1 Revision history for the Perl extension Time::HiRes.
4 - [rt #113856] C90 declaration-after-statement error with darwin threads
7 - MUTEX_LOCK, not PERL_MUTEX_LOCK: blead e5b02b5d
8 - also hrt_ualarm_itimer() is unused: 1cb6cce3
9 - the mutex needs init: 2d639e20
10 - version bump to 1.9732: 730d7fdc
13 - mark unused variable as such: blead a914236c
14 - OS X emulation mutex accidentally unused: da7a6455
15 - remove hrt_ualarm: 6da77c36
16 - pod error fixes: 919ca095
17 - nanosleep and clock_nanosleep cleanups: c8ea02b8..e3ff671b
18 - static funcs in HiRes.xs: 4e922c26
19 - Remove unused variable: 52ffb1b5
22 - TIME_HIRES_DONT_RUN_PROBES=1 to build the probes but not run them
28 - clock_getres(), clock_gettime(), clock_nanosleep() emulation for OS X
32 - No changes since 1.9727_03
34 1.9727_03 [2015-11-06]
35 - use PERL_UNUSED_ARG() in the not-here-croak branches
36 - Makefile.PL/metadata improvements
38 1.9727_02 [2015-08-29]
39 - correct documentation to note that Time::NVtime is NV, not double
41 - avoid a warning about clock_t on OS X / clang 3.6 - 3.6
44 1.9727_01 [2015-08-17]
45 - Remove usage of 5.005 thread APIs in favor of 5.6 and up thread APIs
46 [rt.cpan.org #102799].
47 - Silence a C++ compiler warning [rt.cpan.org #101497].
48 - respect $Config{run} for cross-compiling [rt.cpan.org #92607]
51 - Correct s/us splitting of usleep(1000000) [rt.cpan.org #78266].
52 - Avoid integer overflow in itimer-based alarm() with large
53 argument [rt.cpan.org #87160].
54 - Define PERL_NO_INLINE_FUNCTIONS during probe compilations, to
55 avoid false negatives caused by not linking with the perl core.
56 - Be more careful about context in stat().
57 - Install into "site" directories by default on perl 5.11+
59 - Fix a couple of doc typos [rt.cpan.org #85365].
60 - Fix function name in a doc example [rt.cpan.org #86318].
61 - Provide lstat() that yields high-res timestamps, alongside
62 the existing high-res stat() [rt.cpan.org #78732].
65 - Correct stack discipline in stat(), which was screwing up list
66 operations in expressions containing calls to it [rt.cpan.org
68 - Add missing OUTPUT sections to the XS code [rt.cpan.org #70930].
69 - Skip itimer tests on GNU/Hurd, which has the API but lacks
70 the implementation [rt.cpan.org #72754].
71 - Fix a doubled word in the documentation [rt.cpan.org #72763].
74 - Correct XS parameter list, and therefore prototype, for
75 unimplemented-on-this-platform version of clock_nanosleep()
77 - Declare package variables with "our" rather than "use vars".
78 - Corresponding to "our" usage, check for minimum Perl version
80 - Declare module dependencies.
83 - Remove $ENV{PERL_CORE} logic from test suite, which is no
84 longer desired in the core.
85 - Convert test suite to use Test::More.
86 - Factor out watchdog code from test suite.
87 - In test suite, be consistent about using fully-qualified form
89 - Divide test suite into feature-specific scripts.
90 - Make ualarm timing test less vulnerable to delay-induced false
91 failure, from Dave Mitchell.
94 - Update for changes in build process in the core, patches
95 from BinGOs [rt.cpan.org #58858] and Craig Berry [rt.cpan.org
97 - Fix broken linkage on Windows with gcc 3.4 seen with ActivePerl,
98 report from Christian Walde [rt.cpan.org #61648], fix derived
100 - Jump through hoops to avoid compiler warnings.
103 - Address [rt.cpan.org #54196] alarm and ularm return values are bogus,
104 additional fix from Gisle Aas
105 - Address [rt.cpan.org #55665] "Bad plan" on Windows,
106 report and fix from Jan Dubois
109 - Address [rt.cpan.org #54196] alarm and ularm return values are bogus,
110 report and fix from Nicholas Clark
113 - As with QNX, Haiku has the API of interval timers but not
114 the implementation (bleadperl change #34630), hence skip
115 the tests, via David Mitchell.
118 - .xs code cleanup from Albert Dvornik
119 - in the #39 and #40 do not do us I did, mixing alarm() and
120 sleep(). Now instead spin until enough time has passed.
123 - Skip the tests added in 1.9716 (#39, #40) if there's no subsecond
124 alarm capability, like with the older subsecond alarm tests
127 - Change documentation to agree with reality: there are
128 no interval timers in Win32.
