1 Revision history for the Perl extension Time::HiRes.
4 - file timestamps oddities seen: the atime and mtime
5 can be out of sync (modify first and read second can leave
6 atime < mtime) and mtime can be subsecond while atime is not.
7 So make the test more forgiving.
10 - the ualarm() tests (34-37) assumed that ualarm(N)
11 could never alarm in less than N seconds, widened
12 the acceptable relative range to 0.9..1.5. Addresses
13 [rt.cpan.org #22090] and [rt.cpan.org #22091].
15 - skip the stat() tests in cygwin and win32, because
16 if run on FAT the timestamp granularity is only 2 seconds.
17 Any good way to detect (cygwin or win32) whether we are
18 being run on NTFS or anyplace with better timestamps?
19 Addresses [rt.cpan.org #22089] and [rt.cpan.org #22098].
22 - scan for subsecond resolution timestamps in struct stat,
23 some known possibilities:
25 (1) struct timespec st_atimespec;
31 (4) timestruc_t st_atim;
36 If something like this is found, one can do
39 my @stat = Time::HiRes::stat();
41 or even override the standard stat():
43 use Time::HiRes qw(stat);
45 to get the stat() timestamps
47 my ($atime, $mtime, $ctime) = @stat[8, 9, 10];
49 with subsecond resolution (assuming both the operating
50 system and the filesystem support that kind of thing).
52 Contributions for more systems (especially non-UNIX,
53 e.g. but not limited to, Win32, VMS) gladly accepted.
55 Thanks to H.Merijn Brand, John Peacock, and Craig
56 Berry for brave beta testing.
59 - ualarm() in SuSE 10.1 was overflowing after ~4.2 seconds,
60 possibly due to a glibc bug/feature, workaround by using the
61 setitimer() implementation of ualarm() if either useconds or
62 interval > 999_999 (this case seems to vary between systems:
63 are useconds more than 999_999 for ualarm() defined or not)
64 Added more ualarm() tests to catch various overflow points,
65 hopefully no problems in various platforms.
66 (The problem report by Mark Seger and Jon Paul Sullivan of HP.)
69 - tweak still needed for Const64(), from Jerry Hedden
70 - get a freshly generated ppport.h
71 - update Copyright years
74 - Const64() already appends an 'LL' (or i64), so provide LL and i64
75 forms for the IV_1E[679] (effects Win32 and Cygwin), reported by
77 - the Changes entry for 1.88 talked about [IN]V_1[679],
81 - clean up the g++ warnings in HiRes.xs, all of them
82 about mixing integer and floating point, introduce
83 constants IV_1E[679] and NV_1E[679]
86 - [rt.cpan.org #17442] 'make test' frequently fails under
87 Cygwin Perl v5.8.8, reported and patched by J. R. Hedden
88 (two race condition bugs in the END block in the case the
89 main process dies before the timer process, unearthed
90 by a bug in Cygwin ualarm)
93 - HiRes.t:s/ok 32/ok 33/, from Dominic Dunlop
94 - tighten up the clock() test marginally by requiring non-negative
95 - clock_nanosleep() and clock() doc tweaks
98 - the interface to clock_nanosleep() is more natural
99 when it is like (hires) time() (instead of like nanosleep),
100 and the .xs implementation of clock_nanosleep() in 1.84
102 - the semantics of clock() are not quite so silly as I thought,
103 but still somewhat odd, documented as such
104 - additional enhancements to the clock() documentation
105 - add test for clock_nanosleep() (I cannot test this
106 since none of my systems have the function)
107 - add test for clock()
110 - add clock() which returns the processor time in
111 (floating point) seconds since an arbitrary era
112 - add clock_nanosleep() which suspends the current
113 thread until either absolute time or for relative time
114 - [rt.cpan.org #16486] printf missing value in HiRes.t
115 - add constants CLOCKS_PER_SEC, CLOCK_SOFTTIME, TIMER_ABSTIME
119 - has_symbol() was wrong since e.g. ITIMER_VIRTUAL is exported
120 via @EXPORT_OK even when it is not available. This is heinous.
121 @EXPORT_OK should be determined at Makefile.PL time.
