1 Revision history for the Perl extension Time::HiRes.
4 - No changes since 1.9727_03
7 - use PERL_UNUSED_ARG() in the not-here-croak branches
8 - Makefile.PL/metadata improvements
10 1.9727_02 [2015-08-29]
11 - correct documentation to note that Time::NVtime is NV, not double
13 - avoid a warning about clock_t on OS X / clang 3.6 - 3.6
16 1.9727_01 [2015-08-17]
17 - Remove usage of 5.005 thread APIs in favor of 5.6 and up thread APIs
18 [rt.cpan.org #102799].
19 - Silence a C++ compiler warning [rt.cpan.org #101497].
20 - respect $Config{run} for cross-compiling [rt.cpan.org #92607]
23 - Correct s/us splitting of usleep(1000000) [rt.cpan.org #78266].
24 - Avoid integer overflow in itimer-based alarm() with large
25 argument [rt.cpan.org #87160].
26 - Define PERL_NO_INLINE_FUNCTIONS during probe compilations, to
27 avoid false negatives caused by not linking with the perl core.
28 - Be more careful about context in stat().
29 - Install into "site" directories by default on perl 5.11+
31 - Fix a couple of doc typos [rt.cpan.org #85365].
32 - Fix function name in a doc example [rt.cpan.org #86318].
33 - Provide lstat() that yields high-res timestamps, alongside
34 the existing high-res stat() [rt.cpan.org #78732].
37 - Correct stack discipline in stat(), which was screwing up list
38 operations in expressions containing calls to it [rt.cpan.org
40 - Add missing OUTPUT sections to the XS code [rt.cpan.org #70930].
41 - Skip itimer tests on GNU/Hurd, which has the API but lacks
42 the implementation [rt.cpan.org #72754].
43 - Fix a doubled word in the documentation [rt.cpan.org #72763].
46 - Correct XS parameter list, and therefore prototype, for
47 unimplemented-on-this-platform version of clock_nanosleep()
49 - Declare package variables with "our" rather than "use vars".
50 - Corresponding to "our" usage, check for minimum Perl version
52 - Declare module dependencies.
55 - Remove $ENV{PERL_CORE} logic from test suite, which is no
56 longer desired in the core.
57 - Convert test suite to use Test::More.
58 - Factor out watchdog code from test suite.
59 - In test suite, be consistent about using fully-qualified form
61 - Divide test suite into feature-specific scripts.
62 - Make ualarm timing test less vulnerable to delay-induced false
63 failure, from Dave Mitchell.
66 - Update for changes in build process in the core, patches
67 from BinGOs [rt.cpan.org #58858] and Craig Berry [rt.cpan.org
69 - Fix broken linkage on Windows with gcc 3.4 seen with ActivePerl,
70 report from Christian Walde [rt.cpan.org #61648], fix derived
72 - Jump through hoops to avoid compiler warnings.
75 - Address [rt.cpan.org #54196] alarm and ularm return values are bogus,
76 additional fix from Gisle Aas
77 - Address [rt.cpan.org #55665] "Bad plan" on Windows,
78 report and fix from Jan Dubois
81 - Address [rt.cpan.org #54196] alarm and ularm return values are bogus,
82 report and fix from Nicholas Clark
85 - As with QNX, Haiku has the API of interval timers but not
86 the implementation (bleadperl change #34630), hence skip
87 the tests, via David Mitchell.
90 - .xs code cleanup from Albert Dvornik
91 - in the #39 and #40 do not do us I did, mixing alarm() and
92 sleep(). Now instead spin until enough time has passed.
95 - Skip the tests added in 1.9716 (#39, #40) if there's no subsecond
96 alarm capability, like with the older subsecond alarm tests
99 - Change documentation to agree with reality: there are
100 no interval timers in Win32.
101 - Address [rt.cpan.org #35899] (problem in subsecond sleeps),
102 add two tests to guard against this problem
103 - Address [rt.cpan.org #36600] 'Division by zero' failure in test suite
104 - Address [rt.cpan.org #37340] [PATCH] Address timer process in test
105 - Address [rt.cpan.org#40311 ] bad implementation of hrt_usleep
106 with TIME_HIRES_NANOSLEEP
109 - Silly me: Makefile.PL does need to accept arguments other than mine.
110 Some testing frameworks obviously do this.
111 - Add retrying for tests 34..37, which are the most commonly
112 failing tests. If this helps, consider extending the retry
113 framework to all the tests. [Inspired by Slaven Rezic,
114 [rt.cpan.org #34711] Occasional failures of test 35 or 36 (FreeBSD)]
117 - Under Perl 5.6.* NVgf needs to be "g", reported by Zefram,
118 it seems that ppport.h 3.13 gets this wrong.
