1 Revision history for the Perl extension Time::HiRes.
4 - lost VMS test skippage from Craig Berry
5 - reformat the test code a little
8 - I got the sense of the QNX test the wrong way in an attempt
9 to generalize it for future
12 - casting fixes from Robin Barker for g++ and 64bitint
13 - in QNX skip the itimer tests because though the API
14 is there, the implementation isn't, from Matt Kraai
15 - raise the dead man timer to 180 seconds for really
17 - elaborate the UTF-8 locale warning from Makefile.PL
20 - [rt.cpan.org #29747]: Build failure with perl 5.005_05
21 Fixed by regenerating the ppport.h using Devel::PPPort 3.13.
24 - t/HiRes.t failed in Perl 5.6.2,
25 "action is not of type POSIX::SigAction at t/HiRes.t line 318",
26 reported and fixed by Anton Berezin, the reason was faulty
27 use of sigaction() when restoring the old action.
30 - with bleadperl in VMS the HiRes.t overrun the maximum number
31 of deferred signals because the libc SIGALRM was not strong
32 enough to interrupt select(), and select() got restarted every
33 time, solution is to use POSIX::SigAction if available.
34 A fix from Craig Berry (not 100% there, but helps).
35 - allow for more measuring noise for ualarm() tests 35..37
38 - nanosleep() and clock_nanosleep() detection and use were
39 quite broken; in Linux -lrt needed; fixes from Zefram
40 - [internal] slightly cleaner building of $DEFINE in Makefile.PL,
41 should avoid double/conflicting -D flags
44 - allow 10% of slop in test #14 (testing difference between
45 CORE::time() and Time::HiRes::time()), there seem to be often
46 transient failures from Perl smoke builds on this test
50 - use int main(int argc, char **argv) consistently in Makefile.PL,
52 [rt.cpan.org #23868] nanosleep not detected under Mac OS 10.3.9 starting with Time::HiRes 1.96
53 - if someone still has the locale-broken Perl 5.8.0,
54 suggest that they upgrade their Perl
57 - restore the -DATLEASTFIVEOHOHFIVE, Win32 needed it still
60 - upgrade to ppport.h 3.10_02
61 - remove the -DATLEASTFIVEOHOHFIVE
62 - use the ppport.h PL_ppaddr, PL_statcache, PL_laststatval
63 - use the ppport.h aTHXR for calling Perl stat()
64 - switch into four-digit version since 2.0 is coming up
65 awfully fast but not feeling like a major rewrite
68 - 1.95 broke building in Win32 (since pp_stat is not exported),
69 figured out how to call an op directly in 5.005 (use Perl_ppaddr
71 - backport to Perl 5.004_05 (requires using statcache
72 and laststatval instead of PL_statcache and PL_laststatval)
73 (also checked to work in 5.005_04, 5.6.1, and 5.8.8 with threads)
76 - 1.95 broke builds for threaded Perls, rt.cpan.org tickets:
77 [rt.cpan.org #23694] Time::HiRes fails tests on Solaris and Perl 5.6.1
78 [rt.cpan.org #23712] Time-HiRes 1.95 Fails make on AIX 5.2 with Perl 5.8.8
79 [rt.cpan.org #23730] Time::HiRes 1.95 fails make on MacOS X 10.3.9/perl 5.8.8
80 - use main() prototype consistently in Makefile.PL
83 - integrate core change #29180: Silence VC++ compiler warnings
85 - do not use PL_ppaddr in stat() because that is not available
87 - regenerate fallback/*.inc for older Perls without
88 ExtUtils::Constant because of d_hires_stat, resolves
89 [rt.cpan.org #23694] Time::HiRes fails tests on Solaris and Perl 5.6.1
90 - Make Makefile.PL more defensive against false PERL_CORE
93 - file timestamps oddities seen: the atime and mtime
94 can be out of sync (modify first and read second can leave
95 atime < mtime) and mtime can be subsecond while atime is not.
96 So make the test more forgiving.
99 - the ualarm() tests (34-37) assumed that ualarm(N)
100 could never alarm in less than N seconds, widened
101 the acceptable relative range to 0.9..1.5. Addresses
102 [rt.cpan.org #22090] and [rt.cpan.org #22091].
104 - skip the stat() tests in cygwin and win32, because
105 if run on FAT the timestamp granularity is only 2 seconds.
