1 Revision history for the Perl extension Time::HiRes.
4 - Silly me: Makefile.PL does need to accept arguments other than mine.
5 Some testing frameworks obviously do this.
6 - Add retrying for tests 34..37, which are the most commonly
7 failing tests. If this helps, consider extending the retry
8 framework to all the tests. [Inspired by Slaven Rezic,
9 [rt.cpan.org #34711] Occasional failures of test 35 or 36 (FreeBSD)]
12 - Under Perl 5.6.* NVgf needs to be "g", reported by Zefram,
13 it seems that ppport.h 3.13 gets this wrong.
14 - remove the check in Makefile.PL for 5.7.2, shouldn't be
15 (a) necessary (b) relevant
16 - add logic to Makefile.PL to skip configure/write Makefile
17 step if the "xdefine" file already exists, indicating that
18 the configure step has already been done, one can still
19 force (re)configure by "perl Makefile.PL configure",
20 or of course by "make clean && perl Makefile.PL".
23 - for alarm() and ualarm() [Perl] prefer setitimer() [C]
24 instead of ualarm() [C] since ualarm() [C] cannot portably
25 (and standards-compliantly) be used for more than 999_999
26 microseconds (rt.cpan.org #34655)
27 - it seems that HP-UX has started (at least in 11.31 ia64)
28 #defining the CLOCK_REALTIME et alia (instead of having
30 - document all the diagnostics
33 - move the sub tick in the test file back to where it used to be
34 - in the "consider upgrading" message recommend at least Perl 5.8.8
35 and make the message to appear only for 5.8.0 since 5.8.1 and
36 later have the problem fixed
37 - VOS tweak for Makefile (core perl change #33259)
38 - since the test #17 seems to fail often, relax its limits a bit
41 - lost VMS test skippage from Craig Berry
42 - reformat the test code a little
45 - I got the sense of the QNX test the wrong way in an attempt
46 to generalize it for future
49 - casting fixes from Robin Barker for g++ and 64bitint
50 - in QNX skip the itimer tests because though the API
51 is there, the implementation isn't, from Matt Kraai
52 - raise the dead man timer to 180 seconds for really
54 - elaborate the UTF-8 locale warning from Makefile.PL
57 - [rt.cpan.org #29747]: Build failure with perl 5.005_05
58 Fixed by regenerating the ppport.h using Devel::PPPort 3.13.
61 - t/HiRes.t failed in Perl 5.6.2,
62 "action is not of type POSIX::SigAction at t/HiRes.t line 318",
63 reported and fixed by Anton Berezin, the reason was faulty
64 use of sigaction() when restoring the old action.
67 - with bleadperl in VMS the HiRes.t overrun the maximum number
68 of deferred signals because the libc SIGALRM was not strong
69 enough to interrupt select(), and select() got restarted every
70 time, solution is to use POSIX::SigAction if available.
