1 Revision history for the Perl extension Time::HiRes.
4 - 1.95 broke builds for threaded Perls
5 - use main() prototype consistently in Makefile.PL
8 - integrate core change #29180: Silence VC++ compiler warnings
10 - do not use PL_ppaddr in stat() because that is not available
12 - regenerate fallback/*.inc for older Perls without
13 ExtUtils::Constant because of d_hires_stat, resolves
14 [rt.cpan.org #23694] Time::HiRes fails tests on Solaris and Perl 5.6.1
17 - file timestamps oddities seen: the atime and mtime
18 can be out of sync (modify first and read second can leave
19 atime < mtime) and mtime can be subsecond while atime is not.
20 So make the test more forgiving.
23 - the ualarm() tests (34-37) assumed that ualarm(N)
24 could never alarm in less than N seconds, widened
25 the acceptable relative range to 0.9..1.5. Addresses
26 [rt.cpan.org #22090] and [rt.cpan.org #22091].
28 - skip the stat() tests in cygwin and win32, because
29 if run on FAT the timestamp granularity is only 2 seconds.
30 Any good way to detect (cygwin or win32) whether we are
31 being run on NTFS or anywhere with better timestamps?
32 Addresses [rt.cpan.org #22089] and [rt.cpan.org #22098].
35 - scan for subsecond resolution timestamps in struct stat,
36 some known possibilities:
38 (1) struct timespec st_atimespec;
44 (4) timestruc_t st_atim;
49 If something like this is found, one can do
52 my @stat = Time::HiRes::stat();
54 or even override the standard stat():
56 use Time::HiRes qw(stat);
58 to get the stat() timestamps
60 my ($atime, $mtime, $ctime) = @stat[8, 9, 10];
62 with subsecond resolution (assuming both the operating
63 system and the filesystem support that kind of thing).
65 Contributions for more systems (especially non-UNIX,
66 e.g. but not limited to: Win32, VMS, OS/2) gladly accepted.
68 Thanks to H.Merijn Brand, John Peacock, and Craig
69 Berry for brave beta testing.
72 - ualarm() in SuSE 10.1 was overflowing after ~4.2 seconds,
73 possibly due to a glibc bug/feature, workaround by using the
74 setitimer() implementation of ualarm() if either useconds or
75 interval > 999_999 (this case seems to vary between systems:
76 are useconds more than 999_999 for ualarm() defined or not)
77 Added more ualarm() tests to catch various overflow points,
78 hopefully no problems in various platforms.
79 (The problem report by Mark Seger and Jon Paul Sullivan of HP.)
82 - tweak still needed for Const64(), from Jerry Hedden
83 - get a freshly generated ppport.h
84 - update Copyright years
87 - Const64() already appends an 'LL' (or i64), so provide LL and i64
88 forms for the IV_1E[679] (effects Win32 and Cygwin), reported by
90 - the Changes entry for 1.88 talked about [IN]V_1[679],
94 - clean up the g++ warnings in HiRes.xs, all of them
95 about mixing integer and floating point, introduce
96 constants IV_1E[679] and NV_1E[679]
99 - [rt.cpan.org #17442] 'make test' frequently fails under
100 Cygwin Perl v5.8.8, reported and patched by J. R. Hedden
101 (two race condition bugs in the END block in the case the
102 main process dies before the timer process, unearthed
103 by a bug in Cygwin ualarm)
106 - HiRes.t:s/ok 32/ok 33/, from Dominic Dunlop
107 - tighten up the clock() test marginally by requiring non-negative
108 - clock_nanosleep() and clock() doc tweaks
111 - the interface to clock_nanosleep() is more natural
112 when it is like (hires) time() (instead of like nanosleep),
113 and the .xs implementation of clock_nanosleep() in 1.84
115 - the semantics of clock() are not quite so silly as I thought,
116 but still somewhat odd, documented as such
117 - additional enhancements to the clock() documentation
118 - add test for clock_nanosleep() (I cannot test this
119 since none of my systems have the function)
120 - add test for clock()
123 - add clock() which returns the processor time in
124 (floating point) seconds since an arbitrary era
125 - add clock_nanosleep() which suspends the current
126 thread until either absolute time or for relative time
127 - [rt.cpan.org #16486] printf missing value in HiRes.t
128 - add constants CLOCKS_PER_SEC, CLOCK_SOFTTIME, TIMER_ABSTIME
132 - has_symbol() was wrong since e.g. ITIMER_VIRTUAL is exported
133 via @EXPORT_OK even when it is not available. This is heinous.