106 Any good way to detect (cygwin or win32) whether we are
107 being run on NTFS or anywhere with better timestamps?
108 Addresses [rt.cpan.org #22089] and [rt.cpan.org #22098].
111 - scan for subsecond resolution timestamps in struct stat,
112 some known possibilities:
114 (1) struct timespec st_atimespec;
115 st_atimespec.tv_nsec;
120 (4) timestruc_t st_atim;
125 If something like this is found, one can do
128 my @stat = Time::HiRes::stat();
130 or even override the standard stat():
132 use Time::HiRes qw(stat);
134 to get the stat() timestamps
136 my ($atime, $mtime, $ctime) = @stat[8, 9, 10];
138 with subsecond resolution (assuming both the operating
139 system and the filesystem support that kind of thing).
141 Contributions for more systems (especially non-UNIX,
142 e.g. but not limited to: Win32, VMS, OS/2) gladly accepted.
143 (also more UNIX variants welcome: HP-UX? IRIX?)
145 Thanks to H.Merijn Brand, John Peacock, and Craig
146 Berry for brave beta testing.
149 - ualarm() in SuSE 10.1 was overflowing after ~4.2 seconds,
150 possibly due to a glibc bug/feature (suspected overflow at
151 2**32 microseconds?), workaround by using the setitimer()
152 implementation of ualarm() if either useconds or
153 interval > 999_999 (this case seems to vary between systems:
154 are useconds more than 999_999 for ualarm() defined or not)
155 Added more ualarm() tests to catch various overflow points,
156 hopefully no problems in various platforms.
157 (The problem report by Mark Seger and Jon Paul Sullivan of HP.)
160 - tweak still needed for Const64(), from Jerry Hedden
161 - get a freshly generated ppport.h
162 - update Copyright years
165 - Const64() already appends an 'LL' (or i64), so provide LL and i64
166 forms for the IV_1E[679] (effects Win32 and Cygwin), reported by
168 - the Changes entry for 1.88 talked about [IN]V_1[679],
172 - clean up the g++ warnings in HiRes.xs, all of them
173 about mixing integer and floating point, introduce
174 constants IV_1E[679] and NV_1E[679]
177 - [rt.cpan.org #17442] 'make test' frequently fails under
178 Cygwin Perl v5.8.8, reported and patched by J. R. Hedden
179 (two race condition bugs in the END block in the case the
180 main process dies before the timer process, unearthed
181 by a bug in Cygwin ualarm)
184 - HiRes.t:s/ok 32/ok 33/, from Dominic Dunlop
185 - tighten up the clock() test marginally by requiring non-negative
186 - clock_nanosleep() and clock() doc tweaks
189 - the interface to clock_nanosleep() is more natural
190 when it is like (hires) time() (instead of like nanosleep),
191 and the .xs implementation of clock_nanosleep() in 1.84
193 - the semantics of clock() are not quite so silly as I thought,
194 but still somewhat odd, documented as such
195 - additional enhancements to the clock() documentation
196 - add test for clock_nanosleep() (I cannot test this
197 since none of my systems have the function)
198 - add test for clock()
201 - add clock() which returns the processor time in
202 (floating point) seconds since an arbitrary era
203 - add clock_nanosleep() which suspends the current
204 thread until either absolute time or for relative time
205 - [rt.cpan.org #16486] printf missing value in HiRes.t
206 - add constants CLOCKS_PER_SEC, CLOCK_SOFTTIME, TIMER_ABSTIME
210 - has_symbol() was wrong since e.g. ITIMER_VIRTUAL is exported
211 via @EXPORT_OK even when it is not available. This is heinous.
212 @EXPORT_OK should be determined at Makefile.PL time.