71 A fix from Craig Berry (not 100% there, but helps).
72 - allow for more measuring noise for ualarm() tests 35..37
75 - nanosleep() and clock_nanosleep() detection and use were
76 quite broken; in Linux -lrt needed; fixes from Zefram
77 - [internal] slightly cleaner building of $DEFINE in Makefile.PL,
78 should avoid double/conflicting -D flags
81 - allow 10% of slop in test #14 (testing difference between
82 CORE::time() and Time::HiRes::time()), there seem to be often
83 transient failures from Perl smoke builds on this test
87 - use int main(int argc, char **argv) consistently in Makefile.PL,
89 [rt.cpan.org #23868] nanosleep not detected under Mac OS 10.3.9 starting with Time::HiRes 1.96
90 - if someone still has the locale-broken Perl 5.8.0,
91 suggest that they upgrade their Perl
94 - restore the -DATLEASTFIVEOHOHFIVE, Win32 needed it still
97 - upgrade to ppport.h 3.10_02
98 - remove the -DATLEASTFIVEOHOHFIVE
99 - use the ppport.h PL_ppaddr, PL_statcache, PL_laststatval
100 - use the ppport.h aTHXR for calling Perl stat()
101 - switch into four-digit version since 2.0 is coming up
102 awfully fast but not feeling like a major rewrite
105 - 1.95 broke building in Win32 (since pp_stat is not exported),
106 figured out how to call an op directly in 5.005 (use Perl_ppaddr
107 instead of PL_ppaddr)
108 - backport to Perl 5.004_05 (requires using statcache
109 and laststatval instead of PL_statcache and PL_laststatval)
110 (also checked to work in 5.005_04, 5.6.1, and 5.8.8 with threads)
113 - 1.95 broke builds for threaded Perls, rt.cpan.org tickets:
114 [rt.cpan.org #23694] Time::HiRes fails tests on Solaris and Perl 5.6.1
115 [rt.cpan.org #23712] Time-HiRes 1.95 Fails make on AIX 5.2 with Perl 5.8.8
116 [rt.cpan.org #23730] Time::HiRes 1.95 fails make on MacOS X 10.3.9/perl 5.8.8
117 - use main() prototype consistently in Makefile.PL
120 - integrate core change #29180: Silence VC++ compiler warnings
122 - do not use PL_ppaddr in stat() because that is not available
124 - regenerate fallback/*.inc for older Perls without
125 ExtUtils::Constant because of d_hires_stat, resolves
126 [rt.cpan.org #23694] Time::HiRes fails tests on Solaris and Perl 5.6.1
127 - Make Makefile.PL more defensive against false PERL_CORE
130 - file timestamps oddities seen: the atime and mtime
131 can be out of sync (modify first and read second can leave
132 atime < mtime) and mtime can be subsecond while atime is not.
133 So make the test more forgiving.
136 - the ualarm() tests (34-37) assumed that ualarm(N)
137 could never alarm in less than N seconds, widened
138 the acceptable relative range to 0.9..1.5. Addresses
139 [rt.cpan.org #22090] and [rt.cpan.org #22091].
141 - skip the stat() tests in cygwin and win32, because
142 if run on FAT the timestamp granularity is only 2 seconds.
143 Any good way to detect (cygwin or win32) whether we are
144 being run on NTFS or anywhere with better timestamps?
145 Addresses [rt.cpan.org #22089] and [rt.cpan.org #22098].
148 - scan for subsecond resolution timestamps in struct stat,
149 some known possibilities:
151 (1) struct timespec st_atimespec;
152 st_atimespec.tv_nsec;
157 (4) timestruc_t st_atim;
162 If something like this is found, one can do
165 my @stat = Time::HiRes::stat();
167 or even override the standard stat():
169 use Time::HiRes qw(stat);
171 to get the stat() timestamps
173 my ($atime, $mtime, $ctime) = @stat[8, 9, 10];
175 with subsecond resolution (assuming both the operating
176 system and the filesystem support that kind of thing).
178 Contributions for more systems (especially non-UNIX,
179 e.g. but not limited to: Win32, VMS, OS/2) gladly accepted.
180 (also more UNIX variants welcome: HP-UX? IRIX?)
182 Thanks to H.Merijn Brand, John Peacock, and Craig
183 Berry for brave beta testing.
186 - ualarm() in SuSE 10.1 was overflowing after ~4.2 seconds,
187 possibly due to a glibc bug/feature (suspected overflow at
188 2**32 microseconds?), workaround by using the setitimer()
189 implementation of ualarm() if either useconds or
190 interval > 999_999 (this case seems to vary between systems:
191 are useconds more than 999_999 for ualarm() defined or not)
192 Added more ualarm() tests to catch various overflow points,
193 hopefully no problems in various platforms.
194 (The problem report by Mark Seger and Jon Paul Sullivan of HP.)