134 @EXPORT_OK should be determined at Makefile.PL time.
135 - be more lenient is testing clock_gettime(): allow more slop,
136 and retry up to three times, sleeping a random nap between
138 - human months are one-based (noticed by Anton Berezin)
141 - CLOCK_REALTIME is an enum value (of the clockid_t enum)
142 in HP-UX (and might be so elsewhere, too), debugged by
144 - include const-c.inc as late as possible (from Randy Kobes,
145 [rt.cpan.org #15552] to avoid undefined usleep() on Win32
148 - try to be more robust and consistent in the detection of
149 CLOCK_REALTIME and ITIMER_VIRTUAL in HiRes.t: the proper
154 eval 'import Time::HiRes qw($symbol)';
155 return 0 unless $@ eq '';
156 return exists ${"Time::HiRes::$symbol"};
163 in the test. All these moves are needed because
165 1) one cannot directly do eval 'Time::HiRes::FOO_BAR'
166 because FOO_BAR might have a true value of zero
167 (or in the general case an empty string or even undef)
169 2) In case FOO_BAR is not available in this platform,
170 &FOO_BAR avoids the bareword warning
172 - wait more (1.5 seconds instead of 0.1) for the CLOCK_REALTIME test
173 but expect the 'customary' slop of 0.20 instead of 0.25
174 - fixed inside a comment HAS_POLL -> TIME_HIRES_NANOSLEEP
175 - at the end of HiRest.t tell how close we were to termination
178 - Gisle noticed a mistake (using HAS_NANOSLEEP) in 1.79
181 - try nanosleep for emulating usleep -- may help in some weird
182 embedded realtime places which have nanosleep but neither usleep
183 nor select nor poll (doesn't have to be weird embedded realtime
184 place, though -- in many places usleep is nanosleep anyway)
185 - try poll for emulating usleep -- this may help some obscure/old
186 SVR4 places that have neither usleep nor select
187 - a redundant test guard in HiRes.t
190 - ITIMER_VIRTUAL detection in HiRes.t had problems (that we cannot
191 in the general case fail already at 'use' phase is suboptimal)
192 - fixes to the documentation of clock_gettime() and clock_getres()
195 - add support for the POSIX clock_gettime() and clock_getres(),
196 if available, either as library calls or as syscalls
197 - be more defensive about missing functionality: break out
198 early (during 'use') if no e.g. clock_getres() is available,
199 and protect our back by trapping those cases also in HiRes.xs
200 - the test added in 1.76 could cause an endless loop e.g. in Solaris,
201 due to mixing of sleep() and alarm() (bad programmer, no cookie!)
204 - testing for nanosleep had wrong logic which caused nanosleep
205 to become undefined for e.g. Mac OS X
206 - added a test for a core dump that was introduced by Perl 5.8.0
207 safe signals and was fixed for the time of 5.8.1 (one report of
208 the core dump was [perl #20920]), the test skipped pre-5.8.1.
209 - *cough* s/unanosleep/nanosleep/g; *cough*
212 - installation patch from Gisle Aas: in Perls 5.8.x and later
213 use MakeMaker INSTALLDIRS value of 'perl' instead of 'site'.