213 - be more lenient is testing clock_gettime(): allow more slop,
214 and retry up to three times, sleeping a random nap between
216 - human months are one-based (noticed by Anton Berezin)
219 - CLOCK_REALTIME is an enum value (of the clockid_t enum)
220 in HP-UX (and might be so elsewhere, too), debugged by
222 - include const-c.inc as late as possible (from Randy Kobes,
223 [rt.cpan.org #15552] to avoid undefined usleep() on Win32
226 - try to be more robust and consistent in the detection of
227 CLOCK_REALTIME and ITIMER_VIRTUAL in HiRes.t: the proper
232 eval 'import Time::HiRes qw($symbol)';
233 return 0 unless $@ eq '';
234 return exists ${"Time::HiRes::$symbol"};
241 in the test. All these moves are needed because
243 1) one cannot directly do eval 'Time::HiRes::FOO_BAR'
244 because FOO_BAR might have a true value of zero
245 (or in the general case an empty string or even undef)
247 2) In case FOO_BAR is not available in this platform,
248 &FOO_BAR avoids the bareword warning
250 - wait more (1.5 seconds instead of 0.1) for the CLOCK_REALTIME test
251 but expect the 'customary' slop of 0.20 instead of 0.25
252 - fixed inside a comment HAS_POLL -> TIME_HIRES_NANOSLEEP
253 - at the end of HiRest.t tell how close we were to termination
256 - Gisle noticed a mistake (using HAS_NANOSLEEP) in 1.79
259 - try nanosleep for emulating usleep -- may help in some weird
260 embedded realtime places which have nanosleep but neither usleep
261 nor select nor poll (doesn't have to be weird embedded realtime
262 place, though -- in many places usleep is nanosleep anyway)
263 - try poll for emulating usleep -- this may help some obscure/old
264 SVR4 places that have neither usleep nor select
265 - a redundant test guard in HiRes.t
268 - ITIMER_VIRTUAL detection in HiRes.t had problems (that we cannot
269 in the general case fail already at 'use' phase is suboptimal)
270 - fixes to the documentation of clock_gettime() and clock_getres()
273 - add support for the POSIX clock_gettime() and clock_getres(),
274 if available, either as library calls or as syscalls
275 - be more defensive about missing functionality: break out
276 early (during 'use') if no e.g. clock_getres() is available,
277 and protect our back by trapping those cases also in HiRes.xs
278 - the test added in 1.76 could cause an endless loop e.g. in Solaris,
279 due to mixing of sleep() and alarm() (bad programmer, no cookie!)
282 - testing for nanosleep had wrong logic which caused nanosleep
283 to become undefined for e.g. Mac OS X
284 - added a test for a core dump that was introduced by Perl 5.8.0
285 safe signals and was fixed for the time of 5.8.1 (one report of
286 the core dump was [perl #20920]), the test skipped pre-5.8.1.
287 - *cough* s/unanosleep/nanosleep/g; *cough*
290 - installation patch from Gisle Aas: in Perls 5.8.x and later
291 use MakeMaker INSTALLDIRS value of 'perl' instead of 'site'.
294 - [cpan #14608] Solaris 8 perl 5.005_03 File::Spec module does not have method rel2abs
295 (the workaround is not to use rel2abs, should not be necessary)
296 - [cpan #14642] U2time wrongly exported on the C API
297 (patch supplied by the reporter, SALVA@cpan.org)
298 - add release dates to Changes
301 - Time::HiRes::nanosleep support for Solaris [PATCH]
302 (POSIX::uname() not available if building with core perl,
303 from Gisle Aas, via perl5-porters, perl change #25295)
306 - going back to the 1.68 loader setup (using DynaLoader)
307 since too many weird things starting breaking
308 - fix a typo in José Auguste-Etienne's name
311 - a thinko in the nanosleep() detection
312 - move more changes stuff from the README to Changes
313 - add -w to the Makefile.PL
316 - oops in 1.69 about @ISA (not affecting anything but silly)
317 - add copyright 2005 to HiRes.pm
318 - add copyright and license to HiRes.xs
319 - add copyrights 2003, 2004, 2005 to README
322 - actually run a test for nanosleep
323 (if there is no $Config{d_nanosleep}) since e.g. in AIX 4.2
324 it seems that one can link in nanosleep() but then calling
325 it fails instantly and sets errno to ENOSYS (Not implemented).