119 - remove the check in Makefile.PL for 5.7.2, shouldn't be
120 (a) necessary (b) relevant
121 - add logic to Makefile.PL to skip configure/write Makefile
122 step if the "xdefine" file already exists, indicating that
123 the configure step has already been done, one can still
124 force (re)configure by "perl Makefile.PL configure",
125 or of course by "make clean && perl Makefile.PL".
128 - for alarm() and ualarm() [Perl] prefer setitimer() [C]
129 instead of ualarm() [C] since ualarm() [C] cannot portably
130 (and standards-compliantly) be used for more than 999_999
131 microseconds (rt.cpan.org #34655)
132 - it seems that HP-UX has started (at least in 11.31 ia64)
133 #defining the CLOCK_REALTIME et alia (instead of having
135 - document all the diagnostics
138 - move the sub tick in the test file back to where it used to be
139 - in the "consider upgrading" message recommend at least Perl 5.8.8
140 and make the message to appear only for 5.8.0 since 5.8.1 and
141 later have the problem fixed
142 - VOS tweak for Makefile (core perl change #33259)
143 - since the test #17 seems to fail often, relax its limits a bit
146 - lost VMS test skippage from Craig Berry
147 - reformat the test code a little
150 - I got the sense of the QNX test the wrong way in an attempt
151 to generalize it for future
154 - casting fixes from Robin Barker for g++ and 64bitint
155 - in QNX skip the itimer tests because though the API
156 is there, the implementation isn't, from Matt Kraai
157 - raise the dead man timer to 180 seconds for really
159 - elaborate the UTF-8 locale warning from Makefile.PL
162 - [rt.cpan.org #29747]: Build failure with perl 5.005_05
163 Fixed by regenerating the ppport.h using Devel::PPPort 3.13.
166 - t/HiRes.t failed in Perl 5.6.2,
167 "action is not of type POSIX::SigAction at t/HiRes.t line 318",
168 reported and fixed by Anton Berezin, the reason was faulty
169 use of sigaction() when restoring the old action.
172 - with bleadperl in VMS the HiRes.t overrun the maximum number
173 of deferred signals because the libc SIGALRM was not strong
174 enough to interrupt select(), and select() got restarted every
175 time, solution is to use POSIX::SigAction if available.
176 A fix from Craig Berry (not 100% there, but helps).
177 - allow for more measuring noise for ualarm() tests 35..37
180 - nanosleep() and clock_nanosleep() detection and use were
181 quite broken; in Linux -lrt needed; fixes from Zefram
182 - [internal] slightly cleaner building of $DEFINE in Makefile.PL,
183 should avoid double/conflicting -D flags
186 - allow 10% of slop in test #14 (testing difference between
187 CORE::time() and Time::HiRes::time()), there seem to be often
188 transient failures from Perl smoke builds on this test
192 - use int main(int argc, char **argv) consistently in Makefile.PL,
194 [rt.cpan.org #23868] nanosleep not detected under Mac OS 10.3.9 starting with Time::HiRes 1.96
195 - if someone still has the locale-broken Perl 5.8.0,
196 suggest that they upgrade their Perl
199 - restore the -DATLEASTFIVEOHOHFIVE, Win32 needed it still
202 - upgrade to ppport.h 3.10_02
203 - remove the -DATLEASTFIVEOHOHFIVE
204 - use the ppport.h PL_ppaddr, PL_statcache, PL_laststatval
205 - use the ppport.h aTHXR for calling Perl stat()
206 - switch into four-digit version since 2.0 is coming up
207 awfully fast but not feeling like a major rewrite
210 - 1.95 broke building in Win32 (since pp_stat is not exported),
211 figured out how to call an op directly in 5.005 (use Perl_ppaddr
212 instead of PL_ppaddr)
213 - backport to Perl 5.004_05 (requires using statcache
214 and laststatval instead of PL_statcache and PL_laststatval)
215 (also checked to work in 5.005_04, 5.6.1, and 5.8.8 with threads)
218 - 1.95 broke builds for threaded Perls, rt.cpan.org tickets:
219 [rt.cpan.org #23694] Time::HiRes fails tests on Solaris and Perl 5.6.1
220 [rt.cpan.org #23712] Time-HiRes 1.95 Fails make on AIX 5.2 with Perl 5.8.8
221 [rt.cpan.org #23730] Time::HiRes 1.95 fails make on MacOS X 10.3.9/perl 5.8.8
222 - use main() prototype consistently in Makefile.PL
225 - integrate core change #29180: Silence VC++ compiler warnings
227 - do not use PL_ppaddr in stat() because that is not available
229 - regenerate fallback/*.inc for older Perls without
230 ExtUtils::Constant because of d_hires_stat, resolves
231 [rt.cpan.org #23694] Time::HiRes fails tests on Solaris and Perl 5.6.1
232 - Make Makefile.PL more defensive against false PERL_CORE
235 - file timestamps oddities seen: the atime and mtime
236 can be out of sync (modify first and read second can leave
237 atime < mtime) and mtime can be subsecond while atime is not.