216 - [cpan #14608] Solaris 8 perl 5.005_03 File::Spec module does not have method rel2abs
217 (the workaround is not to use rel2abs, should not be necessary)
218 - [cpan #14642] U2time wrongly exported on the C API
219 (patch supplied by the reporter, SALVA@cpan.org)
220 - add release dates to Changes
223 - Time::HiRes::nanosleep support for Solaris [PATCH]
224 (POSIX::uname() not available if building with core perl,
225 from Gisle Aas, via perl5-porters, perl change #25295)
228 - going back to the 1.68 loader setup (using DynaLoader)
229 since too many weird things starting breaking
230 - fix a typo in José Auguste-Etienne's name
233 - a thinko in the nanosleep() detection
234 - move more changes stuff from the README to Changes
235 - add -w to the Makefile.PL
238 - oops in 1.69 about @ISA (not affecting anything but silly)
239 - add copyright 2005 to HiRes.pm
240 - add copyright and license to HiRes.xs
241 - add copyrights 2003, 2004, 2005 to README
244 - actually run a test for nanosleep
245 (if there is no $Config{d_nanosleep}) since e.g. in AIX 4.2
246 it seems that one can link in nanosleep() but then calling
247 it fails instantly and sets errno to ENOSYS (Not implemented).
248 This may be fixable in the AIX case by figuring out the right
249 (realtime POSIX?) libs and whatnot, but in the general case
250 running a real test case is better. (Of course, this change
251 will no doubt run into portability problems because of the
252 execution step...) Note that because of hysterical raisins
253 most Perls do NOT have $Config{d_nanosleep} (scanning for
254 it by Configure would in many platforms require linking in
255 things like -lrt, which would in many platforms be a bad idea
257 (from José Auguste-Etienne)
258 - support XSLoader also since it's much faster
259 (from Alexey Tourbin)
260 - add SEE ALSO (BSD::Resource and Time::TAI64)
263 - somehow 1.67 had a lot of doubled lines (a major cut-and-paste
264 error suspected), but miraculously it still worked since the
265 doubling took place below the __END__ token
266 - undef Pause() before defining it to avoid redefinition warnings
267 during compilation in case perl.h had already defined Pause()
268 (part of perl change #24271)
272 - (internal) don't ignore the return value of gettimeofday()
273 - (external) return undef or an empty if the C gettimeofday() fails
274 (affects Time::HiRes gettimeofday() and the hires time())
278 - fix the 'hierachy' typo in Makefile.PL [rt.cpan.org #8492]
279 - should now build in Solaris [rt.cpan.org #7165] (since 1.64)
280 - should now build in Cygwin [rt.cpan.org #7535] (since 1.64)
281 - close also [rt.cpan.org #5933] "Time::HiRes::time does not
282 pick up time adjustments like ntp" since ever reproducing it
283 (and therefore verifying a possible fix) in the same environment
284 has become rather unlikely
287 - one should not mix u?alarm and sleep (the tests modified
288 by 1.65, #12 and #13, hung in Solaris), now we just busy
289 loop executing an empty block
290 - in the documentation underline the unspecificity of mixing
292 - small spelling fixes
295 - regenerate ppport.h with Devel::PPPort 3.03,
296 now the MY_CXT_CLONE is defined in ppport.h,
297 we no more need to do that.
299 - the test #12 would often hang in sigsuspend() (at least that's
300 where Mac OS X' ktrace shows it hanging). With the sleep()s
301 changed to sleep(1)s, the tests still pass but no hang after
302 a few hundred repeats.
305 - Win32 and any ithread build: ppport.h didn't define
306 MY_CXT_CLONE, which seems to be a Time-HiRes-ism.
309 - Skip testing if under PERL_CORE and Time::HiRes has not
310 been Configured (from Marcus Holland-Moritz, core change
312 - Use ppport.h generated by Devel::PPPort 3.01,
313 allowing cutting away our own portability code.
314 - Don't use $ENV{PERL_CORE} for < 5.6.0.