129 - Address [rt.cpan.org #35899] (problem in subsecond sleeps),
130 add two tests to guard against this problem
131 - Address [rt.cpan.org #36600] 'Division by zero' failure in test suite
132 - Address [rt.cpan.org #37340] [PATCH] Address timer process in test
133 - Address [rt.cpan.org#40311 ] bad implementation of hrt_usleep
134 with TIME_HIRES_NANOSLEEP
137 - Silly me: Makefile.PL does need to accept arguments other than mine.
138 Some testing frameworks obviously do this.
139 - Add retrying for tests 34..37, which are the most commonly
140 failing tests. If this helps, consider extending the retry
141 framework to all the tests. [Inspired by Slaven Rezic,
142 [rt.cpan.org #34711] Occasional failures of test 35 or 36 (FreeBSD)]
145 - Under Perl 5.6.* NVgf needs to be "g", reported by Zefram,
146 it seems that ppport.h 3.13 gets this wrong.
147 - remove the check in Makefile.PL for 5.7.2, shouldn't be
148 (a) necessary (b) relevant
149 - add logic to Makefile.PL to skip configure/write Makefile
150 step if the "xdefine" file already exists, indicating that
151 the configure step has already been done, one can still
152 force (re)configure by "perl Makefile.PL configure",
153 or of course by "make clean && perl Makefile.PL".
156 - for alarm() and ualarm() [Perl] prefer setitimer() [C]
157 instead of ualarm() [C] since ualarm() [C] cannot portably
158 (and standards-compliantly) be used for more than 999_999
159 microseconds (rt.cpan.org #34655)
160 - it seems that HP-UX has started (at least in 11.31 ia64)
161 #defining the CLOCK_REALTIME et alia (instead of having
163 - document all the diagnostics
166 - move the sub tick in the test file back to where it used to be
167 - in the "consider upgrading" message recommend at least Perl 5.8.8
168 and make the message to appear only for 5.8.0 since 5.8.1 and
169 later have the problem fixed
170 - VOS tweak for Makefile (core perl change #33259)
171 - since the test #17 seems to fail often, relax its limits a bit
174 - lost VMS test skippage from Craig Berry
175 - reformat the test code a little
178 - I got the sense of the QNX test the wrong way in an attempt
179 to generalize it for future
182 - casting fixes from Robin Barker for g++ and 64bitint
183 - in QNX skip the itimer tests because though the API
184 is there, the implementation isn't, from Matt Kraai
185 - raise the dead man timer to 180 seconds for really
187 - elaborate the UTF-8 locale warning from Makefile.PL
190 - [rt.cpan.org #29747]: Build failure with perl 5.005_05
191 Fixed by regenerating the ppport.h using Devel::PPPort 3.13.
194 - t/HiRes.t failed in Perl 5.6.2,
195 "action is not of type POSIX::SigAction at t/HiRes.t line 318",
196 reported and fixed by Anton Berezin, the reason was faulty
197 use of sigaction() when restoring the old action.
200 - with bleadperl in VMS the HiRes.t overrun the maximum number
201 of deferred signals because the libc SIGALRM was not strong
202 enough to interrupt select(), and select() got restarted every
203 time, solution is to use POSIX::SigAction if available.
