1 Revision history for the Perl extension Time::HiRes.
4 - the ualarm() tests (34-37) assumed that ualarm(N)
5 could never alarm in less than N seconds, widened
6 the acceptable relative range to 0.9..1.5. Addresses
7 [rt.cpan.org #22090] and [rt.cpan.org #22091].
9 - skip the stat() tests in cygwin and win32, because
10 if run on FAT the timestamp granularity is only 2 seconds.
11 Any good way to detect (cygwin or win32) whether we are
12 being run on NTFS or anyplace with better timestamps?
13 Addresses [rt.cpan.org #22089] and [rt.cpan.org #22098].
16 - scan for subsecond resolution timestamps in struct stat,
17 some known possibilities:
19 (1) struct timespec st_atimespec;
25 (4) timestruc_t st_atim;
30 If something like this is found, one can do
33 my @stat = Time::HiRes::stat();
35 or even override the standard stat():
37 use Time::HiRes qw(stat);
39 to get the stat() timestamps
41 my ($atime, $mtime, $ctime) = @stat[8, 9, 10];
43 with subsecond resolution (assuming both the operating
44 system and the filesystem support that kind of thing).
46 Contributions for more systems (especially non-UNIX,
47 e.g. but not limited to, Win32, VMS) gladly accepted.
49 Thanks to H.Merijn Brand, John Peacock, and Craig
50 Berry for brave beta testing.
53 - ualarm() in SuSE 10.1 was overflowing after ~4.2 seconds,
54 possibly due to a glibc bug/feature, workaround by using the
55 setitimer() implementation of ualarm() if either useconds or
56 interval > 999_999 (this case seems to vary between systems:
57 are useconds more than 999_999 for ualarm() defined or not)
58 Added more ualarm() tests to catch various overflow points,
59 hopefully no problems in various platforms.
60 (The problem report by Mark Seger and Jon Paul Sullivan of HP.)
63 - tweak still needed for Const64(), from Jerry Hedden
64 - get a freshly generated ppport.h
65 - update Copyright years
68 - Const64() already appends an 'LL' (or i64), so provide LL and i64
69 forms for the IV_1E[679] (effects Win32 and Cygwin), reported by
71 - the Changes entry for 1.88 talked about [IN]V_1[679],
75 - clean up the g++ warnings in HiRes.xs, all of them
76 about mixing integer and floating point, introduce
77 constants IV_1E[679] and NV_1E[679]
80 - [rt.cpan.org #17442] 'make test' frequently fails under
81 Cygwin Perl v5.8.8, reported and patched by J. R. Hedden
82 (two race condition bugs in the END block in the case the
83 main process dies before the timer process, unearthed
84 by a bug in Cygwin ualarm)
87 - HiRes.t:s/ok 32/ok 33/, from Dominic Dunlop
88 - tighten up the clock() test marginally by requiring non-negative
89 - clock_nanosleep() and clock() doc tweaks
92 - the interface to clock_nanosleep() is more natural
93 when it is like (hires) time() (instead of like nanosleep),
94 and the .xs implementation of clock_nanosleep() in 1.84
96 - the semantics of clock() are not quite so silly as I thought,
97 but still somewhat odd, documented as such
98 - additional enhancements to the clock() documentation
99 - add test for clock_nanosleep() (I cannot test this
100 since none of my systems have the function)
101 - add test for clock()
104 - add clock() which returns the processor time in
105 (floating point) seconds since an arbitrary era
106 - add clock_nanosleep() which suspends the current
107 thread until either absolute time or for relative time
108 - [rt.cpan.org #16486] printf missing value in HiRes.t
109 - add constants CLOCKS_PER_SEC, CLOCK_SOFTTIME, TIMER_ABSTIME
113 - has_symbol() was wrong since e.g. ITIMER_VIRTUAL is exported
114 via @EXPORT_OK even when it is not available. This is heinous.
115 @EXPORT_OK should be determined at Makefile.PL time.