315 - Don't use "for my $i" for <= 5.003.
316 - Don't use Pause() for <= 5.003.
317 - Can't use newSVpvf for <= 5.003.
318 (most of the changes from Marcus)
321 - Win32: reset reading from the performance counters every
322 five minutes to better track wall clock time (thanks to
323 PC timers being often quite bad), should help long-running
327 - Win32: Patch from Steve Hay
328 [PATCH] Re: [perl #30755] [Win32] Different results from Time::HiRes::gettimeofdayunder the debugger
329 to [perl #30755] reported by Nigel Sandever
331 - Cygwin: Use the Win32 recalibration code also in Cygwin if the
332 <w32api/windows.h> APIs are available. Cygwin testing by
333 Yitzchak Scott-Thoennes.
335 - Solaris: use -lposix4 to get nanosleep for Solaris 2.6,
336 after that keep using -lrt, patch from Alan Burlison,
337 bug reported in [cpan #7165]
340 - Change the Win32 recalibration limit to 0.5 seconds and tweak
341 the documentation to blather less about the gory details of the
342 Win32 implementation and more about the complications in general
343 of meddling with the system clock.
346 - Document the 1.57 change better.
349 - Win32/Cygwin/MinGW: if the performance counter drifts by more
350 than two seconds from the system clock (due to ntp adjustments,
351 for example), recalibrate our internal counter: from Jan Dubois,
352 based on [cpan #5933] by Jerry D. Hedden.
355 - Give a clearer message if the tests timeout (perl change #22253)
356 - Don't use /tmp or its moral equivalents (perl bug #15036,
360 - Windows: mingw32 patch from Mike Pomraning (use Perl's Const64()
361 instead of VC-specific i64 suffix)
364 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
367 - Windows: higher resolution time() by using the Windows
368 performance counter API, from Jan Dubois and Anton Shcherbinin.
369 The exact new higher resolution depends on the hardware,
370 but it should be quite a bit better than using the basic
374 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
375 the subtest 18. No known analysis nor fix, but added
376 an alarm (that requires fork() and alarm()) to the test.
379 - doc tweaks from mjd (perl change #20456)
380 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
383 - add a message (for non-core builds) to Makefile.PL about
384 the LC_ALL=C workaround
385 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
386 - the nanosleep() probe was broken (perl change #20061)
387 - use existence instead of definedness for feature probes
389 - MPE/iX tweak (perl change #20042)
390 - do not use HAS_NANOSLEEP (perl change #19898)
393 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
394 - OS/2 can always mix subsecond sleeps with signals
395 (part of perl change #19789)
398 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
402 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
403 - unnecessary (nanosleep is in libc anyway)
404 - harmful (-lrt slows down execution)
405 - incompatible (with many distributions' pthreads)
408 - do not create files in blib directories under core
409 (perl change #19160, from rgs)
410 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
413 - guarantee that $xdefine in HiRes.t is always defined
414 (perl change #19109, from IlyaZ)
415 - a cleaner way to detect PERL_CORE (perl change #19111,
419 - add hints/irix.pl to turn off overly POSIX flags that
420 cause hide struct timespec to be hidden (and compilation
421 to fail) (bleadperl change #19085)
422 - documentation tweaks
425 - add c:/temp to the list of temp directories to probe
426 so that cygwin (and win*?) builds are happy. This was
427 needed at least in my cygwin 1.3.20/w2k setup.
430 - modernize the constants code (from Nicholas Clark)
433 - At some point the ability to figure our the correct incdir
434 for EXTERN.h (either a core perl build, or an installed perl)
435 had broken (which lead into all test compiles failing with
436 a core perl build, but thanks to the robustness of Makefile.PL
437 nothing of this was visible). The brokenness seemed to be
438 caused by $ENV{PERL_CORE} not being on for core builds?
439 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
440 right, and both styles of build should work again.