238 So make the test more forgiving.
241 - the ualarm() tests (34-37) assumed that ualarm(N)
242 could never alarm in less than N seconds, widened
243 the acceptable relative range to 0.9..1.5. Addresses
244 [rt.cpan.org #22090] and [rt.cpan.org #22091].
246 - skip the stat() tests in cygwin and win32, because
247 if run on FAT the timestamp granularity is only 2 seconds.
248 Any good way to detect (cygwin or win32) whether we are
249 being run on NTFS or anywhere with better timestamps?
250 Addresses [rt.cpan.org #22089] and [rt.cpan.org #22098].
253 - scan for subsecond resolution timestamps in struct stat,
254 some known possibilities:
256 (1) struct timespec st_atimespec;
257 st_atimespec.tv_nsec;
262 (4) timestruc_t st_atim;
267 If something like this is found, one can do
270 my @stat = Time::HiRes::stat();
272 or even override the standard stat():
274 use Time::HiRes qw(stat);
276 to get the stat() timestamps
278 my ($atime, $mtime, $ctime) = @stat[8, 9, 10];
280 with subsecond resolution (assuming both the operating
281 system and the filesystem support that kind of thing).
283 Contributions for more systems (especially non-UNIX,
284 e.g. but not limited to: Win32, VMS, OS/2) gladly accepted.
285 (also more UNIX variants welcome: HP-UX? IRIX?)
287 Thanks to H.Merijn Brand, John Peacock, and Craig
288 Berry for brave beta testing.
291 - ualarm() in SuSE 10.1 was overflowing after ~4.2 seconds,
292 possibly due to a glibc bug/feature (suspected overflow at
293 2**32 microseconds?), workaround by using the setitimer()
294 implementation of ualarm() if either useconds or
295 interval > 999_999 (this case seems to vary between systems:
296 are useconds more than 999_999 for ualarm() defined or not)
297 Added more ualarm() tests to catch various overflow points,
298 hopefully no problems in various platforms.
299 (The problem report by Mark Seger and Jon Paul Sullivan of HP.)
302 - tweak still needed for Const64(), from Jerry Hedden
303 - get a freshly generated ppport.h
304 - update Copyright years
307 - Const64() already appends an 'LL' (or i64), so provide LL and i64
308 forms for the IV_1E[679] (effects Win32 and Cygwin), reported by
310 - the Changes entry for 1.88 talked about [IN]V_1[679],
314 - clean up the g++ warnings in HiRes.xs, all of them
315 about mixing integer and floating point, introduce
316 constants IV_1E[679] and NV_1E[679]
319 - [rt.cpan.org #17442] 'make test' frequently fails under
320 Cygwin Perl v5.8.8, reported and patched by J. R. Hedden
321 (two race condition bugs in the END block in the case the
322 main process dies before the timer process, unearthed
323 by a bug in Cygwin ualarm)
326 - HiRes.t:s/ok 32/ok 33/, from Dominic Dunlop
327 - tighten up the clock() test marginally by requiring non-negative
328 - clock_nanosleep() and clock() doc tweaks
331 - the interface to clock_nanosleep() is more natural
332 when it is like (hires) time() (instead of like nanosleep),
333 and the .xs implementation of clock_nanosleep() in 1.84
335 - the semantics of clock() are not quite so silly as I thought,
336 but still somewhat odd, documented as such
337 - additional enhancements to the clock() documentation
338 - add test for clock_nanosleep() (I cannot test this
339 since none of my systems have the function)
340 - add test for clock()
343 - add clock() which returns the processor time in
344 (floating point) seconds since an arbitrary era
345 - add clock_nanosleep() which suspends the current
346 thread until either absolute time or for relative time
347 - [rt.cpan.org #16486] printf missing value in HiRes.t
348 - add constants CLOCKS_PER_SEC, CLOCK_SOFTTIME, TIMER_ABSTIME
352 - has_symbol() was wrong since e.g. ITIMER_VIRTUAL is exported
353 via @EXPORT_OK even when it is not available. This is heinous.