197 - tweak still needed for Const64(), from Jerry Hedden
198 - get a freshly generated ppport.h
199 - update Copyright years
202 - Const64() already appends an 'LL' (or i64), so provide LL and i64
203 forms for the IV_1E[679] (effects Win32 and Cygwin), reported by
205 - the Changes entry for 1.88 talked about [IN]V_1[679],
209 - clean up the g++ warnings in HiRes.xs, all of them
210 about mixing integer and floating point, introduce
211 constants IV_1E[679] and NV_1E[679]
214 - [rt.cpan.org #17442] 'make test' frequently fails under
215 Cygwin Perl v5.8.8, reported and patched by J. R. Hedden
216 (two race condition bugs in the END block in the case the
217 main process dies before the timer process, unearthed
218 by a bug in Cygwin ualarm)
221 - HiRes.t:s/ok 32/ok 33/, from Dominic Dunlop
222 - tighten up the clock() test marginally by requiring non-negative
223 - clock_nanosleep() and clock() doc tweaks
226 - the interface to clock_nanosleep() is more natural
227 when it is like (hires) time() (instead of like nanosleep),
228 and the .xs implementation of clock_nanosleep() in 1.84
230 - the semantics of clock() are not quite so silly as I thought,
231 but still somewhat odd, documented as such
232 - additional enhancements to the clock() documentation
233 - add test for clock_nanosleep() (I cannot test this
234 since none of my systems have the function)
235 - add test for clock()
238 - add clock() which returns the processor time in
239 (floating point) seconds since an arbitrary era
240 - add clock_nanosleep() which suspends the current
241 thread until either absolute time or for relative time
242 - [rt.cpan.org #16486] printf missing value in HiRes.t
243 - add constants CLOCKS_PER_SEC, CLOCK_SOFTTIME, TIMER_ABSTIME
247 - has_symbol() was wrong since e.g. ITIMER_VIRTUAL is exported
248 via @EXPORT_OK even when it is not available. This is heinous.
249 @EXPORT_OK should be determined at Makefile.PL time.
250 - be more lenient is testing clock_gettime(): allow more slop,
251 and retry up to three times, sleeping a random nap between
253 - human months are one-based (noticed by Anton Berezin)
256 - CLOCK_REALTIME is an enum value (of the clockid_t enum)
257 in HP-UX (and might be so elsewhere, too), debugged by
259 - include const-c.inc as late as possible (from Randy Kobes,
260 [rt.cpan.org #15552] to avoid undefined usleep() on Win32
263 - try to be more robust and consistent in the detection of
264 CLOCK_REALTIME and ITIMER_VIRTUAL in HiRes.t: the proper
269 eval 'import Time::HiRes qw($symbol)';
270 return 0 unless $@ eq '';
271 return exists ${"Time::HiRes::$symbol"};
278 in the test. All these moves are needed because
280 1) one cannot directly do eval 'Time::HiRes::FOO_BAR'
281 because FOO_BAR might have a true value of zero
282 (or in the general case an empty string or even undef)
284 2) In case FOO_BAR is not available in this platform,
285 &FOO_BAR avoids the bareword warning
287 - wait more (1.5 seconds instead of 0.1) for the CLOCK_REALTIME test
288 but expect the 'customary' slop of 0.20 instead of 0.25
289 - fixed inside a comment HAS_POLL -> TIME_HIRES_NANOSLEEP
290 - at the end of HiRest.t tell how close we were to termination
293 - Gisle noticed a mistake (using HAS_NANOSLEEP) in 1.79
296 - try nanosleep for emulating usleep -- may help in some weird
297 embedded realtime places which have nanosleep but neither usleep
298 nor select nor poll (doesn't have to be weird embedded realtime
299 place, though -- in many places usleep is nanosleep anyway)
300 - try poll for emulating usleep -- this may help some obscure/old
301 SVR4 places that have neither usleep nor select
302 - a redundant test guard in HiRes.t
305 - ITIMER_VIRTUAL detection in HiRes.t had problems (that we cannot
306 in the general case fail already at 'use' phase is suboptimal)
307 - fixes to the documentation of clock_gettime() and clock_getres()
310 - add support for the POSIX clock_gettime() and clock_getres(),
311 if available, either as library calls or as syscalls
312 - be more defensive about missing functionality: break out
313 early (during 'use') if no e.g. clock_getres() is available,
314 and protect our back by trapping those cases also in HiRes.xs
315 - the test added in 1.76 could cause an endless loop e.g. in Solaris,
316 due to mixing of sleep() and alarm() (bad programmer, no cookie!)