443 - Nicholas Clark noticed that the my_catdir() emulation function
444 was broken (which means that we didn't really work for Perls
446 - inspired by fixing the above made the whole Makefile.PL -w
448 - tightened up the Makefile.PL output, less whitespace
451 - fix from Craig Berry for better building in VMS with PERL_CORE
454 - no functional changes
455 - move lib/Time/HiRes.pm as Hires.pm
456 - libraries scanning was slightly broken (always scanned
457 for a library even when $Config{libs} already had it)
460 - Ray Zimmerman ran into a race condition in Mac OS X.
461 A 0.01-second alarm fired before the test expected.
462 The test first slept indefinitely (blocking for signals)
463 and only after that tested for the signal having been sent.
464 Since the signal had already been sent, the test #12 never
465 completed. The solution: test first, then block.
466 - default to being silent on all probing attempts, set the
467 environment variable VERBOSE to a true value to see the
468 details (the probing command and the possible errors)
471 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
472 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
473 cannot be upgraded from CPAN to override the 5.8.0 version
475 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
476 because the include directories did not adjust themselves
477 if $ENV{PERL_CORE} (Hugo van der Sanden)
478 - add documentation about the restart of select() under alarm()
481 - small documentation tweaks
485 - better VMS operation (Craig Berry)
488 - our time machine is accelerating: now works with Perl 5.004_01
489 (tried with 5.003_07 and 5.002 but I get segmentation faults
490 from running the Makefile.PL with those in Tru64 4.0D)
493 - backward compatibility (pre-5.6.0) tweaks:
494 - no XSLoader in 5.00503, use DynaLoader instead
495 - no SvPV_nolen, either
496 - no PerlProc_pause(), either
497 - now tested with 5.00404 and 5.00503
498 - Makefile.PL requires 5.00404 (no more 5.002)
499 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
500 (this means that one can mix subsecond sleeps with alarms)
501 - because of nanosleep we probe for -lrt and -lposix4
502 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
503 is available by exportable constants Time::HiRes::d_func
504 (since older Perl do not have them in %Config, and even
505 5.8.0 does not probe for nanosleep)
508 - backward compatibility (pre-5.6.1) tweaks:
509 - define NV if no NVTYPE
510 - define IVdf if needed (note: the Devel::PPPort
511 in 5.8.0 does not try hard hard enough since
512 the IVSIZE might not be defined)
513 - define NVgf if needed
514 - grab the typemap from 5.8.0 for the NV stuff
516 1.31 and 1.32 add more backward compatibility (now all the way
517 back to Perl 5.00404), and using nanosleep() (if available) for
522 - release 1.29_02 as 1.30
524 1.30 adds all the changes made during the Perl 5.6->5.7->5.8
525 development cycle. Most notably portability across platforms has been
526 enhanced, and the interval timers (setitimer, getitimer) have been
527 added. Note that the version of Time::HiRes that is included in Perl
528 5.8.0 calls itself 1.20_00, but it is equivalent to this Time::HiRes
529 version. Note also that in 1.30 Wegscheid turns over the maintenance
530 to Jarkko Hietaniemi.