204 A fix from Craig Berry (not 100% there, but helps).
205 - allow for more measuring noise for ualarm() tests 35..37
208 - nanosleep() and clock_nanosleep() detection and use were
209 quite broken; in Linux -lrt needed; fixes from Zefram
210 - [internal] slightly cleaner building of $DEFINE in Makefile.PL,
211 should avoid double/conflicting -D flags
214 - allow 10% of slop in test #14 (testing difference between
215 CORE::time() and Time::HiRes::time()), there seem to be often
216 transient failures from Perl smoke builds on this test
220 - use int main(int argc, char **argv) consistently in Makefile.PL,
222 [rt.cpan.org #23868] nanosleep not detected under Mac OS 10.3.9 starting with Time::HiRes 1.96
223 - if someone still has the locale-broken Perl 5.8.0,
224 suggest that they upgrade their Perl
227 - restore the -DATLEASTFIVEOHOHFIVE, Win32 needed it still
230 - upgrade to ppport.h 3.10_02
231 - remove the -DATLEASTFIVEOHOHFIVE
232 - use the ppport.h PL_ppaddr, PL_statcache, PL_laststatval
233 - use the ppport.h aTHXR for calling Perl stat()
234 - switch into four-digit version since 2.0 is coming up
235 awfully fast but not feeling like a major rewrite
238 - 1.95 broke building in Win32 (since pp_stat is not exported),
239 figured out how to call an op directly in 5.005 (use Perl_ppaddr
240 instead of PL_ppaddr)
241 - backport to Perl 5.004_05 (requires using statcache
242 and laststatval instead of PL_statcache and PL_laststatval)
243 (also checked to work in 5.005_04, 5.6.1, and 5.8.8 with threads)
246 - 1.95 broke builds for threaded Perls, rt.cpan.org tickets:
247 [rt.cpan.org #23694] Time::HiRes fails tests on Solaris and Perl 5.6.1
248 [rt.cpan.org #23712] Time-HiRes 1.95 Fails make on AIX 5.2 with Perl 5.8.8
249 [rt.cpan.org #23730] Time::HiRes 1.95 fails make on MacOS X 10.3.9/perl 5.8.8
250 - use main() prototype consistently in Makefile.PL
253 - integrate core change #29180: Silence VC++ compiler warnings
255 - do not use PL_ppaddr in stat() because that is not available
257 - regenerate fallback/*.inc for older Perls without
258 ExtUtils::Constant because of d_hires_stat, resolves
259 [rt.cpan.org #23694] Time::HiRes fails tests on Solaris and Perl 5.6.1
260 - Make Makefile.PL more defensive against false PERL_CORE
263 - file timestamps oddities seen: the atime and mtime
264 can be out of sync (modify first and read second can leave
265 atime < mtime) and mtime can be subsecond while atime is not.
266 So make the test more forgiving.
269 - the ualarm() tests (34-37) assumed that ualarm(N)
270 could never alarm in less than N seconds, widened
271 the acceptable relative range to 0.9..1.5. Addresses
272 [rt.cpan.org #22090] and [rt.cpan.org #22091].
274 - skip the stat() tests in cygwin and win32, because
275 if run on FAT the timestamp granularity is only 2 seconds.
276 Any good way to detect (cygwin or win32) whether we are
277 being run on NTFS or anywhere with better timestamps?
278 Addresses [rt.cpan.org #22089] and [rt.cpan.org #22098].
281 - scan for subsecond resolution timestamps in struct stat,
282 some known possibilities:
284 (1) struct timespec st_atimespec;
285 st_atimespec.tv_nsec;
290 (4) timestruc_t st_atim;
295 If something like this is found, one can do
298 my @stat = Time::HiRes::stat();
300 or even override the standard stat():
302 use Time::HiRes qw(stat);
304 to get the stat() timestamps
306 my ($atime, $mtime, $ctime) = @stat[8, 9, 10];
308 with subsecond resolution (assuming both the operating
309 system and the filesystem support that kind of thing).
311 Contributions for more systems (especially non-UNIX,
312 e.g. but not limited to: Win32, VMS, OS/2) gladly accepted.
313 (also more UNIX variants welcome: HP-UX? IRIX?)
315 Thanks to H.Merijn Brand, John Peacock, and Craig
316 Berry for brave beta testing.
319 - ualarm() in SuSE 10.1 was overflowing after ~4.2 seconds,
320 possibly due to a glibc bug/feature (suspected overflow at
321 2**32 microseconds?), workaround by using the setitimer()
322 implementation of ualarm() if either useconds or
323 interval > 999_999 (this case seems to vary between systems:
324 are useconds more than 999_999 for ualarm() defined or not)
325 Added more ualarm() tests to catch various overflow points,
326 hopefully no problems in various platforms.
327 (The problem report by Mark Seger and Jon Paul Sullivan of HP.)
330 - tweak still needed for Const64(), from Jerry Hedden
331 - get a freshly generated ppport.h
332 - update Copyright years
335 - Const64() already appends an 'LL' (or i64), so provide LL and i64
336 forms for the IV_1E[679] (effects Win32 and Cygwin), reported by
338 - the Changes entry for 1.88 talked about [IN]V_1[679],
342 - clean up the g++ warnings in HiRes.xs, all of them
343 about mixing integer and floating point, introduce
344 constants IV_1E[679] and NV_1E[679]
347 - [rt.cpan.org #17442] 'make test' frequently fails under
348 Cygwin Perl v5.8.8, reported and patched by J. R. Hedden
349 (two race condition bugs in the END block in the case the
350 main process dies before the timer process, unearthed
351 by a bug in Cygwin ualarm)
354 - HiRes.t:s/ok 32/ok 33/, from Dominic Dunlop
355 - tighten up the clock() test marginally by requiring non-negative
356 - clock_nanosleep() and clock() doc tweaks
359 - the interface to clock_nanosleep() is more natural
360 when it is like (hires) time() (instead of like nanosleep),
361 and the .xs implementation of clock_nanosleep() in 1.84
363 - the semantics of clock() are not quite so silly as I thought,
364 but still somewhat odd, documented as such
365 - additional enhancements to the clock() documentation
366 - add test for clock_nanosleep() (I cannot test this
367 since none of my systems have the function)
368 - add test for clock()
371 - add clock() which returns the processor time in
372 (floating point) seconds since an arbitrary era
373 - add clock_nanosleep() which suspends the current
374 thread until either absolute time or for relative time
375 - [rt.cpan.org #16486] printf missing value in HiRes.t
376 - add constants CLOCKS_PER_SEC, CLOCK_SOFTTIME, TIMER_ABSTIME
380 - has_symbol() was wrong since e.g. ITIMER_VIRTUAL is exported
381 via @EXPORT_OK even when it is not available. This is heinous.