319 - testing for nanosleep had wrong logic which caused nanosleep
320 to become undefined for e.g. Mac OS X
321 - added a test for a core dump that was introduced by Perl 5.8.0
322 safe signals and was fixed for the time of 5.8.1 (one report of
323 the core dump was [perl #20920]), the test skipped pre-5.8.1.
324 - *cough* s/unanosleep/nanosleep/g; *cough*
327 - installation patch from Gisle Aas: in Perls 5.8.x and later
328 use MakeMaker INSTALLDIRS value of 'perl' instead of 'site'.
331 - [cpan #14608] Solaris 8 perl 5.005_03 File::Spec module does not have method rel2abs
332 (the workaround is not to use rel2abs, should not be necessary)
333 - [cpan #14642] U2time wrongly exported on the C API
334 (patch supplied by the reporter, SALVA@cpan.org)
335 - add release dates to Changes
338 - Time::HiRes::nanosleep support for Solaris [PATCH]
339 (POSIX::uname() not available if building with core perl,
340 from Gisle Aas, via perl5-porters, perl change #25295)
343 - going back to the 1.68 loader setup (using DynaLoader)
344 since too many weird things starting breaking
345 - fix a typo in José Auguste-Etienne's name
348 - a thinko in the nanosleep() detection
349 - move more changes stuff from the README to Changes
350 - add -w to the Makefile.PL
353 - oops in 1.69 about @ISA (not affecting anything but silly)
354 - add copyright 2005 to HiRes.pm
355 - add copyright and license to HiRes.xs
356 - add copyrights 2003, 2004, 2005 to README
359 - actually run a test for nanosleep
360 (if there is no $Config{d_nanosleep}) since e.g. in AIX 4.2
361 it seems that one can link in nanosleep() but then calling
362 it fails instantly and sets errno to ENOSYS (Not implemented).
363 This may be fixable in the AIX case by figuring out the right
364 (realtime POSIX?) libs and whatnot, but in the general case
365 running a real test case is better. (Of course, this change
366 will no doubt run into portability problems because of the
367 execution step...) Note that because of hysterical raisins
368 most Perls do NOT have $Config{d_nanosleep} (scanning for
369 it by Configure would in many platforms require linking in
370 things like -lrt, which would in many platforms be a bad idea
372 (from José Auguste-Etienne)
373 - support XSLoader also since it's much faster
374 (from Alexey Tourbin)
375 - add SEE ALSO (BSD::Resource and Time::TAI64)
378 - somehow 1.67 had a lot of doubled lines (a major cut-and-paste
379 error suspected), but miraculously it still worked since the
380 doubling took place below the __END__ token
381 - undef Pause() before defining it to avoid redefinition warnings
382 during compilation in case perl.h had already defined Pause()
383 (part of perl change #24271)
387 - (internal) don't ignore the return value of gettimeofday()
388 - (external) return undef or an empty if the C gettimeofday() fails
389 (affects Time::HiRes gettimeofday() and the hires time())
393 - fix the 'hierachy' typo in Makefile.PL [rt.cpan.org #8492]
394 - should now build in Solaris [rt.cpan.org #7165] (since 1.64)
395 - should now build in Cygwin [rt.cpan.org #7535] (since 1.64)
396 - close also [rt.cpan.org #5933] "Time::HiRes::time does not
397 pick up time adjustments like ntp" since ever reproducing it
398 (and therefore verifying a possible fix) in the same environment
399 has become rather unlikely
402 - one should not mix u?alarm and sleep (the tests modified
403 by 1.65, #12 and #13, hung in Solaris), now we just busy
404 loop executing an empty block
405 - in the documentation underline the unspecificity of mixing
407 - small spelling fixes
410 - regenerate ppport.h with Devel::PPPort 3.03,
411 now the MY_CXT_CLONE is defined in ppport.h,
412 we no more need to do that.