354 @EXPORT_OK should be determined at Makefile.PL time.
355 - be more lenient is testing clock_gettime(): allow more slop,
356 and retry up to three times, sleeping a random nap between
358 - human months are one-based (noticed by Anton Berezin)
361 - CLOCK_REALTIME is an enum value (of the clockid_t enum)
362 in HP-UX (and might be so elsewhere, too), debugged by
364 - include const-c.inc as late as possible (from Randy Kobes,
365 [rt.cpan.org #15552] to avoid undefined usleep() on Win32
368 - try to be more robust and consistent in the detection of
369 CLOCK_REALTIME and ITIMER_VIRTUAL in HiRes.t: the proper
374 eval 'import Time::HiRes qw($symbol)';
375 return 0 unless $@ eq '';
376 return exists ${"Time::HiRes::$symbol"};
383 in the test. All these moves are needed because
385 1) one cannot directly do eval 'Time::HiRes::FOO_BAR'
386 because FOO_BAR might have a true value of zero
387 (or in the general case an empty string or even undef)
389 2) In case FOO_BAR is not available in this platform,
390 &FOO_BAR avoids the bareword warning
392 - wait more (1.5 seconds instead of 0.1) for the CLOCK_REALTIME test
393 but expect the 'customary' slop of 0.20 instead of 0.25
394 - fixed inside a comment HAS_POLL -> TIME_HIRES_NANOSLEEP
395 - at the end of HiRest.t tell how close we were to termination
398 - Gisle noticed a mistake (using HAS_NANOSLEEP) in 1.79
401 - try nanosleep for emulating usleep -- may help in some weird
402 embedded realtime places which have nanosleep but neither usleep
403 nor select nor poll (doesn't have to be weird embedded realtime
404 place, though -- in many places usleep is nanosleep anyway)
405 - try poll for emulating usleep -- this may help some obscure/old
406 SVR4 places that have neither usleep nor select
407 - a redundant test guard in HiRes.t
410 - ITIMER_VIRTUAL detection in HiRes.t had problems (that we cannot
411 in the general case fail already at 'use' phase is suboptimal)
412 - fixes to the documentation of clock_gettime() and clock_getres()
415 - add support for the POSIX clock_gettime() and clock_getres(),
416 if available, either as library calls or as syscalls
417 - be more defensive about missing functionality: break out
418 early (during 'use') if no e.g. clock_getres() is available,
419 and protect our back by trapping those cases also in HiRes.xs
420 - the test added in 1.76 could cause an endless loop e.g. in Solaris,
421 due to mixing of sleep() and alarm() (bad programmer, no cookie!)
424 - testing for nanosleep had wrong logic which caused nanosleep
425 to become undefined for e.g. Mac OS X
426 - added a test for a core dump that was introduced by Perl 5.8.0
427 safe signals and was fixed for the time of 5.8.1 (one report of
428 the core dump was [perl #20920]), the test skipped pre-5.8.1.
429 - *cough* s/unanosleep/nanosleep/g; *cough*
432 - installation patch from Gisle Aas: in Perls 5.8.x and later
433 use MakeMaker INSTALLDIRS value of 'perl' instead of 'site'.
436 - [cpan #14608] Solaris 8 perl 5.005_03 File::Spec module does not have method rel2abs
437 (the workaround is not to use rel2abs, should not be necessary)
438 - [cpan #14642] U2time wrongly exported on the C API
439 (patch supplied by the reporter, SALVA@cpan.org)
440 - add release dates to Changes
443 - Time::HiRes::nanosleep support for Solaris [PATCH]
444 (POSIX::uname() not available if building with core perl,
445 from Gisle Aas, via perl5-porters, perl change #25295)
448 - going back to the 1.68 loader setup (using DynaLoader)
449 since too many weird things starting breaking
450 - fix a typo in José Auguste-Etienne's name
453 - a thinko in the nanosleep() detection
454 - move more changes stuff from the README to Changes
455 - add -w to the Makefile.PL
458 - oops in 1.69 about @ISA (not affecting anything but silly)
459 - add copyright 2005 to HiRes.pm
460 - add copyright and license to HiRes.xs
461 - add copyrights 2003, 2004, 2005 to README
464 - actually run a test for nanosleep
465 (if there is no $Config{d_nanosleep}) since e.g. in AIX 4.2
466 it seems that one can link in nanosleep() but then calling
467 it fails instantly and sets errno to ENOSYS (Not implemented).