326 This may be fixable in the AIX case by figuring out the right
327 (realtime POSIX?) libs and whatnot, but in the general case
328 running a real test case is better. (Of course, this change
329 will no doubt run into portability problems because of the
330 execution step...) Note that because of hysterical raisins
331 most Perls do NOT have $Config{d_nanosleep} (scanning for
332 it by Configure would in many platforms require linking in
333 things like -lrt, which would in many platforms be a bad idea
335 (from José Auguste-Etienne)
336 - support XSLoader also since it's much faster
337 (from Alexey Tourbin)
338 - add SEE ALSO (BSD::Resource and Time::TAI64)
341 - somehow 1.67 had a lot of doubled lines (a major cut-and-paste
342 error suspected), but miraculously it still worked since the
343 doubling took place below the __END__ token
344 - undef Pause() before defining it to avoid redefinition warnings
345 during compilation in case perl.h had already defined Pause()
346 (part of perl change #24271)
350 - (internal) don't ignore the return value of gettimeofday()
351 - (external) return undef or an empty if the C gettimeofday() fails
352 (affects Time::HiRes gettimeofday() and the hires time())
356 - fix the 'hierachy' typo in Makefile.PL [rt.cpan.org #8492]
357 - should now build in Solaris [rt.cpan.org #7165] (since 1.64)
358 - should now build in Cygwin [rt.cpan.org #7535] (since 1.64)
359 - close also [rt.cpan.org #5933] "Time::HiRes::time does not
360 pick up time adjustments like ntp" since ever reproducing it
361 (and therefore verifying a possible fix) in the same environment
362 has become rather unlikely
365 - one should not mix u?alarm and sleep (the tests modified
366 by 1.65, #12 and #13, hung in Solaris), now we just busy
367 loop executing an empty block
368 - in the documentation underline the unspecificity of mixing
370 - small spelling fixes
373 - regenerate ppport.h with Devel::PPPort 3.03,
374 now the MY_CXT_CLONE is defined in ppport.h,
375 we no more need to do that.
377 - the test #12 would often hang in sigsuspend() (at least that's
378 where Mac OS X' ktrace shows it hanging). With the sleep()s
379 changed to sleep(1)s, the tests still pass but no hang after
380 a few hundred repeats.
383 - Win32 and any ithread build: ppport.h didn't define
384 MY_CXT_CLONE, which seems to be a Time-HiRes-ism.
387 - Skip testing if under PERL_CORE and Time::HiRes has not
388 been Configured (from Marcus Holland-Moritz, core change
390 - Use ppport.h generated by Devel::PPPort 3.01,
391 allowing cutting away our own portability code.
392 - Don't use $ENV{PERL_CORE} for < 5.6.0.
393 - Don't use "for my $i" for <= 5.003.
394 - Don't use Pause() for <= 5.003.
395 - Can't use newSVpvf for <= 5.003.
396 (most of the changes from Marcus)
399 - Win32: reset reading from the performance counters every
400 five minutes to better track wall clock time (thanks to
401 PC timers being often quite bad), should help long-running
405 - Win32: Patch from Steve Hay
406 [PATCH] Re: [perl #30755] [Win32] Different results from Time::HiRes::gettimeofdayunder the debugger
407 to [perl #30755] reported by Nigel Sandever
409 - Cygwin: Use the Win32 recalibration code also in Cygwin if the
410 <w32api/windows.h> APIs are available. Cygwin testing by
411 Yitzchak Scott-Thoennes.
413 - Solaris: use -lposix4 to get nanosleep for Solaris 2.6,
414 after that keep using -lrt, patch from Alan Burlison,
415 bug reported in [cpan #7165]
418 - Change the Win32 recalibration limit to 0.5 seconds and tweak
419 the documentation to blather less about the gory details of the
420 Win32 implementation and more about the complications in general
421 of meddling with the system clock.
424 - Document the 1.57 change better.
427 - Win32/Cygwin/MinGW: if the performance counter drifts by more
428 than two seconds from the system clock (due to ntp adjustments,
429 for example), recalibrate our internal counter: from Jan Dubois,
430 based on [cpan #5933] by Jerry D. Hedden.
433 - Give a clearer message if the tests timeout (perl change #22253)
434 - Don't use /tmp or its moral equivalents (perl bug #15036,
438 - Windows: mingw32 patch from Mike Pomraning (use Perl's Const64()
439 instead of VC-specific i64 suffix)
442 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
445 - Windows: higher resolution time() by using the Windows
446 performance counter API, from Jan Dubois and Anton Shcherbinin.
447 The exact new higher resolution depends on the hardware,
448 but it should be quite a bit better than using the basic
452 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
453 the subtest 18. No known analysis nor fix, but added
454 an alarm (that requires fork() and alarm()) to the test.