414 - the test #12 would often hang in sigsuspend() (at least that's
415 where Mac OS X' ktrace shows it hanging). With the sleep()s
416 changed to sleep(1)s, the tests still pass but no hang after
417 a few hundred repeats.
420 - Win32 and any ithread build: ppport.h didn't define
421 MY_CXT_CLONE, which seems to be a Time-HiRes-ism.
424 - Skip testing if under PERL_CORE and Time::HiRes has not
425 been Configured (from Marcus Holland-Moritz, core change
427 - Use ppport.h generated by Devel::PPPort 3.01,
428 allowing cutting away our own portability code.
429 - Don't use $ENV{PERL_CORE} for < 5.6.0.
430 - Don't use "for my $i" for <= 5.003.
431 - Don't use Pause() for <= 5.003.
432 - Can't use newSVpvf for <= 5.003.
433 (most of the changes from Marcus)
436 - Win32: reset reading from the performance counters every
437 five minutes to better track wall clock time (thanks to
438 PC timers being often quite bad), should help long-running
442 - Win32: Patch from Steve Hay
443 [PATCH] Re: [perl #30755] [Win32] Different results from Time::HiRes::gettimeofdayunder the debugger
444 to [perl #30755] reported by Nigel Sandever
446 - Cygwin: Use the Win32 recalibration code also in Cygwin if the
447 <w32api/windows.h> APIs are available. Cygwin testing by
448 Yitzchak Scott-Thoennes.
450 - Solaris: use -lposix4 to get nanosleep for Solaris 2.6,
451 after that keep using -lrt, patch from Alan Burlison,
452 bug reported in [cpan #7165]
455 - Change the Win32 recalibration limit to 0.5 seconds and tweak
456 the documentation to blather less about the gory details of the
457 Win32 implementation and more about the complications in general
458 of meddling with the system clock.
461 - Document the 1.57 change better.
464 - Win32/Cygwin/MinGW: if the performance counter drifts by more
465 than two seconds from the system clock (due to ntp adjustments,
466 for example), recalibrate our internal counter: from Jan Dubois,
467 based on [cpan #5933] by Jerry D. Hedden.
470 - Give a clearer message if the tests timeout (perl change #22253)
471 - Don't use /tmp or its moral equivalents (perl bug #15036,
475 - Windows: mingw32 patch from Mike Pomraning (use Perl's Const64()
476 instead of VC-specific i64 suffix)
479 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
482 - Windows: higher resolution time() by using the Windows
483 performance counter API, from Jan Dubois and Anton Shcherbinin.
484 The exact new higher resolution depends on the hardware,
485 but it should be quite a bit better than using the basic
489 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
490 the subtest 18. No known analysis nor fix, but added
491 an alarm (that requires fork() and alarm()) to the test.
494 - doc tweaks from mjd (perl change #20456)
495 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
498 - add a message (for non-core builds) to Makefile.PL about
499 the LC_ALL=C workaround
500 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
501 - the nanosleep() probe was broken (perl change #20061)
502 - use existence instead of definedness for feature probes
504 - MPE/iX tweak (perl change #20042)
505 - do not use HAS_NANOSLEEP (perl change #19898)
508 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
509 - OS/2 can always mix subsecond sleeps with signals
510 (part of perl change #19789)
513 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
517 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
518 - unnecessary (nanosleep is in libc anyway)
519 - harmful (-lrt slows down execution)
520 - incompatible (with many distributions' pthreads)
523 - do not create files in blib directories under core
524 (perl change #19160, from rgs)
525 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
528 - guarantee that $xdefine in HiRes.t is always defined
529 (perl change #19109, from IlyaZ)
530 - a cleaner way to detect PERL_CORE (perl change #19111,
534 - add hints/irix.pl to turn off overly POSIX flags that
535 cause hide struct timespec to be hidden (and compilation
536 to fail) (bleadperl change #19085)
537 - documentation tweaks
540 - add c:/temp to the list of temp directories to probe
541 so that cygwin (and win*?) builds are happy. This was
542 needed at least in my cygwin 1.3.20/w2k setup.