468 This may be fixable in the AIX case by figuring out the right
469 (realtime POSIX?) libs and whatnot, but in the general case
470 running a real test case is better. (Of course, this change
471 will no doubt run into portability problems because of the
472 execution step...) Note that because of hysterical raisins
473 most Perls do NOT have $Config{d_nanosleep} (scanning for
474 it by Configure would in many platforms require linking in
475 things like -lrt, which would in many platforms be a bad idea
477 (from José Auguste-Etienne)
478 - support XSLoader also since it's much faster
479 (from Alexey Tourbin)
480 - add SEE ALSO (BSD::Resource and Time::TAI64)
483 - somehow 1.67 had a lot of doubled lines (a major cut-and-paste
484 error suspected), but miraculously it still worked since the
485 doubling took place below the __END__ token
486 - undef Pause() before defining it to avoid redefinition warnings
487 during compilation in case perl.h had already defined Pause()
488 (part of perl change #24271)
492 - (internal) don't ignore the return value of gettimeofday()
493 - (external) return undef or an empty if the C gettimeofday() fails
494 (affects Time::HiRes gettimeofday() and the hires time())
498 - fix the 'hierachy' typo in Makefile.PL [rt.cpan.org #8492]
499 - should now build in Solaris [rt.cpan.org #7165] (since 1.64)
500 - should now build in Cygwin [rt.cpan.org #7535] (since 1.64)
501 - close also [rt.cpan.org #5933] "Time::HiRes::time does not
502 pick up time adjustments like ntp" since ever reproducing it
503 (and therefore verifying a possible fix) in the same environment
504 has become rather unlikely
507 - one should not mix u?alarm and sleep (the tests modified
508 by 1.65, #12 and #13, hung in Solaris), now we just busy
509 loop executing an empty block
510 - in the documentation underline the unspecificity of mixing
512 - small spelling fixes
515 - regenerate ppport.h with Devel::PPPort 3.03,
516 now the MY_CXT_CLONE is defined in ppport.h,
517 we no more need to do that.
519 - the test #12 would often hang in sigsuspend() (at least that's
520 where Mac OS X' ktrace shows it hanging). With the sleep()s
521 changed to sleep(1)s, the tests still pass but no hang after
522 a few hundred repeats.
525 - Win32 and any ithread build: ppport.h didn't define
526 MY_CXT_CLONE, which seems to be a Time-HiRes-ism.
529 - Skip testing if under PERL_CORE and Time::HiRes has not
530 been Configured (from Marcus Holland-Moritz, core change
532 - Use ppport.h generated by Devel::PPPort 3.01,
533 allowing cutting away our own portability code.
534 - Don't use $ENV{PERL_CORE} for < 5.6.0.
535 - Don't use "for my $i" for <= 5.003.
536 - Don't use Pause() for <= 5.003.
537 - Can't use newSVpvf for <= 5.003.
538 (most of the changes from Marcus)
541 - Win32: reset reading from the performance counters every
542 five minutes to better track wall clock time (thanks to
543 PC timers being often quite bad), should help long-running
547 - Win32: Patch from Steve Hay
548 [PATCH] Re: [perl #30755] [Win32] Different results from Time::HiRes::gettimeofdayunder the debugger
549 to [perl #30755] reported by Nigel Sandever
551 - Cygwin: Use the Win32 recalibration code also in Cygwin if the
552 <w32api/windows.h> APIs are available. Cygwin testing by
553 Yitzchak Scott-Thoennes.
555 - Solaris: use -lposix4 to get nanosleep for Solaris 2.6,
556 after that keep using -lrt, patch from Alan Burlison,
557 bug reported in [cpan #7165]
560 - Change the Win32 recalibration limit to 0.5 seconds and tweak
561 the documentation to blather less about the gory details of the
562 Win32 implementation and more about the complications in general
563 of meddling with the system clock.
566 - Document the 1.57 change better.
569 - Win32/Cygwin/MinGW: if the performance counter drifts by more
570 than two seconds from the system clock (due to ntp adjustments,
571 for example), recalibrate our internal counter: from Jan Dubois,
572 based on [cpan #5933] by Jerry D. Hedden.
575 - Give a clearer message if the tests timeout (perl change #22253)
576 - Don't use /tmp or its moral equivalents (perl bug #15036,
580 - Windows: mingw32 patch from Mike Pomraning (use Perl's Const64()
581 instead of VC-specific i64 suffix)
584 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
587 - Windows: higher resolution time() by using the Windows
588 performance counter API, from Jan Dubois and Anton Shcherbinin.