122 - be more lenient is testing clock_gettime(): allow more slop,
123 and retry up to three times, sleeping a random nap between
125 - human months are one-based (noticed by Anton Berezin)
128 - CLOCK_REALTIME is an enum value (of the clockid_t enum)
129 in HP-UX (and might be so elsewhere, too), debugged by
131 - include const-c.inc as late as possible (from Randy Kobes,
132 [rt.cpan.org #15552] to avoid undefined usleep() on Win32
135 - try to be more robust and consistent in the detection of
136 CLOCK_REALTIME and ITIMER_VIRTUAL in HiRes.t: the proper
141 eval 'import Time::HiRes qw($symbol)';
142 return 0 unless $@ eq '';
143 return exists ${"Time::HiRes::$symbol"};
150 in the test. All these moves are needed because
152 1) one cannot directly do eval 'Time::HiRes::FOO_BAR'
153 because FOO_BAR might have a true value of zero
154 (or in the general case an empty string or even undef)
156 2) In case FOO_BAR is not available in this platform,
157 &FOO_BAR avoids the bareword warning
159 - wait more (1.5 seconds instead of 0.1) for the CLOCK_REALTIME test
160 but expect the 'customary' slop of 0.20 instead of 0.25
161 - fixed inside a comment HAS_POLL -> TIME_HIRES_NANOSLEEP
162 - at the end of HiRest.t tell how close we were to termination
165 - Gisle noticed a mistake (using HAS_NANOSLEEP) in 1.79
168 - try nanosleep for emulating usleep -- may help in some weird
169 embedded realtime places which have nanosleep but neither usleep
170 nor select nor poll (doesn't have to be weird embedded realtime
171 place, though -- in many places usleep is nanosleep anyway)
172 - try poll for emulating usleep -- this may help some obscure/old
173 SVR4 places that have neither usleep nor select
174 - a redundant test guard in HiRes.t
177 - ITIMER_VIRTUAL detection in HiRes.t had problems (that we cannot
178 in the general case fail already at 'use' phase is suboptimal)
179 - fixes to the documentation of clock_gettime() and clock_getres()
182 - add support for the POSIX clock_gettime() and clock_getres(),
183 if available, either as library calls or as syscalls
184 - be more defensive about missing functionality: break out
185 early (during 'use') if no e.g. clock_getres() is available,
186 and protect our back by trapping those cases also in HiRes.xs
187 - the test added in 1.76 could cause an endless loop e.g. in Solaris,
188 due to mixing of sleep() and alarm() (bad programmer, no cookie!)
191 - testing for nanosleep had wrong logic which caused nanosleep
192 to become undefined for e.g. Mac OS X
193 - added a test for a core dump that was introduced by Perl 5.8.0
194 safe signals and was fixed for the time of 5.8.1 (one report of
195 the core dump was [perl #20920]), the test skipped pre-5.8.1.
196 - *cough* s/unanosleep/nanosleep/g; *cough*
199 - installation patch from Gisle Aas: in Perls 5.8.x and later
200 use MakeMaker INSTALLDIRS value of 'perl' instead of 'site'.
203 - [cpan #14608] Solaris 8 perl 5.005_03 File::Spec module does not have method rel2abs
204 (the workaround is not to use rel2abs, should not be necessary)
205 - [cpan #14642] U2time wrongly exported on the C API
206 (patch supplied by the reporter, SALVA@cpan.org)
207 - add release dates to Changes
210 - Time::HiRes::nanosleep support for Solaris [PATCH]
211 (POSIX::uname() not available if building with core perl,
212 from Gisle Aas, via perl5-porters, perl change #25295)
215 - going back to the 1.68 loader setup (using DynaLoader)
216 since too many weird things starting breaking
217 - fix a typo in José Auguste-Etienne's name
220 - a thinko in the nanosleep() detection
221 - move more changes stuff from the README to Changes
222 - add -w to the Makefile.PL
225 - oops in 1.69 about @ISA (not affecting anything but silly)
226 - add copyright 2005 to HiRes.pm
227 - add copyright and license to HiRes.xs
228 - add copyrights 2003, 2004, 2005 to README
231 - actually run a test for nanosleep
232 (if there is no $Config{d_nanosleep}) since e.g. in AIX 4.2
233 it seems that one can link in nanosleep() but then calling
234 it fails instantly and sets errno to ENOSYS (Not implemented).
