1 Revision history for Perl extension Time::HiRes.
4 - Gisle noticed a mistake (using HAS_NANOSLEEP) in 1.79
7 - try nanosleep for emulating usleep -- may help in some weird
8 embedded realtime places which have nanosleep but neither usleep
9 nor select nor poll (doesn't have to be weird embedded realtime
10 place, though -- in many places usleep is nanosleep anyway)
11 - try poll for emulating usleep -- this may help some obscure/old
12 SVR4 places that have neither usleep nor select
13 - a redundant test guard in HiRes.t
16 - ITIMER_VIRTUAL detection in HiRes.t had problems (that we cannot
17 in the general case fail already at 'use' phase is suboptimal)
18 - fixes to the documentation of clock_gettime() and clock_getres()
21 - add support for the POSIX clock_gettime() and clock_getres(),
22 if available, either as library calls or as syscalls
23 - be more defensive about missing functionality: break out
24 early (during 'use') if no e.g. clock_getres() is available,
25 and protect our back by trapping those cases also in HiRes.xs
26 - the test added in 1.76 could cause an endless loop e.g. in Solaris,
27 due to mixing of sleep() and alarm() (bad programmer, no cookie!)
30 - testing for nanosleep had wrong logic which caused nanosleep
31 to become undefined for e.g. Mac OS X
32 - added a test for a core dump that was introduced by Perl 5.8.0
33 safe signals and was fixed for the time of 5.8.1 (one report of
34 the core dump was [perl #20920]), the test skipped pre-5.8.1.
35 - *cough* s/unanosleep/nanosleep/g; *cough*
38 - installation patch from Gisle Aas: in Perls 5.8.x and later
39 use MakeMaker INSTALLDIRS value of 'perl' instead of 'site'.
42 - [cpan #14608] Solaris 8 perl 5.005_03 File::Spec module does not have method rel2abs
43 (the workaround is not to use rel2abs, should not be necessary)
44 - [cpan #14642] U2time wrongly exported on the C API
45 (patch supplied by the reporter, SALVA@cpan.org)
46 - add release dates to Changes
49 - Time::HiRes::nanosleep support for Solaris [PATCH]
50 (POSIX::uname() not available if building with core perl,
51 from Gisle Aas, via perl5-porters, perl change #25295)
54 - going back to the 1.68 loader setup (using DynaLoader)
55 since too many weird things starting breaking
56 - fix a typo in José Auguste-Etienne's name
59 - a thinko in the nanosleep() detection
60 - move more changes stuff from the README to Changes
61 - add -w to the Makefile.PL
64 - oops in 1.69 about @ISA (not affecting anything but silly)
65 - add copyright 2005 to HiRes.pm
66 - add copyright and license to HiRes.xs
67 - add copyrights 2003, 2004, 2005 to README
70 - actually run a test for nanosleep
71 (if there is no $Config{d_nanosleep}) since e.g. in AIX 4.2
72 it seems that one can link in nanosleep() but then calling
73 it fails instantly and sets errno to ENOSYS (Not implemented).
74 This may be fixable in the AIX case by figuring out the right
75 (realtime POSIX?) libs and whatnot, but in the general case
76 running a real test case is better. (Of course, this change
77 will no doubt run into portability problems because of the
78 execution step...) Note that because of hysterical raisins
79 most Perls do NOT have $Config{d_nanosleep} (scanning for
80 it by Configure would in many platforms require linking in
81 things like -lrt, which would in many platforms be a bad idea
83 (from José Auguste-Etienne)
84 - support XSLoader also since it's much faster
86 - add SEE ALSO (BSD::Resource and Time::TAI64)
89 - somehow 1.67 had a lot of doubled lines (a major cut-and-paste
90 error suspected), but miraculously it still worked since the
91 doubling took place below the __END__ token
92 - undef Pause() before defining it to avoid redefinition warnings
93 during compilation in case perl.h had already defined Pause()
94 (part of perl change #24271)
98 - (internal) don't ignore the return value of gettimeofday()
99 - (external) return undef or an empty if the C gettimeofday() fails
100 (affects Time::HiRes gettimeofday() and the hires time())
104 - fix the 'hierachy' typo in Makefile.PL [rt.cpan.org #8492]
105 - should now build in Solaris [rt.cpan.org #7165] (since 1.64)
106 - should now build in Cygwin [rt.cpan.org #7535] (since 1.64)
107 - 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
108 has become rather unlikely
111 - one should not mix u?alarm and sleep (the tests modified
112 by 1.65, #12 and #13, hung in Solaris), now we just busy
113 loop executing an empty block
114 - in the documentation underline the unspecificity of mixing
116 - small spelling fixes
119 - regenerate ppport.h with Devel::PPPort 3.03,
120 now the MY_CXT_CLONE is defined in ppport.h,
121 we no more need to do that.