589 The exact new higher resolution depends on the hardware,
590 but it should be quite a bit better than using the basic
594 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
595 the subtest 18. No known analysis nor fix, but added
596 an alarm (that requires fork() and alarm()) to the test.
599 - doc tweaks from mjd (perl change #20456)
600 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
603 - add a message (for non-core builds) to Makefile.PL about
604 the LC_ALL=C workaround
605 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
606 - the nanosleep() probe was broken (perl change #20061)
607 - use existence instead of definedness for feature probes
609 - MPE/iX tweak (perl change #20042)
610 - do not use HAS_NANOSLEEP (perl change #19898)
613 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
614 - OS/2 can always mix subsecond sleeps with signals
615 (part of perl change #19789)
618 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
622 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
623 - unnecessary (nanosleep is in libc anyway)
624 - harmful (-lrt slows down execution)
625 - incompatible (with many distributions' pthreads)
628 - do not create files in blib directories under core
629 (perl change #19160, from rgs)
630 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
633 - guarantee that $xdefine in HiRes.t is always defined
634 (perl change #19109, from IlyaZ)
635 - a cleaner way to detect PERL_CORE (perl change #19111,
639 - add hints/irix.pl to turn off overly POSIX flags that
640 cause hide struct timespec to be hidden (and compilation
641 to fail) (bleadperl change #19085)
642 - documentation tweaks
645 - add c:/temp to the list of temp directories to probe
646 so that cygwin (and win*?) builds are happy. This was
647 needed at least in my cygwin 1.3.20/w2k setup.
650 - modernize the constants code (from Nicholas Clark)
653 - At some point the ability to figure our the correct incdir
654 for EXTERN.h (either a core perl build, or an installed perl)
655 had broken (which lead into all test compiles failing with
656 a core perl build, but thanks to the robustness of Makefile.PL
657 nothing of this was visible). The brokenness seemed to be
658 caused by $ENV{PERL_CORE} not being on for core builds?
659 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
660 right, and both styles of build should work again.
663 - Nicholas Clark noticed that the my_catdir() emulation function
664 was broken (which means that we didn't really work for Perls
666 - inspired by fixing the above made the whole Makefile.PL -w
668 - tightened up the Makefile.PL output, less whitespace
671 - fix from Craig Berry for better building in VMS with PERL_CORE
674 - no functional changes
675 - move lib/Time/HiRes.pm as Hires.pm
676 - libraries scanning was slightly broken (always scanned
677 for a library even when $Config{libs} already had it)
680 - Ray Zimmerman ran into a race condition in Mac OS X.
681 A 0.01-second alarm fired before the test expected.
682 The test first slept indefinitely (blocking for signals)
683 and only after that tested for the signal having been sent.
684 Since the signal had already been sent, the test #12 never
685 completed. The solution: test first, then block.
686 - default to being silent on all probing attempts, set the
687 environment variable VERBOSE to a true value to see the
688 details (the probing command and the possible errors)
691 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
692 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
693 cannot be upgraded from CPAN to override the 5.8.0 version
695 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
696 because the include directories did not adjust themselves
697 if $ENV{PERL_CORE} (Hugo van der Sanden)
698 - add documentation about the restart of select() under alarm()
701 - small documentation tweaks
705 - better VMS operation (Craig Berry)
708 - our time machine is accelerating: now works with Perl 5.004_01
709 (tried with 5.003_07 and 5.002 but I get segmentation faults
710 from running the Makefile.PL with those in Tru64 4.0D)
713 - backward compatibility (pre-5.6.0) tweaks:
714 - no XSLoader in 5.00503, use DynaLoader instead
715 - no SvPV_nolen, either
716 - no PerlProc_pause(), either
717 - now tested with 5.00404 and 5.00503
718 - Makefile.PL requires 5.00404 (no more 5.002)
719 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
720 (this means that one can mix subsecond sleeps with alarms)
721 - because of nanosleep we probe for -lrt and -lposix4
722 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
723 is available by exportable constants Time::HiRes::d_func
724 (since older Perl do not have them in %Config, and even
725 5.8.0 does not probe for nanosleep)
728 - backward compatibility (pre-5.6.1) tweaks:
729 - define NV if no NVTYPE
730 - define IVdf if needed (note: the Devel::PPPort
731 in 5.8.0 does not try hard hard enough since
732 the IVSIZE might not be defined)
733 - define NVgf if needed
734 - grab the typemap from 5.8.0 for the NV stuff
736 1.31 and 1.32 add more backward compatibility (now all the way
737 back to Perl 5.00404), and using nanosleep() (if available) for
742 - release 1.29_02 as 1.30
744 1.30 adds all the changes made during the Perl 5.6->5.7->5.8
745 development cycle. Most notably portability across platforms has been
746 enhanced, and the interval timers (setitimer, getitimer) have been
747 added. Note that the version of Time::HiRes that is included in Perl
748 5.8.0 calls itself 1.20_00, but it is equivalent to this Time::HiRes
749 version. Note also that in 1.30 Wegscheid turns over the maintenance
750 to Jarkko Hietaniemi.