382 @EXPORT_OK should be determined at Makefile.PL time.
383 - be more lenient is testing clock_gettime(): allow more slop,
384 and retry up to three times, sleeping a random nap between
386 - human months are one-based (noticed by Anton Berezin)
389 - CLOCK_REALTIME is an enum value (of the clockid_t enum)
390 in HP-UX (and might be so elsewhere, too), debugged by
392 - include const-c.inc as late as possible (from Randy Kobes,
393 [rt.cpan.org #15552] to avoid undefined usleep() on Win32
396 - try to be more robust and consistent in the detection of
397 CLOCK_REALTIME and ITIMER_VIRTUAL in HiRes.t: the proper
402 eval 'import Time::HiRes qw($symbol)';
403 return 0 unless $@ eq '';
404 return exists ${"Time::HiRes::$symbol"};
411 in the test. All these moves are needed because
413 1) one cannot directly do eval 'Time::HiRes::FOO_BAR'
414 because FOO_BAR might have a true value of zero
415 (or in the general case an empty string or even undef)
417 2) In case FOO_BAR is not available in this platform,
418 &FOO_BAR avoids the bareword warning
420 - wait more (1.5 seconds instead of 0.1) for the CLOCK_REALTIME test
421 but expect the 'customary' slop of 0.20 instead of 0.25
422 - fixed inside a comment HAS_POLL -> TIME_HIRES_NANOSLEEP
423 - at the end of HiRest.t tell how close we were to termination
426 - Gisle noticed a mistake (using HAS_NANOSLEEP) in 1.79
429 - try nanosleep for emulating usleep -- may help in some weird
430 embedded realtime places which have nanosleep but neither usleep
431 nor select nor poll (doesn't have to be weird embedded realtime
432 place, though -- in many places usleep is nanosleep anyway)
433 - try poll for emulating usleep -- this may help some obscure/old
434 SVR4 places that have neither usleep nor select
435 - a redundant test guard in HiRes.t
438 - ITIMER_VIRTUAL detection in HiRes.t had problems (that we cannot
439 in the general case fail already at 'use' phase is suboptimal)
440 - fixes to the documentation of clock_gettime() and clock_getres()
443 - add support for the POSIX clock_gettime() and clock_getres(),
444 if available, either as library calls or as syscalls
445 - be more defensive about missing functionality: break out
446 early (during 'use') if no e.g. clock_getres() is available,
447 and protect our back by trapping those cases also in HiRes.xs
448 - the test added in 1.76 could cause an endless loop e.g. in Solaris,
449 due to mixing of sleep() and alarm() (bad programmer, no cookie!)
452 - testing for nanosleep had wrong logic which caused nanosleep
453 to become undefined for e.g. Mac OS X
454 - added a test for a core dump that was introduced by Perl 5.8.0
455 safe signals and was fixed for the time of 5.8.1 (one report of
456 the core dump was [perl #20920]), the test skipped pre-5.8.1.
457 - *cough* s/unanosleep/nanosleep/g; *cough*
460 - installation patch from Gisle Aas: in Perls 5.8.x and later
461 use MakeMaker INSTALLDIRS value of 'perl' instead of 'site'.
464 - [cpan #14608] Solaris 8 perl 5.005_03 File::Spec module does not have method rel2abs
465 (the workaround is not to use rel2abs, should not be necessary)
466 - [cpan #14642] U2time wrongly exported on the C API
467 (patch supplied by the reporter, SALVA@cpan.org)
468 - add release dates to Changes
471 - Time::HiRes::nanosleep support for Solaris [PATCH]
472 (POSIX::uname() not available if building with core perl,
473 from Gisle Aas, via perl5-porters, perl change #25295)
476 - going back to the 1.68 loader setup (using DynaLoader)
477 since too many weird things starting breaking
478 - fix a typo in José Auguste-Etienne's name
481 - a thinko in the nanosleep() detection
482 - move more changes stuff from the README to Changes
483 - add -w to the Makefile.PL
486 - oops in 1.69 about @ISA (not affecting anything but silly)
487 - add copyright 2005 to HiRes.pm
488 - add copyright and license to HiRes.xs
489 - add copyrights 2003, 2004, 2005 to README
492 - actually run a test for nanosleep
493 (if there is no $Config{d_nanosleep}) since e.g. in AIX 4.2
494 it seems that one can link in nanosleep() but then calling
495 it fails instantly and sets errno to ENOSYS (Not implemented).