123 - the test #12 would often hang in sigsuspend() (at least that's
124 where Mac OS X' ktrace shows it hanging). With the sleep()s
125 changed to sleep(1)s, the tests still pass but no hang after
126 a few hundred repeats.
129 - Win32 and any ithread build: ppport.h didn't define
130 MY_CXT_CLONE, which seems to be a Time-HiResism.
133 - Skip testing if under PERL_CORE and Time::HiRes has not
134 been Configured (from Marcus Holland-Moritz, core change
136 - Use ppport.h generated by Devel::PPPort 3.01,
137 allowing cutting away our own portability code.
138 - Don't use $ENV{PERL_CORE} for < 5.6.0.
139 - Don't use "for my $i" for <= 5.003.
140 - Don't use Pause() for <= 5.003.
141 - Can't use newSVpvf for <= 5.003.
142 (most of the changes from Marcus)
145 - Win32: reset reading from the performance counters every
146 five minutes to better track wall clock time (thanks to
147 PC timers being often quite bad), should help long-running
151 - Win32: Patch from Steve Hay
152 [PATCH] Re: [perl #30755] [Win32] Different results from Time::HiRes::gettimeofdayunder the debugger
153 to [perl #30755] reported by Nigel Sandever
155 - Cygwin: Use the Win32 recalibration code also in Cygwin if the
156 <w32api/windows.h> APIs are available. Cygwin testing by
157 Yitzchak Scott-Thoennes.
159 - Solaris: use -lposix4 to get nanosleep for Solaris 2.6,
160 after that keep using -lrt, patch from Alan Burlison,
161 bug reported in [cpan #7165]
164 - Change the Win32 recalibration limit to 0.5 seconds and tweak
165 the documentation to blather less about the gory details of the
166 Win32 implementation and more about the complications in general
167 of meddling with the system clock.
170 - Document the 1.57 change better.
173 - Win32/Cygwin/MinGW: if the performance counter drifts by more
174 than two seconds from the system clock (due to ntp adjustments,
175 for example), recalibrate our internal counter: from Jan Dubois,
176 based on [cpan #5933] by Jerry D. Hedden.
179 - Give a clearer message if the tests timeout (perl change #22253)
180 - Don't use /tmp or its moral equivalents (perl bug #15036,
184 - Windows: mingw32 patch from Mike Pomraning (use Perl's Const64()
185 instead of VC-specific i64 suffix)
188 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
191 - Windows: higher resolution time() by using the Windows
192 performance counter API, from Jan Dubois and Anton Shcherbinin.
193 The exact new higher resolution depends on the hardware,
194 but it should be quite a bit better than using the basic
198 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
199 the subtest 18. No known analysis nor fix, but added
200 an alarm (that requires fork() and alarm()) to the test.
203 - doc tweaks from mjd (perl change #20456)
204 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
207 - add a message (for non-core builds) to Makefile.PL about
208 the LC_ALL=C workaround
209 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
210 - the nanosleep() probe was broken (perl change #20061)
211 - use existence instead of definedness for feature probes
213 - MPE/iX tweak (perl change #20042)
214 - do not use HAS_NANOSLEEP (perl change #19898)
217 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
218 - OS/2 can always mix subsecond sleeps with signals
219 (part of perl change #19789)
222 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
226 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
227 - unnecessary (nanosleep is in libc anyway)
228 - harmful (-lrt slows down execution)
229 - incompatible (with many distributions' pthreads)
232 - do not create files in blib directories under core
233 (perl change #19160, from rgs)
234 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
237 - guarantee that $xdefine in HiRes.t is always defined
238 (perl change #19109, from IlyaZ)
239 - a cleaner way to detect PERL_CORE (perl change #19111,
243 - add hints/irix.pl to turn off overly POSIX flags that
244 cause hide struct timespec to be hidden (and compilation
245 to fail) (bleadperl change #19085)
246 - documentation tweaks
249 - add c:/temp to the list of temp directories to probe
250 so that cygwin (and win*?) builds are happy. This was
251 needed at least in my cygwin 1.3.20/w2k setup.