116 - be more lenient is testing clock_gettime(): allow more slop,
117 and retry up to three times, sleeping a random nap between
119 - human months are one-based (noticed by Anton Berezin)
122 - CLOCK_REALTIME is an enum value (of the clockid_t enum)
123 in HP-UX (and might be so elsewhere, too), debugged by
125 - include const-c.inc as late as possible (from Randy Kobes,
126 [rt.cpan.org #15552] to avoid undefined usleep() on Win32
129 - try to be more robust and consistent in the detection of
130 CLOCK_REALTIME and ITIMER_VIRTUAL in HiRes.t: the proper
135 eval 'import Time::HiRes qw($symbol)';
136 return 0 unless $@ eq '';
137 return exists ${"Time::HiRes::$symbol"};
144 in the test. All these moves are needed because
146 1) one cannot directly do eval 'Time::HiRes::FOO_BAR'
147 because FOO_BAR might have a true value of zero
148 (or in the general case an empty string or even undef)
150 2) In case FOO_BAR is not available in this platform,
151 &FOO_BAR avoids the bareword warning
153 - wait more (1.5 seconds instead of 0.1) for the CLOCK_REALTIME test
154 but expect the 'customary' slop of 0.20 instead of 0.25
155 - fixed inside a comment HAS_POLL -> TIME_HIRES_NANOSLEEP
156 - at the end of HiRest.t tell how close we were to termination
159 - Gisle noticed a mistake (using HAS_NANOSLEEP) in 1.79
162 - try nanosleep for emulating usleep -- may help in some weird
163 embedded realtime places which have nanosleep but neither usleep
164 nor select nor poll (doesn't have to be weird embedded realtime
165 place, though -- in many places usleep is nanosleep anyway)
166 - try poll for emulating usleep -- this may help some obscure/old
167 SVR4 places that have neither usleep nor select
168 - a redundant test guard in HiRes.t
171 - ITIMER_VIRTUAL detection in HiRes.t had problems (that we cannot
172 in the general case fail already at 'use' phase is suboptimal)
173 - fixes to the documentation of clock_gettime() and clock_getres()
176 - add support for the POSIX clock_gettime() and clock_getres(),
177 if available, either as library calls or as syscalls
178 - be more defensive about missing functionality: break out
179 early (during 'use') if no e.g. clock_getres() is available,
180 and protect our back by trapping those cases also in HiRes.xs
181 - the test added in 1.76 could cause an endless loop e.g. in Solaris,
182 due to mixing of sleep() and alarm() (bad programmer, no cookie!)
185 - testing for nanosleep had wrong logic which caused nanosleep
186 to become undefined for e.g. Mac OS X
187 - added a test for a core dump that was introduced by Perl 5.8.0
188 safe signals and was fixed for the time of 5.8.1 (one report of
189 the core dump was [perl #20920]), the test skipped pre-5.8.1.
190 - *cough* s/unanosleep/nanosleep/g; *cough*
193 - installation patch from Gisle Aas: in Perls 5.8.x and later
194 use MakeMaker INSTALLDIRS value of 'perl' instead of 'site'.
197 - [cpan #14608] Solaris 8 perl 5.005_03 File::Spec module does not have method rel2abs
198 (the workaround is not to use rel2abs, should not be necessary)
199 - [cpan #14642] U2time wrongly exported on the C API
200 (patch supplied by the reporter, SALVA@cpan.org)
201 - add release dates to Changes
204 - Time::HiRes::nanosleep support for Solaris [PATCH]
205 (POSIX::uname() not available if building with core perl,
206 from Gisle Aas, via perl5-porters, perl change #25295)
209 - going back to the 1.68 loader setup (using DynaLoader)
210 since too many weird things starting breaking
211 - fix a typo in José Auguste-Etienne's name
214 - a thinko in the nanosleep() detection
215 - move more changes stuff from the README to Changes
216 - add -w to the Makefile.PL
219 - oops in 1.69 about @ISA (not affecting anything but silly)
220 - add copyright 2005 to HiRes.pm
221 - add copyright and license to HiRes.xs
222 - add copyrights 2003, 2004, 2005 to README
225 - actually run a test for nanosleep
226 (if there is no $Config{d_nanosleep}) since e.g. in AIX 4.2
227 it seems that one can link in nanosleep() but then calling
228 it fails instantly and sets errno to ENOSYS (Not implemented).