496 This may be fixable in the AIX case by figuring out the right
497 (realtime POSIX?) libs and whatnot, but in the general case
498 running a real test case is better. (Of course, this change
499 will no doubt run into portability problems because of the
500 execution step...) Note that because of hysterical raisins
501 most Perls do NOT have $Config{d_nanosleep} (scanning for
502 it by Configure would in many platforms require linking in
503 things like -lrt, which would in many platforms be a bad idea
505 (from José Auguste-Etienne)
506 - support XSLoader also since it's much faster
507 (from Alexey Tourbin)
508 - add SEE ALSO (BSD::Resource and Time::TAI64)
511 - somehow 1.67 had a lot of doubled lines (a major cut-and-paste
512 error suspected), but miraculously it still worked since the
513 doubling took place below the __END__ token
514 - undef Pause() before defining it to avoid redefinition warnings
515 during compilation in case perl.h had already defined Pause()
516 (part of perl change #24271)
520 - (internal) don't ignore the return value of gettimeofday()
521 - (external) return undef or an empty if the C gettimeofday() fails
522 (affects Time::HiRes gettimeofday() and the hires time())
526 - fix the 'hierachy' typo in Makefile.PL [rt.cpan.org #8492]
527 - should now build in Solaris [rt.cpan.org #7165] (since 1.64)
528 - should now build in Cygwin [rt.cpan.org #7535] (since 1.64)
529 - close also [rt.cpan.org #5933] "Time::HiRes::time does not
530 pick up time adjustments like ntp" since ever reproducing it
531 (and therefore verifying a possible fix) in the same environment
532 has become rather unlikely
535 - one should not mix u?alarm and sleep (the tests modified
536 by 1.65, #12 and #13, hung in Solaris), now we just busy
537 loop executing an empty block
538 - in the documentation underline the unspecificity of mixing
540 - small spelling fixes
543 - regenerate ppport.h with Devel::PPPort 3.03,
544 now the MY_CXT_CLONE is defined in ppport.h,
545 we no more need to do that.
547 - the test #12 would often hang in sigsuspend() (at least that's
548 where Mac OS X' ktrace shows it hanging). With the sleep()s
549 changed to sleep(1)s, the tests still pass but no hang after
550 a few hundred repeats.
553 - Win32 and any ithread build: ppport.h didn't define
554 MY_CXT_CLONE, which seems to be a Time-HiRes-ism.
557 - Skip testing if under PERL_CORE and Time::HiRes has not
558 been Configured (from Marcus Holland-Moritz, core change
560 - Use ppport.h generated by Devel::PPPort 3.01,
561 allowing cutting away our own portability code.
562 - Don't use $ENV{PERL_CORE} for < 5.6.0.
563 - Don't use "for my $i" for <= 5.003.
564 - Don't use Pause() for <= 5.003.
565 - Can't use newSVpvf for <= 5.003.
566 (most of the changes from Marcus)
569 - Win32: reset reading from the performance counters every
570 five minutes to better track wall clock time (thanks to
571 PC timers being often quite bad), should help long-running
575 - Win32: Patch from Steve Hay
576 [PATCH] Re: [perl #30755] [Win32] Different results from Time::HiRes::gettimeofdayunder the debugger
577 to [perl #30755] reported by Nigel Sandever
579 - Cygwin: Use the Win32 recalibration code also in Cygwin if the
580 <w32api/windows.h> APIs are available. Cygwin testing by
581 Yitzchak Scott-Thoennes.
583 - Solaris: use -lposix4 to get nanosleep for Solaris 2.6,
584 after that keep using -lrt, patch from Alan Burlison,
585 bug reported in [cpan #7165]
588 - Change the Win32 recalibration limit to 0.5 seconds and tweak
589 the documentation to blather less about the gory details of the
590 Win32 implementation and more about the complications in general
591 of meddling with the system clock.
594 - Document the 1.57 change better.
597 - Win32/Cygwin/MinGW: if the performance counter drifts by more
598 than two seconds from the system clock (due to ntp adjustments,
599 for example), recalibrate our internal counter: from Jan Dubois,
600 based on [cpan #5933] by Jerry D. Hedden.