254 - modernize the constants code (from Nicholas Clark)
257 - At some point the ability to figure our the correct incdir
258 for EXTERN.h (either a core perl build, or an installed perl)
259 had broken (which lead into all test compiles failing with
260 a core perl build, but thanks to the robustness of Makefile.PL
261 nothing of this was visible). The brokenness seemed to be
262 caused by $ENV{PERL_CORE} not being on for core builds?
263 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
264 right, and both styles of build should work again.
267 - Nicholas Clark noticed that the my_catdir() emulation function
268 was broken (which means that we didn't really work for Perls
270 - inspired by fixing the above made the whole Makefile.PL -w
272 - tightened up the Makefile.PL output, less whitespace
275 - fix from Craig Berry for better building in VMS with PERL_CORE
278 - no functional changes
279 - move lib/Time/HiRes.pm as Hires.pm
280 - libraries scanning was slightly broken (always scanned
281 for a library even when $Config{libs} already had it)
284 - Ray Zimmerman ran into a race condition in Mac OS X.
285 A 0.01-second alarm fired before the test expected.
286 The test first slept indefinitely (blocking for signals)
287 and only after that tested for the signal having been sent.
288 Since the signal had already been sent, the test #12 never
289 completed. The solution: test first, then block.
290 - default to being silent on all probing attempts, set the
291 environment variable VERBOSE to a true value to see the
292 details (the probing command and the possible errors)
295 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
296 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
297 cannot be upgraded from CPAN to override the 5.8.0 version
299 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
300 because the include directories did not adjust themselves
301 if $ENV{PERL_CORE} (Hugo van der Sanden)
302 - add documentation about the restart of select() under alarm()
305 - small documentation tweaks
309 - better VMS operation (Craig Berry)
312 - our time machine is accelerating: now works with Perl 5.004_01
313 (tried with 5.003_07 and 5.002 but I get segmentation faults
314 from running the Makefile.PL with those in Tru64 4.0D)
317 - backward compatibility (pre-5.6.0) tweaks:
318 - no XSLoader in 5.00503, use DynaLoader instead
319 - no SvPV_nolen, either
320 - no PerlProc_pause(), either
321 - now tested with 5.00404 and 5.00503
322 - Makefile.PL requires 5.00404 (no more 5.002)
323 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
324 (this means that one can mix subsecond sleeps with alarms)
325 - because of nanosleep we probe for -lrt and -lposix4
326 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
327 is available by exportable constants Time::HiRes::d_func
328 (since older Perl do not have them in %Config, and even
329 5.8.0 does not probe for nanosleep)
332 - backward compatibility (pre-5.6.1) tweaks:
333 - define NV if no NVTYPE
334 - define IVdf if needed (note: the Devel::PPPort
335 in 5.8.0 does not try hard hard enough since
336 the IVSIZE might not be defined)
337 - define NVgf if needed
338 - grab the typemap from 5.8.0 for the NV stuff
340 1.31 and 1.32 add more backward compatibility (now all the way
341 back to Perl 5.00404), and using nanosleep() (if available) for
346 - release 1.29_02 as 1.30
348 1.30 adds all the changes made during the Perl 5.6->5.7->5.8
349 development cycle. Most notably portability across platforms has been
350 enhanced, and the interval timers (setitimer, getitimer) have been
351 added. Note that the version of Time::HiRes that is included in Perl
352 5.8.0 calls itself 1.20_00, but it is equivalent to this Time::HiRes
353 version. Note also that in 1.30 Wegscheid turns over the maintenance
354 to Jarkko Hietaniemi.