229 This may be fixable in the AIX case by figuring out the right
230 (realtime POSIX?) libs and whatnot, but in the general case
231 running a real test case is better. (Of course, this change
232 will no doubt run into portability problems because of the
233 execution step...) Note that because of hysterical raisins
234 most Perls do NOT have $Config{d_nanosleep} (scanning for
235 it by Configure would in many platforms require linking in
236 things like -lrt, which would in many platforms be a bad idea
238 (from José Auguste-Etienne)
239 - support XSLoader also since it's much faster
240 (from Alexey Tourbin)
241 - add SEE ALSO (BSD::Resource and Time::TAI64)
244 - somehow 1.67 had a lot of doubled lines (a major cut-and-paste
245 error suspected), but miraculously it still worked since the
246 doubling took place below the __END__ token
247 - undef Pause() before defining it to avoid redefinition warnings
248 during compilation in case perl.h had already defined Pause()
249 (part of perl change #24271)
253 - (internal) don't ignore the return value of gettimeofday()
254 - (external) return undef or an empty if the C gettimeofday() fails
255 (affects Time::HiRes gettimeofday() and the hires time())
259 - fix the 'hierachy' typo in Makefile.PL [rt.cpan.org #8492]
260 - should now build in Solaris [rt.cpan.org #7165] (since 1.64)
261 - should now build in Cygwin [rt.cpan.org #7535] (since 1.64)
262 - close also [rt.cpan.org #5933] "Time::HiRes::time does not
263 pick up time adjustments like ntp" since ever reproducing it
264 (and therefore verifying a possible fix) in the same environment
265 has become rather unlikely
268 - one should not mix u?alarm and sleep (the tests modified
269 by 1.65, #12 and #13, hung in Solaris), now we just busy
270 loop executing an empty block
271 - in the documentation underline the unspecificity of mixing
273 - small spelling fixes
276 - regenerate ppport.h with Devel::PPPort 3.03,
277 now the MY_CXT_CLONE is defined in ppport.h,
278 we no more need to do that.
280 - the test #12 would often hang in sigsuspend() (at least that's
281 where Mac OS X' ktrace shows it hanging). With the sleep()s
282 changed to sleep(1)s, the tests still pass but no hang after
283 a few hundred repeats.
286 - Win32 and any ithread build: ppport.h didn't define
287 MY_CXT_CLONE, which seems to be a Time-HiRes-ism.
290 - Skip testing if under PERL_CORE and Time::HiRes has not
291 been Configured (from Marcus Holland-Moritz, core change
293 - Use ppport.h generated by Devel::PPPort 3.01,
294 allowing cutting away our own portability code.
295 - Don't use $ENV{PERL_CORE} for < 5.6.0.
296 - Don't use "for my $i" for <= 5.003.
297 - Don't use Pause() for <= 5.003.
298 - Can't use newSVpvf for <= 5.003.
299 (most of the changes from Marcus)
302 - Win32: reset reading from the performance counters every
303 five minutes to better track wall clock time (thanks to
304 PC timers being often quite bad), should help long-running
308 - Win32: Patch from Steve Hay
309 [PATCH] Re: [perl #30755] [Win32] Different results from Time::HiRes::gettimeofdayunder the debugger
310 to [perl #30755] reported by Nigel Sandever
312 - Cygwin: Use the Win32 recalibration code also in Cygwin if the
313 <w32api/windows.h> APIs are available. Cygwin testing by
314 Yitzchak Scott-Thoennes.
316 - Solaris: use -lposix4 to get nanosleep for Solaris 2.6,
317 after that keep using -lrt, patch from Alan Burlison,
318 bug reported in [cpan #7165]
321 - Change the Win32 recalibration limit to 0.5 seconds and tweak
322 the documentation to blather less about the gory details of the
323 Win32 implementation and more about the complications in general
324 of meddling with the system clock.
327 - Document the 1.57 change better.
330 - Win32/Cygwin/MinGW: if the performance counter drifts by more
331 than two seconds from the system clock (due to ntp adjustments,
332 for example), recalibrate our internal counter: from Jan Dubois,
333 based on [cpan #5933] by Jerry D. Hedden.
336 - Give a clearer message if the tests timeout (perl change #22253)
337 - Don't use /tmp or its moral equivalents (perl bug #15036,
341 - Windows: mingw32 patch from Mike Pomraning (use Perl's Const64()
342 instead of VC-specific i64 suffix)
345 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
348 - Windows: higher resolution time() by using the Windows
349 performance counter API, from Jan Dubois and Anton Shcherbinin.
350 The exact new higher resolution depends on the hardware,
351 but it should be quite a bit better than using the basic
355 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
356 the subtest 18. No known analysis nor fix, but added
357 an alarm (that requires fork() and alarm()) to the test.