603 - Give a clearer message if the tests timeout (perl change #22253)
604 - Don't use /tmp or its moral equivalents (perl bug #15036,
608 - Windows: mingw32 patch from Mike Pomraning (use Perl's Const64()
609 instead of VC-specific i64 suffix)
612 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
615 - Windows: higher resolution time() by using the Windows
616 performance counter API, from Jan Dubois and Anton Shcherbinin.
617 The exact new higher resolution depends on the hardware,
618 but it should be quite a bit better than using the basic
622 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
623 the subtest 18. No known analysis nor fix, but added
624 an alarm (that requires fork() and alarm()) to the test.
627 - doc tweaks from mjd (perl change #20456)
628 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
631 - add a message (for non-core builds) to Makefile.PL about
632 the LC_ALL=C workaround
633 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
634 - the nanosleep() probe was broken (perl change #20061)
635 - use existence instead of definedness for feature probes
637 - MPE/iX tweak (perl change #20042)
638 - do not use HAS_NANOSLEEP (perl change #19898)
641 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
642 - OS/2 can always mix subsecond sleeps with signals
643 (part of perl change #19789)
646 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
650 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
651 - unnecessary (nanosleep is in libc anyway)
652 - harmful (-lrt slows down execution)
653 - incompatible (with many distributions' pthreads)
656 - do not create files in blib directories under core
657 (perl change #19160, from rgs)
658 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
661 - guarantee that $xdefine in HiRes.t is always defined
662 (perl change #19109, from IlyaZ)
663 - a cleaner way to detect PERL_CORE (perl change #19111,
667 - add hints/irix.pl to turn off overly POSIX flags that
668 cause hide struct timespec to be hidden (and compilation
669 to fail) (bleadperl change #19085)
670 - documentation tweaks
673 - add c:/temp to the list of temp directories to probe
674 so that cygwin (and win*?) builds are happy. This was
675 needed at least in my cygwin 1.3.20/w2k setup.
678 - modernize the constants code (from Nicholas Clark)
681 - At some point the ability to figure our the correct incdir
682 for EXTERN.h (either a core perl build, or an installed perl)
683 had broken (which lead into all test compiles failing with
684 a core perl build, but thanks to the robustness of Makefile.PL
685 nothing of this was visible). The brokenness seemed to be
686 caused by $ENV{PERL_CORE} not being on for core builds?
687 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
688 right, and both styles of build should work again.
691 - Nicholas Clark noticed that the my_catdir() emulation function
692 was broken (which means that we didn't really work for Perls
694 - inspired by fixing the above made the whole Makefile.PL -w
696 - tightened up the Makefile.PL output, less whitespace
699 - fix from Craig Berry for better building in VMS with PERL_CORE
702 - no functional changes
703 - move lib/Time/HiRes.pm as Hires.pm
704 - libraries scanning was slightly broken (always scanned
705 for a library even when $Config{libs} already had it)
708 - Ray Zimmerman ran into a race condition in Mac OS X.
709 A 0.01-second alarm fired before the test expected.
710 The test first slept indefinitely (blocking for signals)
711 and only after that tested for the signal having been sent.
712 Since the signal had already been sent, the test #12 never
713 completed. The solution: test first, then block.
714 - default to being silent on all probing attempts, set the
715 environment variable VERBOSE to a true value to see the
716 details (the probing command and the possible errors)
719 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
720 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
721 cannot be upgraded from CPAN to override the 5.8.0 version
723 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
724 because the include directories did not adjust themselves
725 if $ENV{PERL_CORE} (Hugo van der Sanden)
726 - add documentation about the restart of select() under alarm()
729 - small documentation tweaks
733 - better VMS operation (Craig Berry)
736 - our time machine is accelerating: now works with Perl 5.004_01
737 (tried with 5.003_07 and 5.002 but I get segmentation faults
738 from running the Makefile.PL with those in Tru64 4.0D)
741 - backward compatibility (pre-5.6.0) tweaks:
742 - no XSLoader in 5.00503, use DynaLoader instead
743 - no SvPV_nolen, either
744 - no PerlProc_pause(), either
745 - now tested with 5.00404 and 5.00503
746 - Makefile.PL requires 5.00404 (no more 5.002)
747 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
748 (this means that one can mix subsecond sleeps with alarms)
749 - because of nanosleep we probe for -lrt and -lposix4
750 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
751 is available by exportable constants Time::HiRes::d_func
752 (since older Perl do not have them in %Config, and even
753 5.8.0 does not probe for nanosleep)
756 - backward compatibility (pre-5.6.1) tweaks:
757 - define NV if no NVTYPE
758 - define IVdf if needed (note: the Devel::PPPort
759 in 5.8.0 does not try hard hard enough since
760 the IVSIZE might not be defined)
761 - define NVgf if needed
762 - grab the typemap from 5.8.0 for the NV stuff
764 1.31 and 1.32 add more backward compatibility (now all the way
765 back to Perl 5.00404), and using nanosleep() (if available) for
770 - release 1.29_02 as 1.30
772 1.30 adds all the changes made during the Perl 5.6->5.7->5.8
773 development cycle. Most notably portability across platforms has been
774 enhanced, and the interval timers (setitimer, getitimer) have been
775 added. Note that the version of Time::HiRes that is included in Perl
776 5.8.0 calls itself 1.20_00, but it is equivalent to this Time::HiRes
777 version. Note also that in 1.30 Wegscheid turns over the maintenance
778 to Jarkko Hietaniemi.