457 - doc tweaks from mjd (perl change #20456)
458 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
461 - add a message (for non-core builds) to Makefile.PL about
462 the LC_ALL=C workaround
463 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
464 - the nanosleep() probe was broken (perl change #20061)
465 - use existence instead of definedness for feature probes
467 - MPE/iX tweak (perl change #20042)
468 - do not use HAS_NANOSLEEP (perl change #19898)
471 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
472 - OS/2 can always mix subsecond sleeps with signals
473 (part of perl change #19789)
476 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
480 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
481 - unnecessary (nanosleep is in libc anyway)
482 - harmful (-lrt slows down execution)
483 - incompatible (with many distributions' pthreads)
486 - do not create files in blib directories under core
487 (perl change #19160, from rgs)
488 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
491 - guarantee that $xdefine in HiRes.t is always defined
492 (perl change #19109, from IlyaZ)
493 - a cleaner way to detect PERL_CORE (perl change #19111,
497 - add hints/irix.pl to turn off overly POSIX flags that
498 cause hide struct timespec to be hidden (and compilation
499 to fail) (bleadperl change #19085)
500 - documentation tweaks
503 - add c:/temp to the list of temp directories to probe
504 so that cygwin (and win*?) builds are happy. This was
505 needed at least in my cygwin 1.3.20/w2k setup.
508 - modernize the constants code (from Nicholas Clark)
511 - At some point the ability to figure our the correct incdir
512 for EXTERN.h (either a core perl build, or an installed perl)
513 had broken (which lead into all test compiles failing with
514 a core perl build, but thanks to the robustness of Makefile.PL
515 nothing of this was visible). The brokenness seemed to be
516 caused by $ENV{PERL_CORE} not being on for core builds?
517 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
518 right, and both styles of build should work again.
521 - Nicholas Clark noticed that the my_catdir() emulation function
522 was broken (which means that we didn't really work for Perls
524 - inspired by fixing the above made the whole Makefile.PL -w
526 - tightened up the Makefile.PL output, less whitespace
529 - fix from Craig Berry for better building in VMS with PERL_CORE
532 - no functional changes
533 - move lib/Time/HiRes.pm as Hires.pm
534 - libraries scanning was slightly broken (always scanned
535 for a library even when $Config{libs} already had it)
538 - Ray Zimmerman ran into a race condition in Mac OS X.
539 A 0.01-second alarm fired before the test expected.
540 The test first slept indefinitely (blocking for signals)
541 and only after that tested for the signal having been sent.
542 Since the signal had already been sent, the test #12 never
543 completed. The solution: test first, then block.
544 - default to being silent on all probing attempts, set the
545 environment variable VERBOSE to a true value to see the
546 details (the probing command and the possible errors)
549 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
550 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
551 cannot be upgraded from CPAN to override the 5.8.0 version
553 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
554 because the include directories did not adjust themselves
555 if $ENV{PERL_CORE} (Hugo van der Sanden)
556 - add documentation about the restart of select() under alarm()
559 - small documentation tweaks
563 - better VMS operation (Craig Berry)
566 - our time machine is accelerating: now works with Perl 5.004_01
567 (tried with 5.003_07 and 5.002 but I get segmentation faults
568 from running the Makefile.PL with those in Tru64 4.0D)
571 - backward compatibility (pre-5.6.0) tweaks:
572 - no XSLoader in 5.00503, use DynaLoader instead
573 - no SvPV_nolen, either
574 - no PerlProc_pause(), either
575 - now tested with 5.00404 and 5.00503
576 - Makefile.PL requires 5.00404 (no more 5.002)
577 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
578 (this means that one can mix subsecond sleeps with alarms)
579 - because of nanosleep we probe for -lrt and -lposix4
580 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
581 is available by exportable constants Time::HiRes::d_func
582 (since older Perl do not have them in %Config, and even
583 5.8.0 does not probe for nanosleep)
586 - backward compatibility (pre-5.6.1) tweaks:
587 - define NV if no NVTYPE
588 - define IVdf if needed (note: the Devel::PPPort
589 in 5.8.0 does not try hard hard enough since
590 the IVSIZE might not be defined)
591 - define NVgf if needed
592 - grab the typemap from 5.8.0 for the NV stuff
594 1.31 and 1.32 add more backward compatibility (now all the way
595 back to Perl 5.00404), and using nanosleep() (if available) for
600 - release 1.29_02 as 1.30
602 1.30 adds all the changes made during the Perl 5.6->5.7->5.8
603 development cycle. Most notably portability across platforms has been
604 enhanced, and the interval timers (setitimer, getitimer) have been
605 added. Note that the version of Time::HiRes that is included in Perl
606 5.8.0 calls itself 1.20_00, but it is equivalent to this Time::HiRes
607 version. Note also that in 1.30 Wegscheid turns over the maintenance
608 to Jarkko Hietaniemi.