235 This may be fixable in the AIX case by figuring out the right
236 (realtime POSIX?) libs and whatnot, but in the general case
237 running a real test case is better. (Of course, this change
238 will no doubt run into portability problems because of the
239 execution step...) Note that because of hysterical raisins
240 most Perls do NOT have $Config{d_nanosleep} (scanning for
241 it by Configure would in many platforms require linking in
242 things like -lrt, which would in many platforms be a bad idea
244 (from José Auguste-Etienne)
245 - support XSLoader also since it's much faster
246 (from Alexey Tourbin)
247 - add SEE ALSO (BSD::Resource and Time::TAI64)
250 - somehow 1.67 had a lot of doubled lines (a major cut-and-paste
251 error suspected), but miraculously it still worked since the
252 doubling took place below the __END__ token
253 - undef Pause() before defining it to avoid redefinition warnings
254 during compilation in case perl.h had already defined Pause()
255 (part of perl change #24271)
259 - (internal) don't ignore the return value of gettimeofday()
260 - (external) return undef or an empty if the C gettimeofday() fails
261 (affects Time::HiRes gettimeofday() and the hires time())
265 - fix the 'hierachy' typo in Makefile.PL [rt.cpan.org #8492]
266 - should now build in Solaris [rt.cpan.org #7165] (since 1.64)
267 - should now build in Cygwin [rt.cpan.org #7535] (since 1.64)
268 - close also [rt.cpan.org #5933] "Time::HiRes::time does not
269 pick up time adjustments like ntp" since ever reproducing it
270 (and therefore verifying a possible fix) in the same environment
271 has become rather unlikely
274 - one should not mix u?alarm and sleep (the tests modified
275 by 1.65, #12 and #13, hung in Solaris), now we just busy
276 loop executing an empty block
277 - in the documentation underline the unspecificity of mixing
279 - small spelling fixes
282 - regenerate ppport.h with Devel::PPPort 3.03,
283 now the MY_CXT_CLONE is defined in ppport.h,
284 we no more need to do that.
286 - the test #12 would often hang in sigsuspend() (at least that's
287 where Mac OS X' ktrace shows it hanging). With the sleep()s
288 changed to sleep(1)s, the tests still pass but no hang after
289 a few hundred repeats.
292 - Win32 and any ithread build: ppport.h didn't define
293 MY_CXT_CLONE, which seems to be a Time-HiRes-ism.
296 - Skip testing if under PERL_CORE and Time::HiRes has not
297 been Configured (from Marcus Holland-Moritz, core change
299 - Use ppport.h generated by Devel::PPPort 3.01,
300 allowing cutting away our own portability code.
301 - Don't use $ENV{PERL_CORE} for < 5.6.0.
302 - Don't use "for my $i" for <= 5.003.
303 - Don't use Pause() for <= 5.003.
304 - Can't use newSVpvf for <= 5.003.
305 (most of the changes from Marcus)
308 - Win32: reset reading from the performance counters every
309 five minutes to better track wall clock time (thanks to
310 PC timers being often quite bad), should help long-running
314 - Win32: Patch from Steve Hay
315 [PATCH] Re: [perl #30755] [Win32] Different results from Time::HiRes::gettimeofdayunder the debugger
316 to [perl #30755] reported by Nigel Sandever
318 - Cygwin: Use the Win32 recalibration code also in Cygwin if the
319 <w32api/windows.h> APIs are available. Cygwin testing by
320 Yitzchak Scott-Thoennes.
322 - Solaris: use -lposix4 to get nanosleep for Solaris 2.6,
323 after that keep using -lrt, patch from Alan Burlison,
324 bug reported in [cpan #7165]
327 - Change the Win32 recalibration limit to 0.5 seconds and tweak
328 the documentation to blather less about the gory details of the
329 Win32 implementation and more about the complications in general
330 of meddling with the system clock.
333 - Document the 1.57 change better.
336 - Win32/Cygwin/MinGW: if the performance counter drifts by more
337 than two seconds from the system clock (due to ntp adjustments,
338 for example), recalibrate our internal counter: from Jan Dubois,
339 based on [cpan #5933] by Jerry D. Hedden.
342 - Give a clearer message if the tests timeout (perl change #22253)
343 - Don't use /tmp or its moral equivalents (perl bug #15036,
347 - Windows: mingw32 patch from Mike Pomraning (use Perl's Const64()
348 instead of VC-specific i64 suffix)
351 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
354 - Windows: higher resolution time() by using the Windows
355 performance counter API, from Jan Dubois and Anton Shcherbinin.
356 The exact new higher resolution depends on the hardware,
357 but it should be quite a bit better than using the basic
361 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
362 the subtest 18. No known analysis nor fix, but added
363 an alarm (that requires fork() and alarm()) to the test.