360 - doc tweaks from mjd (perl change #20456)
361 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
364 - add a message (for non-core builds) to Makefile.PL about
365 the LC_ALL=C workaround
366 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
367 - the nanosleep() probe was broken (perl change #20061)
368 - use existence instead of definedness for feature probes
370 - MPE/iX tweak (perl change #20042)
371 - do not use HAS_NANOSLEEP (perl change #19898)
374 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
375 - OS/2 can always mix subsecond sleeps with signals
376 (part of perl change #19789)
379 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
383 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
384 - unnecessary (nanosleep is in libc anyway)
385 - harmful (-lrt slows down execution)
386 - incompatible (with many distributions' pthreads)
389 - do not create files in blib directories under core
390 (perl change #19160, from rgs)
391 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
394 - guarantee that $xdefine in HiRes.t is always defined
395 (perl change #19109, from IlyaZ)
396 - a cleaner way to detect PERL_CORE (perl change #19111,
400 - add hints/irix.pl to turn off overly POSIX flags that
401 cause hide struct timespec to be hidden (and compilation
402 to fail) (bleadperl change #19085)
403 - documentation tweaks
406 - add c:/temp to the list of temp directories to probe
407 so that cygwin (and win*?) builds are happy. This was
408 needed at least in my cygwin 1.3.20/w2k setup.
411 - modernize the constants code (from Nicholas Clark)
414 - At some point the ability to figure our the correct incdir
415 for EXTERN.h (either a core perl build, or an installed perl)
416 had broken (which lead into all test compiles failing with
417 a core perl build, but thanks to the robustness of Makefile.PL
418 nothing of this was visible). The brokenness seemed to be
419 caused by $ENV{PERL_CORE} not being on for core builds?
420 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
421 right, and both styles of build should work again.
424 - Nicholas Clark noticed that the my_catdir() emulation function
425 was broken (which means that we didn't really work for Perls
427 - inspired by fixing the above made the whole Makefile.PL -w
429 - tightened up the Makefile.PL output, less whitespace
432 - fix from Craig Berry for better building in VMS with PERL_CORE
435 - no functional changes
436 - move lib/Time/HiRes.pm as Hires.pm
437 - libraries scanning was slightly broken (always scanned
438 for a library even when $Config{libs} already had it)
441 - Ray Zimmerman ran into a race condition in Mac OS X.
442 A 0.01-second alarm fired before the test expected.
443 The test first slept indefinitely (blocking for signals)
444 and only after that tested for the signal having been sent.
445 Since the signal had already been sent, the test #12 never
446 completed. The solution: test first, then block.
447 - default to being silent on all probing attempts, set the
448 environment variable VERBOSE to a true value to see the
449 details (the probing command and the possible errors)
452 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
453 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
454 cannot be upgraded from CPAN to override the 5.8.0 version
456 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
457 because the include directories did not adjust themselves
458 if $ENV{PERL_CORE} (Hugo van der Sanden)
459 - add documentation about the restart of select() under alarm()
462 - small documentation tweaks
466 - better VMS operation (Craig Berry)
469 - our time machine is accelerating: now works with Perl 5.004_01
470 (tried with 5.003_07 and 5.002 but I get segmentation faults
471 from running the Makefile.PL with those in Tru64 4.0D)
474 - backward compatibility (pre-5.6.0) tweaks:
475 - no XSLoader in 5.00503, use DynaLoader instead
476 - no SvPV_nolen, either
477 - no PerlProc_pause(), either
478 - now tested with 5.00404 and 5.00503
479 - Makefile.PL requires 5.00404 (no more 5.002)
480 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
481 (this means that one can mix subsecond sleeps with alarms)
482 - because of nanosleep we probe for -lrt and -lposix4
483 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
484 is available by exportable constants Time::HiRes::d_func
485 (since older Perl do not have them in %Config, and even
486 5.8.0 does not probe for nanosleep)
489 - backward compatibility (pre-5.6.1) tweaks:
490 - define NV if no NVTYPE
491 - define IVdf if needed (note: the Devel::PPPort
492 in 5.8.0 does not try hard hard enough since
493 the IVSIZE might not be defined)
494 - define NVgf if needed
495 - grab the typemap from 5.8.0 for the NV stuff
497 1.31 and 1.32 add more backward compatibility (now all the way
498 back to Perl 5.00404), and using nanosleep() (if available) for
503 - release 1.29_02 as 1.30
505 1.30 adds all the changes made during the Perl 5.6->5.7->5.8
506 development cycle. Most notably portability across platforms has been
507 enhanced, and the interval timers (setitimer, getitimer) have been
508 added. Note that the version of Time::HiRes that is included in Perl
509 5.8.0 calls itself 1.20_00, but it is equivalent to this Time::HiRes
510 version. Note also that in 1.30 Wegscheid turns over the maintenance
511 to Jarkko Hietaniemi.