545 - modernize the constants code (from Nicholas Clark)
548 - At some point the ability to figure our the correct incdir
549 for EXTERN.h (either a core perl build, or an installed perl)
550 had broken (which lead into all test compiles failing with
551 a core perl build, but thanks to the robustness of Makefile.PL
552 nothing of this was visible). The brokenness seemed to be
553 caused by $ENV{PERL_CORE} not being on for core builds?
554 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
555 right, and both styles of build should work again.
558 - Nicholas Clark noticed that the my_catdir() emulation function
559 was broken (which means that we didn't really work for Perls
561 - inspired by fixing the above made the whole Makefile.PL -w
563 - tightened up the Makefile.PL output, less whitespace
566 - fix from Craig Berry for better building in VMS with PERL_CORE
569 - no functional changes
570 - move lib/Time/HiRes.pm as Hires.pm
571 - libraries scanning was slightly broken (always scanned
572 for a library even when $Config{libs} already had it)
575 - Ray Zimmerman ran into a race condition in Mac OS X.
576 A 0.01-second alarm fired before the test expected.
577 The test first slept indefinitely (blocking for signals)
578 and only after that tested for the signal having been sent.
579 Since the signal had already been sent, the test #12 never
580 completed. The solution: test first, then block.
581 - default to being silent on all probing attempts, set the
582 environment variable VERBOSE to a true value to see the
583 details (the probing command and the possible errors)
586 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
587 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
588 cannot be upgraded from CPAN to override the 5.8.0 version
590 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
591 because the include directories did not adjust themselves
592 if $ENV{PERL_CORE} (Hugo van der Sanden)
593 - add documentation about the restart of select() under alarm()
596 - small documentation tweaks
600 - better VMS operation (Craig Berry)
603 - our time machine is accelerating: now works with Perl 5.004_01
604 (tried with 5.003_07 and 5.002 but I get segmentation faults
605 from running the Makefile.PL with those in Tru64 4.0D)
608 - backward compatibility (pre-5.6.0) tweaks:
609 - no XSLoader in 5.00503, use DynaLoader instead
610 - no SvPV_nolen, either
611 - no PerlProc_pause(), either
612 - now tested with 5.00404 and 5.00503
613 - Makefile.PL requires 5.00404 (no more 5.002)
614 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
615 (this means that one can mix subsecond sleeps with alarms)
616 - because of nanosleep we probe for -lrt and -lposix4
617 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
618 is available by exportable constants Time::HiRes::d_func
619 (since older Perl do not have them in %Config, and even
620 5.8.0 does not probe for nanosleep)
623 - backward compatibility (pre-5.6.1) tweaks:
624 - define NV if no NVTYPE
625 - define IVdf if needed (note: the Devel::PPPort
626 in 5.8.0 does not try hard hard enough since
627 the IVSIZE might not be defined)
628 - define NVgf if needed
629 - grab the typemap from 5.8.0 for the NV stuff
631 1.31 and 1.32 add more backward compatibility (now all the way
632 back to Perl 5.00404), and using nanosleep() (if available) for
637 - release 1.29_02 as 1.30
639 1.30 adds all the changes made during the Perl 5.6->5.7->5.8
640 development cycle. Most notably portability across platforms has been
641 enhanced, and the interval timers (setitimer, getitimer) have been
642 added. Note that the version of Time::HiRes that is included in Perl
643 5.8.0 calls itself 1.20_00, but it is equivalent to this Time::HiRes
644 version. Note also that in 1.30 Wegscheid turns over the maintenance
645 to Jarkko Hietaniemi.