782 - fix a silly unclosed comment typo in HiRes.xs
783 - document and export REALTIME_REALPROF (Solaris)
787 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
788 (need to patch this also in Perl 5.[89])
789 - remove CVS revision log from HiRes.xs
793 The following numbered patches refer to the Perl 5.7 changes,
794 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
796 - 17558: Add #!./perl to the .t
797 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
798 - 16198: political correctness, from Simon Cozens
799 - 15857: doc tweaks, from Jarkko Hietaniemi
800 - 15593: optimization in .xs, from Paul Green
801 - 14892: pod fixes, from Robin Barker
802 - 14100: VOS fixes, from Paul Green
803 - 13422: XS segfault, from Marc Lehmann
804 - 13378: whether select() gets restarted on signals, depends
805 - 13354: timing constraints, again, from Andy Dougherty
806 - 13278: can't do subsecond alarms with ualarm;
807 break out early if alarms do not seem to be working
808 - 13266: test relaxation (cygwin gets lower hires
809 times than lores ones)
810 - 12846: protect against high load, from Jarkko Hietaniemi
811 - 12837: HiRes.t VMS tweak, from Craig A. Berry
812 - 12797: HiRes.t VMS tweak, from Charles Lane
813 - 12769: HiRes.t VMS tweak, from Craig A. Berry
814 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
815 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
816 - 12692: alarm() ain't gonna work if ualarm() ain't,
817 from Gurusamy Sarathy
818 - 12680: minor VMS tweak, from Charles Lane
819 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
820 - 12609: croak on negative time, from Jarkko Hietaniemi
821 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
822 - 12594: MacOS Classic timeofday, from Chris Nandor
823 - 12473: allow for more than one second for sleep() and usleep()
824 - 12458: test tuning, relax timing constraints,
825 from Jarkko Hietaniemi
826 - 12449: make sleep() and usleep() to return the number
827 of seconds and microseconds actually slept (analogously
828 with the builtin sleep()), also make usleep() croak if
829 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
830 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
831 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
832 - 12196: use ftime() on Win32, from Artur Bergman
833 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
834 - 12105: use GetSystemTime() on Win32, from Artur Bergman
835 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
836 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
837 - 11797: problem in HiRes.t, from John P. Linderman
838 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
839 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
840 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
841 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
842 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
843 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
845 - 10942: MPE/IX test tweaks, from Mark Bixby
846 - 10784: unnecessary pod2man calls, from Andy Dougherty
847 - 10354: ext/ + -Wall, from Doug MacEachern
848 - 10320: fix the BOOT section to call myU2time correctly
849 - 10317: correct casting for AIX< from H. Merijn Brand
850 - 10119: document that the core time() may be rounding, not truncating
851 - 10118: test fix, from John Peacock
852 - 9988: long =item, from Robin Barker
853 - 9714: correct test output
854 - 9708: test also the scalar aspect of getitimer()
855 - 9705: Add interval timers (setitimer, getitimer)
856 - 9692: do not require at least 5.005 using XS
858 The following changes were made on top of the changes
859 made for Time::HiRes during the Perl 5.7 development
860 cycle that culminated in the release of Perl 5.8.0.
862 - add "require 5.005" to the Makefile.PL
863 - remove the REVISION section (CVS log) from HiRes.pm
864 - add jhi's copyright alongside Douglas'
865 - move HiRes.pm to lib/Time/
867 - modify HiRes.t to use $ENV{PERL_CORE}
868 - modify the original Time::HiRes version 1.20 Makefile.PL
869 to work both with Perl 5.8.0 and the new code with pre-5.8.0
870 Perls (tried with 5.6.1)
871 - tiny tweaks and updates in README and TODO
872 - bump the VERSION to 1.29
874 1.20 Wed Feb 24 21:30 1999
875 - make our usleep and ualarm substitutes into hrt_usleep
876 and hrt_ualarm. This helps static links of Perl with other
877 packages that also have usleep, etc. From
878 Ilya Zakharevich <ilya@math.ohio-state.edu>
879 - add C API stuff. From Joshua Pritikin
880 <joshua.pritikin@db.com>
881 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
882 - hopefully correct "-lc" fix for SCO.