515 - fix a silly unclosed comment typo in HiRes.xs
516 - document and export REALTIME_REALPROF (Solaris)
520 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
521 (need to patch this also in Perl 5.[89])
522 - remove CVS revision log from HiRes.xs
526 The following numbered patches refer to the Perl 5.7 changes,
527 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
529 - 17558: Add #!./perl to the .t
530 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
531 - 16198: political correctness, from Simon Cozens
532 - 15857: doc tweaks, from Jarkko Hietaniemi
533 - 15593: optimization in .xs, from Paul Green
534 - 14892: pod fixes, from Robin Barker
535 - 14100: VOS fixes, from Paul Green
536 - 13422: XS segfault, from Marc Lehmann
537 - 13378: whether select() gets restarted on signals, depends
538 - 13354: timing constraints, again, from Andy Dougherty
539 - 13278: can't do subsecond alarms with ualarm;
540 break out early if alarms do not seem to be working
541 - 13266: test relaxation (cygwin gets lower hires
542 times than lores ones)
543 - 12846: protect against high load, from Jarkko Hietaniemi
544 - 12837: HiRes.t VMS tweak, from Craig A. Berry
545 - 12797: HiRes.t VMS tweak, from Charles Lane
546 - 12769: HiRes.t VMS tweak, from Craig A. Berry
547 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
548 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
549 - 12692: alarm() ain't gonna work if ualarm() ain't,
550 from Gurusamy Sarathy
551 - 12680: minor VMS tweak, from Charles Lane
552 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
553 - 12609: croak on negative time, from Jarkko Hietaniemi
554 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
555 - 12594: MacOS Classic timeofday, from Chris Nandor
556 - 12473: allow for more than one second for sleep() and usleep()
557 - 12458: test tuning, relax timing constraints,
558 from Jarkko Hietaniemi
559 - 12449: make sleep() and usleep() to return the number
560 of seconds and microseconds actually slept (analogously
561 with the builtin sleep()), also make usleep() croak if
562 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
563 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
564 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
565 - 12196: use ftime() on Win32, from Artur Bergman
566 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
567 - 12105: use GetSystemTime() on Win32, from Artur Bergman
568 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
569 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
570 - 11797: problem in HiRes.t, from John P. Linderman
571 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
572 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
573 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
574 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
575 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
576 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
578 - 10942: MPE/IX test tweaks, from Mark Bixby
579 - 10784: unnecessary pod2man calls, from Andy Dougherty
580 - 10354: ext/ + -Wall, from Doug MacEachern
581 - 10320: fix the BOOT section to call myU2time correctly
582 - 10317: correct casting for AIX< from H. Merijn Brand
583 - 10119: document that the core time() may be rounding, not truncating
584 - 10118: test fix, from John Peacock
585 - 9988: long =item, from Robin Barker
586 - 9714: correct test output
587 - 9708: test also the scalar aspect of getitimer()
588 - 9705: Add interval timers (setitimer, getitimer)
589 - 9692: do not require at least 5.005 using XS
591 The following changes were made on top of the changes
592 made for Time::HiRes during the Perl 5.7 development
593 cycle that culminated in the release of Perl 5.8.0.
595 - add "require 5.005" to the Makefile.PL
596 - remove the REVISION section (CVS log) from HiRes.pm
597 - add jhi's copyright alongside Douglas'
598 - move HiRes.pm to lib/Time/
600 - modify HiRes.t to use $ENV{PERL_CORE}
601 - modify the original Time::HiRes version 1.20 Makefile.PL
602 to work both with Perl 5.8.0 and the new code with pre-5.8.0
603 Perls (tried with 5.6.1)
604 - tiny tweaks and updates in README and TODO
605 - bump the VERSION to 1.29
607 1.20 Wed Feb 24 21:30 1999
608 - make our usleep and ualarm substitutes into hrt_usleep
609 and hrt_ualarm. This helps static links of Perl with other
610 packages that also have usleep, etc. From
611 Ilya Zakharevich <ilya@math.ohio-state.edu>
612 - add C API stuff. From Joshua Pritikin
613 <joshua.pritikin@db.com>
614 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
615 - hopefully correct "-lc" fix for SCO.