358 - fix a silly unclosed comment typo in HiRes.xs
359 - document and export REALTIME_REALPROF (Solaris)
363 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
364 (need to patch this also in Perl 5.[89])
365 - remove CVS revision log from HiRes.xs
369 The following numbered patches refer to the Perl 5.7 changes,
370 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
372 - 17558: Add #!./perl to the .t
373 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
374 - 16198: political correctness, from Simon Cozens
375 - 15857: doc tweaks, from Jarkko Hietaniemi
376 - 15593: optimization in .xs, from Paul Green
377 - 14892: pod fixes, from Robin Barker
378 - 14100: VOS fixes, from Paul Green
379 - 13422: XS segfault, from Marc Lehmann
380 - 13378: whether select() gets restarted on signals, depends
381 - 13354: timing constraints, again, from Andy Dougherty
382 - 13278: can't do subecond alarms with ualarm;
383 break out early if alarms do not seem to be working
384 - 13266: test relaxation (cygwin gets lower hires
385 times than lores ones)
386 - 12846: protect against high load, from Jarkko Hietaniemi
387 - 12837: HiRes.t VMS tweak, from Craig A. Berry
388 - 12797: HiRes.t VMS tweak, from Charles Lane
389 - 12769: HiRes.t VMS tweak, from Craig A. Berry
390 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
391 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
392 - 12692: alarm() ain't gonna work if ualarm() ain't,
393 from Gurusamy Sarathy
394 - 12680: minor VMS tweak, from Charles Lane
395 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
396 - 12609: croak on negative time, from Jarkko Hietaniemi
397 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
398 - 12594: MacOS Classic timeofday, from Chris Nandor
399 - 12473: allow for more than one second for sleep() and usleep()
400 - 12458: test tuning, relax timing constraints,
401 from Jarkko Hietaniemi
402 - 12449: make sleep() and usleep() to return the number
403 of seconds and microseconds actually slept (analogously
404 with the builtin sleep()), also make usleep() croak if
405 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
406 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
407 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
408 - 12196: use ftime() on Win32, from Artur Bergman
409 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
410 - 12105: use GetSystemTime() on Win32, from Artur Bergman
411 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
412 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
413 - 11797: problem in HiRes.t, from John P. Linderman
414 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
415 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
416 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
417 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
418 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
419 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
421 - 10942: MPE/IX test tweaks, from Mark Bixby
422 - 10784: unnecessary pod2man calls, from Andy Dougherty
423 - 10354: ext/ + -Wall, from Doug MacEachern
424 - 10320: fix the BOOT section to call myU2time correctly
425 - 10317: correct casting for AIX< from H. Merijn Brand
426 - 10119: document that the core time() may be rounding, not truncating
427 - 10118: test fix, from John Peacock
428 - 9988: long =item, from Robin Barker
429 - 9714: correct test output
430 - 9708: test also the scalar aspect of getitimer()
431 - 9705: Add interval timers (setitimer, getitimer)
432 - 9692: do not require at least 5.005 using XS
434 The following changes were made on top of the changes
435 made for Time::HiRes during the Perl 5.7 development
436 cycle that culminated in the release of Perl 5.8.0.
438 - add "require 5.005" to the Makefile.PL
439 - remove the REVISION section (CVS log) from HiRes.pm
440 - add jhi's copyright alongside Douglas'
441 - move HiRes.pm to lib/Time/
443 - modify HiRes.t to use $ENV{PERL_CORE}
444 - modify the original Time::HiRes version 1.20 Makefile.PL
445 to work both with Perl 5.8.0 and the new code with pre-5.8.0
446 Perls (tried with 5.6.1)
447 - tiny tweaks and updates in README and TODO
448 - bump the VERSION to 1.29
450 1.20 Wed Feb 24 21:30 1999
451 - make our usleep and ualarm substitutes into hrt_usleep
452 and hrt_ualarm. This helps static links of Perl with other
453 packages that also have usleep, etc. From
454 Ilya Zakharevich <ilya@math.ohio-state.edu>
455 - add C API stuff. From Joshua Pritikin
456 <joshua.pritikin@db.com>
457 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
458 - hopefully correct "-lc" fix for SCO.
461 1.20 adds a platform neutral set of C accessible routines if you are
462 running 5.005+. All other changes are packaging changes and build
463 fixes(?) for statically linked Perl, SCO, and VMS.