612 - fix a silly unclosed comment typo in HiRes.xs
613 - document and export REALTIME_REALPROF (Solaris)
617 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
618 (need to patch this also in Perl 5.[89])
619 - remove CVS revision log from HiRes.xs
623 The following numbered patches refer to the Perl 5.7 changes,
624 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
626 - 17558: Add #!./perl to the .t
627 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
628 - 16198: political correctness, from Simon Cozens
629 - 15857: doc tweaks, from Jarkko Hietaniemi
630 - 15593: optimization in .xs, from Paul Green
631 - 14892: pod fixes, from Robin Barker
632 - 14100: VOS fixes, from Paul Green
633 - 13422: XS segfault, from Marc Lehmann
634 - 13378: whether select() gets restarted on signals, depends
635 - 13354: timing constraints, again, from Andy Dougherty
636 - 13278: can't do subsecond alarms with ualarm;
637 break out early if alarms do not seem to be working
638 - 13266: test relaxation (cygwin gets lower hires
639 times than lores ones)
640 - 12846: protect against high load, from Jarkko Hietaniemi
641 - 12837: HiRes.t VMS tweak, from Craig A. Berry
642 - 12797: HiRes.t VMS tweak, from Charles Lane
643 - 12769: HiRes.t VMS tweak, from Craig A. Berry
644 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
645 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
646 - 12692: alarm() ain't gonna work if ualarm() ain't,
647 from Gurusamy Sarathy
648 - 12680: minor VMS tweak, from Charles Lane
649 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
650 - 12609: croak on negative time, from Jarkko Hietaniemi
651 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
652 - 12594: MacOS Classic timeofday, from Chris Nandor
653 - 12473: allow for more than one second for sleep() and usleep()
654 - 12458: test tuning, relax timing constraints,
655 from Jarkko Hietaniemi
656 - 12449: make sleep() and usleep() to return the number
657 of seconds and microseconds actually slept (analogously
658 with the builtin sleep()), also make usleep() croak if
659 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
660 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
661 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
662 - 12196: use ftime() on Win32, from Artur Bergman
663 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
664 - 12105: use GetSystemTime() on Win32, from Artur Bergman
665 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
666 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
667 - 11797: problem in HiRes.t, from John P. Linderman
668 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
669 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
670 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
671 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
672 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
673 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
675 - 10942: MPE/IX test tweaks, from Mark Bixby
676 - 10784: unnecessary pod2man calls, from Andy Dougherty
677 - 10354: ext/ + -Wall, from Doug MacEachern
678 - 10320: fix the BOOT section to call myU2time correctly
679 - 10317: correct casting for AIX< from H. Merijn Brand
680 - 10119: document that the core time() may be rounding, not truncating
681 - 10118: test fix, from John Peacock
682 - 9988: long =item, from Robin Barker
683 - 9714: correct test output
684 - 9708: test also the scalar aspect of getitimer()
685 - 9705: Add interval timers (setitimer, getitimer)
686 - 9692: do not require at least 5.005 using XS
688 The following changes were made on top of the changes
689 made for Time::HiRes during the Perl 5.7 development
690 cycle that culminated in the release of Perl 5.8.0.
692 - add "require 5.005" to the Makefile.PL
693 - remove the REVISION section (CVS log) from HiRes.pm
694 - add jhi's copyright alongside Douglas'
695 - move HiRes.pm to lib/Time/
697 - modify HiRes.t to use $ENV{PERL_CORE}
698 - modify the original Time::HiRes version 1.20 Makefile.PL
699 to work both with Perl 5.8.0 and the new code with pre-5.8.0
700 Perls (tried with 5.6.1)
701 - tiny tweaks and updates in README and TODO
702 - bump the VERSION to 1.29
704 1.20 Wed Feb 24 21:30 1999
705 - make our usleep and ualarm substitutes into hrt_usleep
706 and hrt_ualarm. This helps static links of Perl with other
707 packages that also have usleep, etc. From
708 Ilya Zakharevich <ilya@math.ohio-state.edu>
709 - add C API stuff. From Joshua Pritikin
710 <joshua.pritikin@db.com>
711 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
712 - hopefully correct "-lc" fix for SCO.