366 - doc tweaks from mjd (perl change #20456)
367 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
370 - add a message (for non-core builds) to Makefile.PL about
371 the LC_ALL=C workaround
372 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
373 - the nanosleep() probe was broken (perl change #20061)
374 - use existence instead of definedness for feature probes
376 - MPE/iX tweak (perl change #20042)
377 - do not use HAS_NANOSLEEP (perl change #19898)
380 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
381 - OS/2 can always mix subsecond sleeps with signals
382 (part of perl change #19789)
385 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
389 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
390 - unnecessary (nanosleep is in libc anyway)
391 - harmful (-lrt slows down execution)
392 - incompatible (with many distributions' pthreads)
395 - do not create files in blib directories under core
396 (perl change #19160, from rgs)
397 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
400 - guarantee that $xdefine in HiRes.t is always defined
401 (perl change #19109, from IlyaZ)
402 - a cleaner way to detect PERL_CORE (perl change #19111,
406 - add hints/irix.pl to turn off overly POSIX flags that
407 cause hide struct timespec to be hidden (and compilation
408 to fail) (bleadperl change #19085)
409 - documentation tweaks
412 - add c:/temp to the list of temp directories to probe
413 so that cygwin (and win*?) builds are happy. This was
414 needed at least in my cygwin 1.3.20/w2k setup.
417 - modernize the constants code (from Nicholas Clark)
420 - At some point the ability to figure our the correct incdir
421 for EXTERN.h (either a core perl build, or an installed perl)
422 had broken (which lead into all test compiles failing with
423 a core perl build, but thanks to the robustness of Makefile.PL
424 nothing of this was visible). The brokenness seemed to be
425 caused by $ENV{PERL_CORE} not being on for core builds?
426 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
427 right, and both styles of build should work again.
430 - Nicholas Clark noticed that the my_catdir() emulation function
431 was broken (which means that we didn't really work for Perls
433 - inspired by fixing the above made the whole Makefile.PL -w
435 - tightened up the Makefile.PL output, less whitespace
438 - fix from Craig Berry for better building in VMS with PERL_CORE
441 - no functional changes
442 - move lib/Time/HiRes.pm as Hires.pm
443 - libraries scanning was slightly broken (always scanned
444 for a library even when $Config{libs} already had it)
447 - Ray Zimmerman ran into a race condition in Mac OS X.
448 A 0.01-second alarm fired before the test expected.
449 The test first slept indefinitely (blocking for signals)
450 and only after that tested for the signal having been sent.
451 Since the signal had already been sent, the test #12 never
452 completed. The solution: test first, then block.
453 - default to being silent on all probing attempts, set the
454 environment variable VERBOSE to a true value to see the
455 details (the probing command and the possible errors)
458 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
459 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
460 cannot be upgraded from CPAN to override the 5.8.0 version
462 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
463 because the include directories did not adjust themselves
464 if $ENV{PERL_CORE} (Hugo van der Sanden)
465 - add documentation about the restart of select() under alarm()
468 - small documentation tweaks
472 - better VMS operation (Craig Berry)
475 - our time machine is accelerating: now works with Perl 5.004_01
476 (tried with 5.003_07 and 5.002 but I get segmentation faults
477 from running the Makefile.PL with those in Tru64 4.0D)
480 - backward compatibility (pre-5.6.0) tweaks:
481 - no XSLoader in 5.00503, use DynaLoader instead
482 - no SvPV_nolen, either
483 - no PerlProc_pause(), either
484 - now tested with 5.00404 and 5.00503
485 - Makefile.PL requires 5.00404 (no more 5.002)
486 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
487 (this means that one can mix subsecond sleeps with alarms)
488 - because of nanosleep we probe for -lrt and -lposix4
489 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
490 is available by exportable constants Time::HiRes::d_func
491 (since older Perl do not have them in %Config, and even
492 5.8.0 does not probe for nanosleep)
495 - backward compatibility (pre-5.6.1) tweaks:
496 - define NV if no NVTYPE
497 - define IVdf if needed (note: the Devel::PPPort
498 in 5.8.0 does not try hard hard enough since
499 the IVSIZE might not be defined)
500 - define NVgf if needed
501 - grab the typemap from 5.8.0 for the NV stuff
503 1.31 and 1.32 add more backward compatibility (now all the way
504 back to Perl 5.00404), and using nanosleep() (if available) for
509 - release 1.29_02 as 1.30
511 1.30 adds all the changes made during the Perl 5.6->5.7->5.8
512 development cycle. Most notably portability across platforms has been
513 enhanced, and the interval timers (setitimer, getitimer) have been
514 added. Note that the version of Time::HiRes that is included in Perl
515 5.8.0 calls itself 1.20_00, but it is equivalent to this Time::HiRes
516 version. Note also that in 1.30 Wegscheid turns over the maintenance
517 to Jarkko Hietaniemi.