465 1.19 Tue Sep 29 22:30 1998
466 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
467 <seb@stian.demon.co.uk>
468 - change GIMME_V to GIMME to help people with older versions of
470 - fix Win32 version of gettimeofday(). It didn't affect anything,
471 but it confuses people reading the code when the return value
472 is backwards (0 is success).
473 - fix Makefile.PL (more) so that detection of gettimeofday is
476 1.19 has better VMS support.
478 1.18 Mon Jul 6 22:40 1998
479 - add usleep() for Win32.
480 - fix Makefile.PL to fix reported HP/UX feature where unresolved
481 externals still cause an executable to be generated (though no
482 x bit set). Thanks to David Kozinn for report and explanation.
483 Problems with the fix are mine :)
485 1.18 has limited Win32 support (no ualarm). Added usleep for Win32.
486 Probably buggy. I'm sure I'll hear.
488 1.17 Wed Jul 1 20:10 1998
489 - fix setitimer calls so microseconds is not more than 1000000.
490 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
491 - make Win32. We only get gettimeofday (the select hack doesn't
492 seem to work on my Win95 system).
493 - fix test 4 on 01test.t. add test to see if time() and
494 Time::HiRes::time() are close.
496 1.16 Wed Nov 12 21:05 1997
497 - add missing EXTEND in new gettimeofday scalar code.
499 1.16+ should be closer to building out of the box on Linux. Thanks
500 to Gisle Aas for patches, and the ualarm equivalent using setitimer.
502 If your underlying operating system doesn't implement ualarm(), then
503 a fake using setitimer() will be made. If the OS is missing usleep(),
504 a fake one using select() will be made. If a fake can't be made for
505 either ualarm() or usleep(), then the corresponding Perl function will
506 not be available. If the OS is missing gettimeofday(), you will get
507 unresolved externals, either at link- or run-time.
509 This is an improvement; the package used to not even build if
510 you were missing any of these bits. Roderick Schertler
512 <roderick@argon.org> did all the conditional compilation stuff,
513 look at HiRes.pm and the test suites; it's good educational reading.
515 1.15 Mon Nov 10 21:30 1997
516 - HiRes.pm: update pod. Provided by Gisle Aas.
517 - HiRes.xs: if gettimeofday() called in scalar context, do
518 something more useful than before. Provided by Gisle Aas.
519 - README: tell of xsubpp '-nolinenumber' woes. thanks to
520 Edward Henigin <ed@texas.net> for pointing out the problem.
522 1.14 Wed Nov 5 9:40 1997
523 - Makefile.PL: look for setitimer
524 - HiRes.xs: if missing ualarm, but we have setitimer, make up
525 our own setitimer. These were provided by Gisle Aas.
527 1.13 Tue Nov 4 23:30 1997
528 - Makefile.PL: fix autodetect mechanism to do try linking in addition
529 to just compiling; should fix Linux build problem. Fix was provided
532 1.12 Sun Oct 12 12:00:00 1997
533 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
534 you may need to comment this back out if you have an older xsubpp.
535 - HiRes.xs: set PROTOTYPES: DISABLE
537 1.11 Fri Sep 05 16:00:00 1997
539 Had some line commented out that shouldn't have been (testing
542 Previous version was corrupted.
544 1.10 Thu May 22 20:20:00 1997
545 - HiRes.xs, HiRes.pm, t/*:
546 - only compile what we have OS support for (or can
548 - only test what we compiled
549 - gross improvement to the test suite
551 This work was all done by Roderick Schertler
552 <roderick@argon.org>. If you run Linux or
553 one of the other ualarm-less platoforms, and you like this
554 module, let Roderick know; without him, it still wouldn't
555 be working on those boxes...
556 - Makefile.PL: figure out what routines the OS has and
557 only build what we need. These bits were written by Jarkko
558 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
560 1.02 Mon Dec 30 08:00:00 1996
561 - HiRes.pm: update documentation to say what to do when missing
562 ualarm() and friends.
563 - README: update to warn that ualarm() and friends need to exist
565 1.01 Fri Oct 17 08:00:00 1996
566 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
567 - HiRes.pm: put blank line between __END__ and =head1 so that
570 1.00 Tue Sep 03 13:00:00 1996
571 - original version; created by h2xs 1.16