715 1.20 adds a platform neutral set of C accessible routines if you are
716 running 5.005+. All other changes are packaging changes and build
717 fixes(?) for statically linked Perl, SCO, and VMS.
719 1.19 Tue Sep 29 22:30 1998
720 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
721 <seb@stian.demon.co.uk>
722 - change GIMME_V to GIMME to help people with older versions of
724 - fix Win32 version of gettimeofday(). It didn't affect anything,
725 but it confuses people reading the code when the return value
726 is backwards (0 is success).
727 - fix Makefile.PL (more) so that detection of gettimeofday is
730 1.19 has better VMS support.
732 1.18 Mon Jul 6 22:40 1998
733 - add usleep() for Win32.
734 - fix Makefile.PL to fix reported HP/UX feature where unresolved
735 externals still cause an executable to be generated (though no
736 x bit set). Thanks to David Kozinn for report and explanation.
737 Problems with the fix are mine :)
739 1.18 has limited Win32 support (no ualarm). Added usleep for Win32.
740 Probably buggy. I'm sure I'll hear.
742 1.17 Wed Jul 1 20:10 1998
743 - fix setitimer calls so microseconds is not more than 1000000.
744 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
745 - make Win32. We only get gettimeofday (the select hack doesn't
746 seem to work on my Win95 system).
747 - fix test 4 on 01test.t. add test to see if time() and
748 Time::HiRes::time() are close.
750 1.16 Wed Nov 12 21:05 1997
751 - add missing EXTEND in new gettimeofday scalar code.
753 1.16+ should be closer to building out of the box on Linux. Thanks
754 to Gisle Aas for patches, and the ualarm equivalent using setitimer.
756 If your underlying operating system doesn't implement ualarm(), then
757 a fake using setitimer() will be made. If the OS is missing usleep(),
758 a fake one using select() will be made. If a fake can't be made for
759 either ualarm() or usleep(), then the corresponding Perl function will
760 not be available. If the OS is missing gettimeofday(), you will get
761 unresolved externals, either at link- or run-time.
763 This is an improvement; the package used to not even build if
764 you were missing any of these bits. Roderick Schertler
766 <roderick@argon.org> did all the conditional compilation stuff,
767 look at HiRes.pm and the test suites; it's good educational reading.
769 1.15 Mon Nov 10 21:30 1997
770 - HiRes.pm: update pod. Provided by Gisle Aas.
771 - HiRes.xs: if gettimeofday() called in scalar context, do
772 something more useful than before. Provided by Gisle Aas.
773 - README: tell of xsubpp '-nolinenumber' woes. thanks to
774 Edward Henigin <ed@texas.net> for pointing out the problem.
776 1.14 Wed Nov 5 9:40 1997
777 - Makefile.PL: look for setitimer
778 - HiRes.xs: if missing ualarm, but we have setitimer, make up
779 our own setitimer. These were provided by Gisle Aas.
781 1.13 Tue Nov 4 23:30 1997
782 - Makefile.PL: fix autodetect mechanism to do try linking in addition
783 to just compiling; should fix Linux build problem. Fix was provided
786 1.12 Sun Oct 12 12:00:00 1997
787 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
788 you may need to comment this back out if you have an older xsubpp.
789 - HiRes.xs: set PROTOTYPES: DISABLE
791 1.11 Fri Sep 05 16:00:00 1997
793 Had some line commented out that shouldn't have been (testing
796 Previous version was corrupted.
798 1.10 Thu May 22 20:20:00 1997
799 - HiRes.xs, HiRes.pm, t/*:
800 - only compile what we have OS support for (or can
802 - only test what we compiled
803 - gross improvement to the test suite
805 This work was all done by Roderick Schertler
806 <roderick@argon.org>. If you run Linux or
807 one of the other ualarm-less platforms, and you like this
808 module, let Roderick know; without him, it still wouldn't
809 be working on those boxes...
810 - Makefile.PL: figure out what routines the OS has and
811 only build what we need. These bits were written by Jarkko
812 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
814 1.02 Mon Dec 30 08:00:00 1996
815 - HiRes.pm: update documentation to say what to do when missing
816 ualarm() and friends.
817 - README: update to warn that ualarm() and friends need to exist
819 1.01 Fri Oct 17 08:00:00 1996
820 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
821 - HiRes.pm: put blank line between __END__ and =head1 so that
824 1.00 Tue Sep 03 13:00:00 1996
825 - original version; created by h2xs 1.16