521 - fix a silly unclosed comment typo in HiRes.xs
522 - document and export REALTIME_REALPROF (Solaris)
526 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
527 (need to patch this also in Perl 5.[89])
528 - remove CVS revision log from HiRes.xs
532 The following numbered patches refer to the Perl 5.7 changes,
533 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
535 - 17558: Add #!./perl to the .t
536 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
537 - 16198: political correctness, from Simon Cozens
538 - 15857: doc tweaks, from Jarkko Hietaniemi
539 - 15593: optimization in .xs, from Paul Green
540 - 14892: pod fixes, from Robin Barker
541 - 14100: VOS fixes, from Paul Green
542 - 13422: XS segfault, from Marc Lehmann
543 - 13378: whether select() gets restarted on signals, depends
544 - 13354: timing constraints, again, from Andy Dougherty
545 - 13278: can't do subsecond alarms with ualarm;
546 break out early if alarms do not seem to be working
547 - 13266: test relaxation (cygwin gets lower hires
548 times than lores ones)
549 - 12846: protect against high load, from Jarkko Hietaniemi
550 - 12837: HiRes.t VMS tweak, from Craig A. Berry
551 - 12797: HiRes.t VMS tweak, from Charles Lane
552 - 12769: HiRes.t VMS tweak, from Craig A. Berry
553 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
554 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
555 - 12692: alarm() ain't gonna work if ualarm() ain't,
556 from Gurusamy Sarathy
557 - 12680: minor VMS tweak, from Charles Lane
558 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
559 - 12609: croak on negative time, from Jarkko Hietaniemi
560 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
561 - 12594: MacOS Classic timeofday, from Chris Nandor
562 - 12473: allow for more than one second for sleep() and usleep()
563 - 12458: test tuning, relax timing constraints,
564 from Jarkko Hietaniemi
565 - 12449: make sleep() and usleep() to return the number
566 of seconds and microseconds actually slept (analogously
567 with the builtin sleep()), also make usleep() croak if
568 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
569 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
570 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
571 - 12196: use ftime() on Win32, from Artur Bergman
572 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
573 - 12105: use GetSystemTime() on Win32, from Artur Bergman
574 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
575 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
576 - 11797: problem in HiRes.t, from John P. Linderman
577 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
578 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
579 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
580 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
581 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
582 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
584 - 10942: MPE/IX test tweaks, from Mark Bixby
585 - 10784: unnecessary pod2man calls, from Andy Dougherty
586 - 10354: ext/ + -Wall, from Doug MacEachern
587 - 10320: fix the BOOT section to call myU2time correctly
588 - 10317: correct casting for AIX< from H. Merijn Brand
589 - 10119: document that the core time() may be rounding, not truncating
590 - 10118: test fix, from John Peacock
591 - 9988: long =item, from Robin Barker
592 - 9714: correct test output
593 - 9708: test also the scalar aspect of getitimer()
594 - 9705: Add interval timers (setitimer, getitimer)
595 - 9692: do not require at least 5.005 using XS
597 The following changes were made on top of the changes
598 made for Time::HiRes during the Perl 5.7 development
599 cycle that culminated in the release of Perl 5.8.0.
601 - add "require 5.005" to the Makefile.PL
602 - remove the REVISION section (CVS log) from HiRes.pm
603 - add jhi's copyright alongside Douglas'
604 - move HiRes.pm to lib/Time/
606 - modify HiRes.t to use $ENV{PERL_CORE}
607 - modify the original Time::HiRes version 1.20 Makefile.PL
608 to work both with Perl 5.8.0 and the new code with pre-5.8.0
609 Perls (tried with 5.6.1)
610 - tiny tweaks and updates in README and TODO
611 - bump the VERSION to 1.29
613 1.20 Wed Feb 24 21:30 1999
614 - make our usleep and ualarm substitutes into hrt_usleep
615 and hrt_ualarm. This helps static links of Perl with other
616 packages that also have usleep, etc. From
617 Ilya Zakharevich <ilya@math.ohio-state.edu>
618 - add C API stuff. From Joshua Pritikin
619 <joshua.pritikin@db.com>
620 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
621 - hopefully correct "-lc" fix for SCO.
