - 12473: allow for more than one second for sleep() and usleep()
366 - 12458: test tuning, relax timing constraints,
367 from Jarkko Hietaniemi
368 - 12449: make sleep() and usleep() to return the number
369 of seconds and microseconds actually slept (analogously
370 with the builtin sleep()), also make usleep() croak if
371 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
372 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
373 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
374 - 12196: use ftime() on Win32, from Artur Bergman
375 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
376 - 12105: use GetSystemTime() on Win32, from Artur Bergman
377 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
378 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
379 - 11797: problem in HiRes.t, from John P. Linderman
380 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
381 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
382 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
383 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
384 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
385 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
387 - 10942: MPE/IX test tweaks, from Mark Bixby
388 - 10784: unnecessary pod2man calls, from Andy Dougherty
389 - 10354: ext/ + -Wall, from Doug MacEachern
390 - 10320: fix the BOOT section to call myU2time correctly
391 - 10317: correct casting for AIX< from H. Merijn Brand
392 - 10119: document that the core time() may be rounding, not truncating
393 - 10118: test fix, from John Peacock
394 - 9988: long =item, from Robin Barker
395 - 9714: correct test output
396 - 9708: test also the scalar aspect of getitimer()
397 - 9705: Add interval timers (setitimer, getitimer)
398 - 9692: do not require at least 5.005 using XS
400 The following changes were made on top of the changes
401 made for Time::HiRes during the Perl 5.7 development
402 cycle that culminated in the release of Perl 5.8.0.
404 - add "require 5.005" to the Makefile.PL
405 - remove the REVISION section (CVS log) from HiRes.pm
406 - add jhi's copyright alongside Douglas'
407 - move HiRes.pm to lib/Time/
409 - modify HiRes.t to use $ENV{PERL_CORE}
410 - modify the original Time::HiRes version 1.20 Makefile.PL
411 to work both with Perl 5.8.0 and the new code with pre-5.8.0
412 Perls (tried with 5.6.1)
413 - tiny tweaks and updates in README and TODO
414 - bump the VERSION to 1.29
416 1.20 Wed Feb 24 21:30 1999
417 - make our usleep and ualarm substitutes into hrt_usleep
418 and hrt_ualarm. This helps static links of Perl with other
419 packages that also have usleep, etc. From
420 Ilya Zakharevich <ilya@math.ohio-state.edu>
421 - add C API stuff. From Joshua Pritikin
422 <joshua.pritikin@db.com>
423 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
424 - hopefully correct "-lc" fix for SCO.
427 1.20 adds a platform neutral set of C accessible routines if you are
428 running 5.005+. All other changes are packaging changes and build
429 fixes(?) for statically linked Perl, SCO, and VMS.
431 1.19 Tue Sep 29 22:30 1998
432 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
433 <seb@stian.demon.co.uk>
434 - change GIMME_V to GIMME to help people with older versions of
436 - fix Win32 version of gettimeofday(). It didn't affect anything,
437 but it confuses people reading the code when the return value
438 is backwards (0 is success).
439 - fix Makefile.PL (more) so that detection of gettimeofday is
442 1.19 has better VMS support.
444 1.18 Mon Jul 6 22:40 1998
445 - add usleep() for Win32.
446 - fix Makefile.PL to fix reported HP/UX feature where unresolved
447 externals still cause an executable to be generated (though no
448 x bit set). Thanks to David Kozinn for report and explanation.
449 Problems with the fix are mine :)
451 1.18 has limited Win32 support (no ualarm). Added usleep for Win32.
452 Probably buggy. I'm sure I'll hear.
454 1.17 Wed Jul 1 20:10 1998
455 - fix setitimer calls so microseconds is not more than 1000000.
456 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
457 - make Win32. We only get gettimeofday (the select hack doesn't
458 seem to work on my Win95 system).
459 - fix test 4 on 01test.t. add test to see if time() and
460 Time::HiRes::time() are close.
462 1.16 Wed Nov 12 21:05 1997
463 - add missing EXTEND in new gettimeofday scalar code.
465 1.16+ should be closer to building out of the box on Linux. Thanks
466 to Gisle Aas for patches, and the ualarm equivalent using setitimer.
468 If your underlying operating system doesn't implement ualarm(), then
469 a fake using setitimer() will be made. If the OS is missing usleep(),
470 a fake one using select() will be made. If a fake can't be made for
471 either ualarm() or usleep(), then the corresponding Perl function will
472 not be available. If the OS is missing gettimeofday(), you will get
473 unresolved externals, either at link- or run-time.
475 This is an improvement; the package used to not even build if
476 you were missing any of these bits. Roderick Schertler
478 <roderick@argon.org> did all the conditional compilation stuff,
479 look at HiRes.pm and the test suites; it's good educational reading.
481 1.15 Mon Nov 10 21:30 1997
482 - HiRes.pm: update pod. Provided by Gisle Aas.
483 - HiRes.xs: if gettimeofday() called in scalar context, do
484 something more useful than before. Provided by Gisle Aas.
485 - README: tell of xsubpp '-nolinenumber' woes. thanks to
486 Edward Henigin <ed@texas.net> for pointing out the problem.
488 1.14 Wed Nov 5 9:40 1997
489 - Makefile.PL: look for setitimer
490 - HiRes.xs: if missing ualarm, but we have setitimer, make up
491 our own setitimer. These were provided by Gisle Aas.
493 1.13 Tue Nov 4 23:30 1997
494 - Makefile.PL: fix autodetect mechanism to do try linking in addition
495 to just compiling; should fix Linux build problem. Fix was provided
498 1.12 Sun Oct 12 12:00:00 1997
499 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
500 you may need to comment this back out if you have an older xsubpp.
501 - HiRes.xs: set PROTOTYPES: DISABLE
503 1.11 Fri Sep 05 16:00:00 1997
505 Had some line commented out that shouldn't have been (testing
508 Previous version was corrupted.
510 1.10 Thu May 22 20:20:00 1997
511 - HiRes.xs, HiRes.pm, t/*:
512 - only compile what we have OS support for (or can
514 - only test what we compiled
515 - gross improvement to the test suite
517 This work was all done by Roderick Schertler
518 <roderick@argon.org>. If you run Linux or
519 one of the other ualarm-less platoforms, and you like this
520 module, let Roderick know; without him, it still wouldn't
521 be working on those boxes...
522 - Makefile.PL: figure out what routines the OS has and
523 only build what we need. These bits were written by Jarkko
524 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
526 1.02 Mon Dec 30 08:00:00 1996
527 - HiRes.pm: update documentation to say what to do when missing
528 ualarm() and friends.
529 - README: update to warn that ualarm() and friends need to exist
531 1.01 Fri Oct 17 08:00:00 1996
532 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
533 - HiRes.pm: put blank line between __END__ and =head1 so that
536 1.00 Tue Sep 03 13:00:00 1996
537 - original version; created by h2xs 1.16