754 - fix a silly unclosed comment typo in HiRes.xs
755 - document and export REALTIME_REALPROF (Solaris)
759 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
760 (need to patch this also in Perl 5.[89])
761 - remove CVS revision log from HiRes.xs
765 The following numbered patches refer to the Perl 5.7 changes,
766 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
768 - 17558: Add #!./perl to the .t
769 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
770 - 16198: political correctness, from Simon Cozens
771 - 15857: doc tweaks, from Jarkko Hietaniemi
772 - 15593: optimization in .xs, from Paul Green
773 - 14892: pod fixes, from Robin Barker
774 - 14100: VOS fixes, from Paul Green
775 - 13422: XS segfault, from Marc Lehmann
776 - 13378: whether select() gets restarted on signals, depends
777 - 13354: timing constraints, again, from Andy Dougherty
778 - 13278: can't do subsecond alarms with ualarm;
779 break out early if alarms do not seem to be working
780 - 13266: test relaxation (cygwin gets lower hires
781 times than lores ones)
782 - 12846: protect against high load, from Jarkko Hietaniemi
783 - 12837: HiRes.t VMS tweak, from Craig A. Berry
784 - 12797: HiRes.t VMS tweak, from Charles Lane
785 - 12769: HiRes.t VMS tweak, from Craig A. Berry
786 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
787 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
788 - 12692: alarm() ain't gonna work if ualarm() ain't,
789 from Gurusamy Sarathy
790 - 12680: minor VMS tweak, from Charles Lane
791 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
792 - 12609: croak on negative time, from Jarkko Hietaniemi
793 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
794 - 12594: MacOS Classic timeofday, from Chris Nandor
795 - 12473: allow for more than one second for sleep() and usleep()
796 - 12458: test tuning, relax timing constraints,
797 from Jarkko Hietaniemi
798 - 12449: make sleep() and usleep() to return the number
799 of seconds and microseconds actually slept (analogously
800 with the builtin sleep()), also make usleep() croak if
801 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
802 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
803 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
804 - 12196: use ftime() on Win32, from Artur Bergman
805 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
806 - 12105: use GetSystemTime() on Win32, from Artur Bergman
807 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
808 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
809 - 11797: problem in HiRes.t, from John P. Linderman
810 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
811 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
812 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
813 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
814 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
815 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
817 - 10942: MPE/IX test tweaks, from Mark Bixby
818 - 10784: unnecessary pod2man calls, from Andy Dougherty
819 - 10354: ext/ + -Wall, from Doug MacEachern
820 - 10320: fix the BOOT section to call myU2time correctly
821 - 10317: correct casting for AIX< from H. Merijn Brand
822 - 10119: document that the core time() may be rounding, not truncating
823 - 10118: test fix, from John Peacock
824 - 9988: long =item, from Robin Barker
825 - 9714: correct test output
826 - 9708: test also the scalar aspect of getitimer()
827 - 9705: Add interval timers (setitimer, getitimer)
828 - 9692: do not require at least 5.005 using XS
830 The following changes were made on top of the changes
831 made for Time::HiRes during the Perl 5.7 development
832 cycle that culminated in the release of Perl 5.8.0.
834 - add "require 5.005" to the Makefile.PL
835 - remove the REVISION section (CVS log) from HiRes.pm
836 - add jhi's copyright alongside Douglas'
837 - move HiRes.pm to lib/Time/
839 - modify HiRes.t to use $ENV{PERL_CORE}
840 - modify the original Time::HiRes version 1.20 Makefile.PL
841 to work both with Perl 5.8.0 and the new code with pre-5.8.0
842 Perls (tried with 5.6.1)
843 - tiny tweaks and updates in README and TODO
844 - bump the VERSION to 1.29
846 1.20 Wed Feb 24 21:30 1999
847 - make our usleep and ualarm substitutes into hrt_usleep
848 and hrt_ualarm. This helps static links of Perl with other
849 packages that also have usleep, etc. From
850 Ilya Zakharevich <ilya@math.ohio-state.edu>
851 - add C API stuff. From Joshua Pritikin
852 <joshua.pritikin@db.com>
853 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
854 - hopefully correct "-lc" fix for SCO.