618 1.20 adds a platform neutral set of C accessible routines if you are
619 running 5.005+. All other changes are packaging changes and build
620 fixes(?) for statically linked Perl, SCO, and VMS.
622 1.19 Tue Sep 29 22:30 1998
623 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
624 <seb@stian.demon.co.uk>
625 - change GIMME_V to GIMME to help people with older versions of
627 - fix Win32 version of gettimeofday(). It didn't affect anything,
628 but it confuses people reading the code when the return value
629 is backwards (0 is success).
630 - fix Makefile.PL (more) so that detection of gettimeofday is
633 1.19 has better VMS support.
635 1.18 Mon Jul 6 22:40 1998
636 - add usleep() for Win32.
637 - fix Makefile.PL to fix reported HP/UX feature where unresolved
638 externals still cause an executable to be generated (though no
639 x bit set). Thanks to David Kozinn for report and explanation.
640 Problems with the fix are mine :)
642 1.18 has limited Win32 support (no ualarm). Added usleep for Win32.
643 Probably buggy. I'm sure I'll hear.
645 1.17 Wed Jul 1 20:10 1998
646 - fix setitimer calls so microseconds is not more than 1000000.
647 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
648 - make Win32. We only get gettimeofday (the select hack doesn't
649 seem to work on my Win95 system).
650 - fix test 4 on 01test.t. add test to see if time() and
651 Time::HiRes::time() are close.
653 1.16 Wed Nov 12 21:05 1997
654 - add missing EXTEND in new gettimeofday scalar code.
656 1.16+ should be closer to building out of the box on Linux. Thanks
657 to Gisle Aas for patches, and the ualarm equivalent using setitimer.
659 If your underlying operating system doesn't implement ualarm(), then
660 a fake using setitimer() will be made. If the OS is missing usleep(),
661 a fake one using select() will be made. If a fake can't be made for
662 either ualarm() or usleep(), then the corresponding Perl function will
663 not be available. If the OS is missing gettimeofday(), you will get
664 unresolved externals, either at link- or run-time.
666 This is an improvement; the package used to not even build if
667 you were missing any of these bits. Roderick Schertler
669 <roderick@argon.org> did all the conditional compilation stuff,
670 look at HiRes.pm and the test suites; it's good educational reading.
672 1.15 Mon Nov 10 21:30 1997
673 - HiRes.pm: update pod. Provided by Gisle Aas.
674 - HiRes.xs: if gettimeofday() called in scalar context, do
675 something more useful than before. Provided by Gisle Aas.
676 - README: tell of xsubpp '-nolinenumber' woes. thanks to
677 Edward Henigin <ed@texas.net> for pointing out the problem.
679 1.14 Wed Nov 5 9:40 1997
680 - Makefile.PL: look for setitimer
681 - HiRes.xs: if missing ualarm, but we have setitimer, make up
682 our own setitimer. These were provided by Gisle Aas.
684 1.13 Tue Nov 4 23:30 1997
685 - Makefile.PL: fix autodetect mechanism to do try linking in addition
686 to just compiling; should fix Linux build problem. Fix was provided
689 1.12 Sun Oct 12 12:00:00 1997
690 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
691 you may need to comment this back out if you have an older xsubpp.
692 - HiRes.xs: set PROTOTYPES: DISABLE
694 1.11 Fri Sep 05 16:00:00 1997
696 Had some line commented out that shouldn't have been (testing
699 Previous version was corrupted.
701 1.10 Thu May 22 20:20:00 1997
702 - HiRes.xs, HiRes.pm, t/*:
703 - only compile what we have OS support for (or can
705 - only test what we compiled
706 - gross improvement to the test suite
708 This work was all done by Roderick Schertler
709 <roderick@argon.org>. If you run Linux or
710 one of the other ualarm-less platforms, and you like this
711 module, let Roderick know; without him, it still wouldn't
712 be working on those boxes...
713 - Makefile.PL: figure out what routines the OS has and
714 only build what we need. These bits were written by Jarkko
715 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
717 1.02 Mon Dec 30 08:00:00 1996
718 - HiRes.pm: update documentation to say what to do when missing
719 ualarm() and friends.
720 - README: update to warn that ualarm() and friends need to exist
722 1.01 Fri Oct 17 08:00:00 1996
723 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
724 - HiRes.pm: put blank line between __END__ and =head1 so that
727 1.00 Tue Sep 03 13:00:00 1996
728 - original version; created by h2xs 1.16