534 - fix a silly unclosed comment typo in HiRes.xs
535 - document and export REALTIME_REALPROF (Solaris)
539 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
540 (need to patch this also in Perl 5.[89])
541 - remove CVS revision log from HiRes.xs
545 The following numbered patches refer to the Perl 5.7 changes,
546 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
548 - 17558: Add #!./perl to the .t
549 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
550 - 16198: political correctness, from Simon Cozens
551 - 15857: doc tweaks, from Jarkko Hietaniemi
552 - 15593: optimization in .xs, from Paul Green
553 - 14892: pod fixes, from Robin Barker
554 - 14100: VOS fixes, from Paul Green
555 - 13422: XS segfault, from Marc Lehmann
556 - 13378: whether select() gets restarted on signals, depends
557 - 13354: timing constraints, again, from Andy Dougherty
558 - 13278: can't do subsecond alarms with ualarm;
559 break out early if alarms do not seem to be working
560 - 13266: test relaxation (cygwin gets lower hires
561 times than lores ones)
562 - 12846: protect against high load, from Jarkko Hietaniemi
563 - 12837: HiRes.t VMS tweak, from Craig A. Berry
564 - 12797: HiRes.t VMS tweak, from Charles Lane
565 - 12769: HiRes.t VMS tweak, from Craig A. Berry
566 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
567 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
568 - 12692: alarm() ain't gonna work if ualarm() ain't,
569 from Gurusamy Sarathy
570 - 12680: minor VMS tweak, from Charles Lane
571 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
572 - 12609: croak on negative time, from Jarkko Hietaniemi
573 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
574 - 12594: MacOS Classic timeofday, from Chris Nandor
575 - 12473: allow for more than one second for sleep() and usleep()
576 - 12458: test tuning, relax timing constraints,
577 from Jarkko Hietaniemi
578 - 12449: make sleep() and usleep() to return the number
579 of seconds and microseconds actually slept (analogously
580 with the builtin sleep()), also make usleep() croak if
581 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
582 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
583 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
584 - 12196: use ftime() on Win32, from Artur Bergman
585 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
586 - 12105: use GetSystemTime() on Win32, from Artur Bergman
587 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
588 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
589 - 11797: problem in HiRes.t, from John P. Linderman
590 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
591 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
592 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
593 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
594 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
595 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
597 - 10942: MPE/IX test tweaks, from Mark Bixby
598 - 10784: unnecessary pod2man calls, from Andy Dougherty
599 - 10354: ext/ + -Wall, from Doug MacEachern
600 - 10320: fix the BOOT section to call myU2time correctly
601 - 10317: correct casting for AIX< from H. Merijn Brand
602 - 10119: document that the core time() may be rounding, not truncating
603 - 10118: test fix, from John Peacock
604 - 9988: long =item, from Robin Barker
605 - 9714: correct test output
606 - 9708: test also the scalar aspect of getitimer()
607 - 9705: Add interval timers (setitimer, getitimer)
608 - 9692: do not require at least 5.005 using XS
610 The following changes were made on top of the changes
611 made for Time::HiRes during the Perl 5.7 development
612 cycle that culminated in the release of Perl 5.8.0.
614 - add "require 5.005" to the Makefile.PL
615 - remove the REVISION section (CVS log) from HiRes.pm
616 - add jhi's copyright alongside Douglas'
617 - move HiRes.pm to lib/Time/
619 - modify HiRes.t to use $ENV{PERL_CORE}
620 - modify the original Time::HiRes version 1.20 Makefile.PL
621 to work both with Perl 5.8.0 and the new code with pre-5.8.0
622 Perls (tried with 5.6.1)
623 - tiny tweaks and updates in README and TODO
624 - bump the VERSION to 1.29
626 1.20 Wed Feb 24 21:30 1999
627 - make our usleep and ualarm substitutes into hrt_usleep
628 and hrt_ualarm. This helps static links of Perl with other
629 packages that also have usleep, etc. From
630 Ilya Zakharevich <ilya@math.ohio-state.edu>
631 - add C API stuff. From Joshua Pritikin
632 <joshua.pritikin@db.com>
633 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
634 - hopefully correct "-lc" fix for SCO.
637 1.20 adds a platform neutral set of C accessible routines if you are
638 running 5.005+. All other changes are packaging changes and build
639 fixes(?) for statically linked Perl, SCO, and VMS.
641 1.19 Tue Sep 29 22:30 1998
642 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
643 <seb@stian.demon.co.uk>
644 - change GIMME_V to GIMME to help people with older versions of
646 - fix Win32 version of gettimeofday(). It didn't affect anything,
647 but it confuses people reading the code when the return value
648 is backwards (0 is success).