885 1.20 adds a platform neutral set of C accessible routines if you are
886 running 5.005+. All other changes are packaging changes and build
887 fixes(?) for statically linked Perl, SCO, and VMS.
889 1.19 Tue Sep 29 22:30 1998
890 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
891 <seb@stian.demon.co.uk>
892 - change GIMME_V to GIMME to help people with older versions of
894 - fix Win32 version of gettimeofday(). It didn't affect anything,
895 but it confuses people reading the code when the return value
896 is backwards (0 is success).
897 - fix Makefile.PL (more) so that detection of gettimeofday is
900 1.19 has better VMS support.
902 1.18 Mon Jul 6 22:40 1998
903 - add usleep() for Win32.
904 - fix Makefile.PL to fix reported HP/UX feature where unresolved
905 externals still cause an executable to be generated (though no
906 x bit set). Thanks to David Kozinn for report and explanation.
907 Problems with the fix are mine :)
909 1.18 has limited Win32 support (no ualarm). Added usleep for Win32.
910 Probably buggy. I'm sure I'll hear.
912 1.17 Wed Jul 1 20:10 1998
913 - fix setitimer calls so microseconds is not more than 1000000.
914 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
915 - make Win32. We only get gettimeofday (the select hack doesn't
916 seem to work on my Win95 system).
917 - fix test 4 on 01test.t. add test to see if time() and
918 Time::HiRes::time() are close.
920 1.16 Wed Nov 12 21:05 1997
921 - add missing EXTEND in new gettimeofday scalar code.
923 1.16+ should be closer to building out of the box on Linux. Thanks
924 to Gisle Aas for patches, and the ualarm equivalent using setitimer.
926 If your underlying operating system doesn't implement ualarm(), then
927 a fake using setitimer() will be made. If the OS is missing usleep(),
928 a fake one using select() will be made. If a fake can't be made for
929 either ualarm() or usleep(), then the corresponding Perl function will
930 not be available. If the OS is missing gettimeofday(), you will get
931 unresolved externals, either at link- or run-time.
933 This is an improvement; the package used to not even build if
934 you were missing any of these bits. Roderick Schertler
936 <roderick@argon.org> did all the conditional compilation stuff,
937 look at HiRes.pm and the test suites; it's good educational reading.
939 1.15 Mon Nov 10 21:30 1997
940 - HiRes.pm: update pod. Provided by Gisle Aas.
941 - HiRes.xs: if gettimeofday() called in scalar context, do
942 something more useful than before. Provided by Gisle Aas.
943 - README: tell of xsubpp '-nolinenumber' woes. thanks to
944 Edward Henigin <ed@texas.net> for pointing out the problem.
946 1.14 Wed Nov 5 9:40 1997
947 - Makefile.PL: look for setitimer
948 - HiRes.xs: if missing ualarm, but we have setitimer, make up
949 our own setitimer. These were provided by Gisle Aas.
951 1.13 Tue Nov 4 23:30 1997
952 - Makefile.PL: fix autodetect mechanism to do try linking in addition
953 to just compiling; should fix Linux build problem. Fix was provided
956 1.12 Sun Oct 12 12:00:00 1997
957 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
958 you may need to comment this back out if you have an older xsubpp.
959 - HiRes.xs: set PROTOTYPES: DISABLE
961 1.11 Fri Sep 05 16:00:00 1997
963 Had some line commented out that shouldn't have been (testing
966 Previous version was corrupted.
968 1.10 Thu May 22 20:20:00 1997
969 - HiRes.xs, HiRes.pm, t/*:
970 - only compile what we have OS support for (or can
972 - only test what we compiled
973 - gross improvement to the test suite
975 This work was all done by Roderick Schertler
976 <roderick@argon.org>. If you run Linux or
977 one of the other ualarm-less platforms, and you like this
978 module, let Roderick know; without him, it still wouldn't
979 be working on those boxes...
980 - Makefile.PL: figure out what routines the OS has and
981 only build what we need. These bits were written by Jarkko
982 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
984 1.02 Mon Dec 30 08:00:00 1996
985 - HiRes.pm: update documentation to say what to do when missing
986 ualarm() and friends.
987 - README: update to warn that ualarm() and friends need to exist
989 1.01 Fri Oct 17 08:00:00 1996
990 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
991 - HiRes.pm: put blank line between __END__ and =head1 so that
994 1.00 Tue Sep 03 13:00:00 1996
995 - original version; created by h2xs 1.16