649 - fix a silly unclosed comment typo in HiRes.xs
650 - document and export REALTIME_REALPROF (Solaris)
654 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
655 (need to patch this also in Perl 5.[89])
656 - remove CVS revision log from HiRes.xs
660 The following numbered patches refer to the Perl 5.7 changes,
661 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
663 - 17558: Add #!./perl to the .t
664 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
665 - 16198: political correctness, from Simon Cozens
666 - 15857: doc tweaks, from Jarkko Hietaniemi
667 - 15593: optimization in .xs, from Paul Green
668 - 14892: pod fixes, from Robin Barker
669 - 14100: VOS fixes, from Paul Green
670 - 13422: XS segfault, from Marc Lehmann
671 - 13378: whether select() gets restarted on signals, depends
672 - 13354: timing constraints, again, from Andy Dougherty
673 - 13278: can't do subsecond alarms with ualarm;
674 break out early if alarms do not seem to be working
675 - 13266: test relaxation (cygwin gets lower hires
676 times than lores ones)
677 - 12846: protect against high load, from Jarkko Hietaniemi
678 - 12837: HiRes.t VMS tweak, from Craig A. Berry
679 - 12797: HiRes.t VMS tweak, from Charles Lane
680 - 12769: HiRes.t VMS tweak, from Craig A. Berry
681 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
682 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
683 - 12692: alarm() ain't gonna work if ualarm() ain't,
684 from Gurusamy Sarathy
685 - 12680: minor VMS tweak, from Charles Lane
686 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
687 - 12609: croak on negative time, from Jarkko Hietaniemi
688 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
689 - 12594: MacOS Classic timeofday, from Chris Nandor
690 - 12473: allow for more than one second for sleep() and usleep()
691 - 12458: test tuning, relax timing constraints,
692 from Jarkko Hietaniemi
693 - 12449: make sleep() and usleep() to return the number
694 of seconds and microseconds actually slept (analogously
695 with the builtin sleep()), also make usleep() croak if
696 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
697 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
698 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
699 - 12196: use ftime() on Win32, from Artur Bergman
700 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
701 - 12105: use GetSystemTime() on Win32, from Artur Bergman
702 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
703 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
704 - 11797: problem in HiRes.t, from John P. Linderman
705 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
706 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
707 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
708 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
709 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
710 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
712 - 10942: MPE/IX test tweaks, from Mark Bixby
713 - 10784: unnecessary pod2man calls, from Andy Dougherty
714 - 10354: ext/ + -Wall, from Doug MacEachern
715 - 10320: fix the BOOT section to call myU2time correctly
716 - 10317: correct casting for AIX< from H. Merijn Brand
717 - 10119: document that the core time() may be rounding, not truncating
718 - 10118: test fix, from John Peacock
719 - 9988: long =item, from Robin Barker
720 - 9714: correct test output
721 - 9708: test also the scalar aspect of getitimer()
722 - 9705: Add interval timers (setitimer, getitimer)
723 - 9692: do not require at least 5.005 using XS
725 The following changes were made on top of the changes
726 made for Time::HiRes during the Perl 5.7 development
727 cycle that culminated in the release of Perl 5.8.0.
729 - add "require 5.005" to the Makefile.PL
730 - remove the REVISION section (CVS log) from HiRes.pm
731 - add jhi's copyright alongside Douglas'
732 - move HiRes.pm to lib/Time/
734 - modify HiRes.t to use $ENV{PERL_CORE}
735 - modify the original Time::HiRes version 1.20 Makefile.PL
736 to work both with Perl 5.8.0 and the new code with pre-5.8.0
737 Perls (tried with 5.6.1)
738 - tiny tweaks and updates in README and TODO
739 - bump the VERSION to 1.29
741 1.20 Wed Feb 24 21:30 1999
742 - make our usleep and ualarm substitutes into hrt_usleep
743 and hrt_ualarm. This helps static links of Perl with other
744 packages that also have usleep, etc. From
745 Ilya Zakharevich <ilya@math.ohio-state.edu>
746 - add C API stuff. From Joshua Pritikin
747 <joshua.pritikin@db.com>
748 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
749 - hopefully correct "-lc" fix for SCO.
752 1.20 adds a platform neutral set of C accessible routines if you are
753 running 5.005+. All other changes are packaging changes and build
754 fixes(?) for statically linked Perl, SCO, and VMS.
756 1.19 Tue Sep 29 22:30 1998
757 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
758 <seb@stian.demon.co.uk>
759 - change GIMME_V to GIMME to help people with older versions of
761 - fix Win32 version of gettimeofday(). It didn't affect anything,
762 but it confuses people reading the code when the return value
763 is backwards (0 is success).