649 - fix Makefile.PL (more) so that detection of gettimeofday is
652 1.19 has better VMS support.
654 1.18 Mon Jul 6 22:40 1998
655 - add usleep() for Win32.
656 - fix Makefile.PL to fix reported HP/UX feature where unresolved
657 externals still cause an executable to be generated (though no
658 x bit set). Thanks to David Kozinn for report and explanation.
659 Problems with the fix are mine :)
661 1.18 has limited Win32 support (no ualarm). Added usleep for Win32.
662 Probably buggy. I'm sure I'll hear.
664 1.17 Wed Jul 1 20:10 1998
665 - fix setitimer calls so microseconds is not more than 1000000.
666 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
667 - make Win32. We only get gettimeofday (the select hack doesn't
668 seem to work on my Win95 system).
669 - fix test 4 on 01test.t. add test to see if time() and
670 Time::HiRes::time() are close.
672 1.16 Wed Nov 12 21:05 1997
673 - add missing EXTEND in new gettimeofday scalar code.
675 1.16+ should be closer to building out of the box on Linux. Thanks
676 to Gisle Aas for patches, and the ualarm equivalent using setitimer.
678 If your underlying operating system doesn't implement ualarm(), then
679 a fake using setitimer() will be made. If the OS is missing usleep(),
680 a fake one using select() will be made. If a fake can't be made for
681 either ualarm() or usleep(), then the corresponding Perl function will
682 not be available. If the OS is missing gettimeofday(), you will get
683 unresolved externals, either at link- or run-time.
685 This is an improvement; the package used to not even build if
686 you were missing any of these bits. Roderick Schertler
688 <roderick@argon.org> did all the conditional compilation stuff,
689 look at HiRes.pm and the test suites; it's good educational reading.
691 1.15 Mon Nov 10 21:30 1997
692 - HiRes.pm: update pod. Provided by Gisle Aas.
693 - HiRes.xs: if gettimeofday() called in scalar context, do
694 something more useful than before. Provided by Gisle Aas.
695 - README: tell of xsubpp '-nolinenumber' woes. thanks to
696 Edward Henigin <ed@texas.net> for pointing out the problem.
698 1.14 Wed Nov 5 9:40 1997
699 - Makefile.PL: look for setitimer
700 - HiRes.xs: if missing ualarm, but we have setitimer, make up
701 our own setitimer. These were provided by Gisle Aas.
703 1.13 Tue Nov 4 23:30 1997
704 - Makefile.PL: fix autodetect mechanism to do try linking in addition
705 to just compiling; should fix Linux build problem. Fix was provided
708 1.12 Sun Oct 12 12:00:00 1997
709 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
710 you may need to comment this back out if you have an older xsubpp.
711 - HiRes.xs: set PROTOTYPES: DISABLE
713 1.11 Fri Sep 05 16:00:00 1997
715 Had some line commented out that shouldn't have been (testing
718 Previous version was corrupted.
720 1.10 Thu May 22 20:20:00 1997
721 - HiRes.xs, HiRes.pm, t/*:
722 - only compile what we have OS support for (or can
724 - only test what we compiled
725 - gross improvement to the test suite
727 This work was all done by Roderick Schertler
728 <roderick@argon.org>. If you run Linux or
729 one of the other ualarm-less platforms, and you like this
730 module, let Roderick know; without him, it still wouldn't
731 be working on those boxes...
732 - Makefile.PL: figure out what routines the OS has and
733 only build what we need. These bits were written by Jarkko
734 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
736 1.02 Mon Dec 30 08:00:00 1996
737 - HiRes.pm: update documentation to say what to do when missing
738 ualarm() and friends.
739 - README: update to warn that ualarm() and friends need to exist
741 1.01 Fri Oct 17 08:00:00 1996
742 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
743 - HiRes.pm: put blank line between __END__ and =head1 so that
746 1.00 Tue Sep 03 13:00:00 1996
747 - original version; created by h2xs 1.16