857 1.20 adds a platform neutral set of C accessible routines if you are
858 running 5.005+. All other changes are packaging changes and build
859 fixes(?) for statically linked Perl, SCO, and VMS.
861 1.19 Tue Sep 29 22:30 1998
862 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
863 <seb@stian.demon.co.uk>
864 - change GIMME_V to GIMME to help people with older versions of
866 - fix Win32 version of gettimeofday(). It didn't affect anything,
867 but it confuses people reading the code when the return value
868 is backwards (0 is success).
869 - fix Makefile.PL (more) so that detection of gettimeofday is
872 1.19 has better VMS support.
874 1.18 Mon Jul 6 22:40 1998
875 - add usleep() for Win32.
876 - fix Makefile.PL to fix reported HP/UX feature where unresolved
877 externals still cause an executable to be generated (though no
878 x bit set). Thanks to David Kozinn for report and explanation.
879 Problems with the fix are mine :)
881 1.18 has limited Win32 support (no ualarm). Added usleep for Win32.
882 Probably buggy. I'm sure I'll hear.
884 1.17 Wed Jul 1 20:10 1998
885 - fix setitimer calls so microseconds is not more than 1000000.
886 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
887 - make Win32. We only get gettimeofday (the select hack doesn't
888 seem to work on my Win95 system).
889 - fix test 4 on 01test.t. add test to see if time() and
890 Time::HiRes::time() are close.
892 1.16 Wed Nov 12 21:05 1997
893 - add missing EXTEND in new gettimeofday scalar code.
895 1.16+ should be closer to building out of the box on Linux. Thanks
896 to Gisle Aas for patches, and the ualarm equivalent using setitimer.
898 If your underlying operating system doesn't implement ualarm(), then
899 a fake using setitimer() will be made. If the OS is missing usleep(),
900 a fake one using select() will be made. If a fake can't be made for
901 either ualarm() or usleep(), then the corresponding Perl function will
902 not be available. If the OS is missing gettimeofday(), you will get
903 unresolved externals, either at link- or run-time.
905 This is an improvement; the package used to not even build if
906 you were missing any of these bits. Roderick Schertler
908 <roderick@argon.org> did all the conditional compilation stuff,
909 look at HiRes.pm and the test suites; it's good educational reading.
911 1.15 Mon Nov 10 21:30 1997
912 - HiRes.pm: update pod. Provided by Gisle Aas.
913 - HiRes.xs: if gettimeofday() called in scalar context, do
914 something more useful than before. Provided by Gisle Aas.
915 - README: tell of xsubpp '-nolinenumber' woes. thanks to
916 Edward Henigin <ed@texas.net> for pointing out the problem.
918 1.14 Wed Nov 5 9:40 1997
919 - Makefile.PL: look for setitimer
920 - HiRes.xs: if missing ualarm, but we have setitimer, make up
921 our own setitimer. These were provided by Gisle Aas.
923 1.13 Tue Nov 4 23:30 1997
924 - Makefile.PL: fix autodetect mechanism to do try linking in addition
925 to just compiling; should fix Linux build problem. Fix was provided
928 1.12 Sun Oct 12 12:00:00 1997
929 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
930 you may need to comment this back out if you have an older xsubpp.
931 - HiRes.xs: set PROTOTYPES: DISABLE
933 1.11 Fri Sep 05 16:00:00 1997
935 Had some line commented out that shouldn't have been (testing
938 Previous version was corrupted.
940 1.10 Thu May 22 20:20:00 1997
941 - HiRes.xs, HiRes.pm, t/*:
942 - only compile what we have OS support for (or can
944 - only test what we compiled
945 - gross improvement to the test suite
947 This work was all done by Roderick Schertler
948 <roderick@argon.org>. If you run Linux or
949 one of the other ualarm-less platforms, and you like this
950 module, let Roderick know; without him, it still wouldn't
951 be working on those boxes...
952 - Makefile.PL: figure out what routines the OS has and
953 only build what we need. These bits were written by Jarkko
954 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
956 1.02 Mon Dec 30 08:00:00 1996
957 - HiRes.pm: update documentation to say what to do when missing
958 ualarm() and friends.
959 - README: update to warn that ualarm() and friends need to exist
961 1.01 Fri Oct 17 08:00:00 1996
962 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
963 - HiRes.pm: put blank line between __END__ and =head1 so that
966 1.00 Tue Sep 03 13:00:00 1996
967 - original version; created by h2xs 1.16