764 - fix Makefile.PL (more) so that detection of gettimeofday is
767 1.19 has better VMS support.
769 1.18 Mon Jul 6 22:40 1998
770 - add usleep() for Win32.
771 - fix Makefile.PL to fix reported HP/UX feature where unresolved
772 externals still cause an executable to be generated (though no
773 x bit set). Thanks to David Kozinn for report and explanation.
774 Problems with the fix are mine :)
776 1.18 has limited Win32 support (no ualarm). Added usleep for Win32.
777 Probably buggy. I'm sure I'll hear.
779 1.17 Wed Jul 1 20:10 1998
780 - fix setitimer calls so microseconds is not more than 1000000.
781 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
782 - make Win32. We only get gettimeofday (the select hack doesn't
783 seem to work on my Win95 system).
784 - fix test 4 on 01test.t. add test to see if time() and
785 Time::HiRes::time() are close.
787 1.16 Wed Nov 12 21:05 1997
788 - add missing EXTEND in new gettimeofday scalar code.
790 1.16+ should be closer to building out of the box on Linux. Thanks
791 to Gisle Aas for patches, and the ualarm equivalent using setitimer.
793 If your underlying operating system doesn't implement ualarm(), then
794 a fake using setitimer() will be made. If the OS is missing usleep(),
795 a fake one using select() will be made. If a fake can't be made for
796 either ualarm() or usleep(), then the corresponding Perl function will
797 not be available. If the OS is missing gettimeofday(), you will get
798 unresolved externals, either at link- or run-time.
800 This is an improvement; the package used to not even build if
801 you were missing any of these bits. Roderick Schertler
803 <roderick@argon.org> did all the conditional compilation stuff,
804 look at HiRes.pm and the test suites; it's good educational reading.
806 1.15 Mon Nov 10 21:30 1997
807 - HiRes.pm: update pod. Provided by Gisle Aas.
808 - HiRes.xs: if gettimeofday() called in scalar context, do
809 something more useful than before. Provided by Gisle Aas.
810 - README: tell of xsubpp '-nolinenumber' woes. thanks to
811 Edward Henigin <ed@texas.net> for pointing out the problem.
813 1.14 Wed Nov 5 9:40 1997
814 - Makefile.PL: look for setitimer
815 - HiRes.xs: if missing ualarm, but we have setitimer, make up
816 our own setitimer. These were provided by Gisle Aas.
818 1.13 Tue Nov 4 23:30 1997
819 - Makefile.PL: fix autodetect mechanism to do try linking in addition
820 to just compiling; should fix Linux build problem. Fix was provided
823 1.12 Sun Oct 12 12:00:00 1997
824 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
825 you may need to comment this back out if you have an older xsubpp.
826 - HiRes.xs: set PROTOTYPES: DISABLE
828 1.11 Fri Sep 05 16:00:00 1997
830 Had some line commented out that shouldn't have been (testing
833 Previous version was corrupted.
835 1.10 Thu May 22 20:20:00 1997
836 - HiRes.xs, HiRes.pm, t/*:
837 - only compile what we have OS support for (or can
839 - only test what we compiled
840 - gross improvement to the test suite
842 This work was all done by Roderick Schertler
843 <roderick@argon.org>. If you run Linux or
844 one of the other ualarm-less platforms, and you like this
845 module, let Roderick know; without him, it still wouldn't
846 be working on those boxes...
847 - Makefile.PL: figure out what routines the OS has and
848 only build what we need. These bits were written by Jarkko
849 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
851 1.02 Mon Dec 30 08:00:00 1996
852 - HiRes.pm: update documentation to say what to do when missing
853 ualarm() and friends.
854 - README: update to warn that ualarm() and friends need to exist
856 1.01 Fri Oct 17 08:00:00 1996
857 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
858 - HiRes.pm: put blank line between __END__ and =head1 so that
861 1.00 Tue Sep 03 13:00:00 1996
862 - original version; created by h2xs 1.16