624 1.20 adds a platform neutral set of C accessible routines if you are
625 running 5.005+. All other changes are packaging changes and build
626 fixes(?) for statically linked Perl, SCO, and VMS.
628 1.19 Tue Sep 29 22:30 1998
629 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
630 <seb@stian.demon.co.uk>
631 - change GIMME_V to GIMME to help people with older versions of
633 - fix Win32 version of gettimeofday(). It didn't affect anything,
634 but it confuses people reading the code when the return value
635 is backwards (0 is success).
636 - fix Makefile.PL (more) so that detection of gettimeofday is
639 1.19 has better VMS support.
641 1.18 Mon Jul 6 22:40 1998
642 - add usleep() for Win32.
643 - fix Makefile.PL to fix reported HP/UX feature where unresolved
644 externals still cause an executable to be generated (though no
645 x bit set). Thanks to David Kozinn for report and explanation.
646 Problems with the fix are mine :)
648 1.18 has limited Win32 support (no ualarm). Added usleep for Win32.
649 Probably buggy. I'm sure I'll hear.
651 1.17 Wed Jul 1 20:10 1998
652 - fix setitimer calls so microseconds is not more than 1000000.
653 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
654 - make Win32. We only get gettimeofday (the select hack doesn't
655 seem to work on my Win95 system).
656 - fix test 4 on 01test.t. add test to see if time() and
657 Time::HiRes::time() are close.
659 1.16 Wed Nov 12 21:05 1997
660 - add missing EXTEND in new gettimeofday scalar code.
662 1.16+ should be closer to building out of the box on Linux. Thanks
663 to Gisle Aas for patches, and the ualarm equivalent using setitimer.
665 If your underlying operating system doesn't implement ualarm(), then
666 a fake using setitimer() will be made. If the OS is missing usleep(),
667 a fake one using select() will be made. If a fake can't be made for
668 either ualarm() or usleep(), then the corresponding Perl function will
669 not be available. If the OS is missing gettimeofday(), you will get
670 unresolved externals, either at link- or run-time.
672 This is an improvement; the package used to not even build if
673 you were missing any of these bits. Roderick Schertler
675 <roderick@argon.org> did all the conditional compilation stuff,
676 look at HiRes.pm and the test suites; it's good educational reading.
678 1.15 Mon Nov 10 21:30 1997
679 - HiRes.pm: update pod. Provided by Gisle Aas.
680 - HiRes.xs: if gettimeofday() called in scalar context, do
681 something more useful than before. Provided by Gisle Aas.
682 - README: tell of xsubpp '-nolinenumber' woes. thanks to
683 Edward Henigin <ed@texas.net> for pointing out the problem.
685 1.14 Wed Nov 5 9:40 1997
686 - Makefile.PL: look for setitimer
687 - HiRes.xs: if missing ualarm, but we have setitimer, make up
688 our own setitimer. These were provided by Gisle Aas.
690 1.13 Tue Nov 4 23:30 1997
691 - Makefile.PL: fix autodetect mechanism to do try linking in addition
692 to just compiling; should fix Linux build problem. Fix was provided
695 1.12 Sun Oct 12 12:00:00 1997
696 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
697 you may need to comment this back out if you have an older xsubpp.
698 - HiRes.xs: set PROTOTYPES: DISABLE
700 1.11 Fri Sep 05 16:00:00 1997
702 Had some line commented out that shouldn't have been (testing
705 Previous version was corrupted.
707 1.10 Thu May 22 20:20:00 1997
708 - HiRes.xs, HiRes.pm, t/*:
709 - only compile what we have OS support for (or can
711 - only test what we compiled
712 - gross improvement to the test suite
714 This work was all done by Roderick Schertler
715 <roderick@argon.org>. If you run Linux or
716 one of the other ualarm-less platforms, and you like this
717 module, let Roderick know; without him, it still wouldn't
718 be working on those boxes...
719 - Makefile.PL: figure out what routines the OS has and
720 only build what we need. These bits were written by Jarkko
721 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
723 1.02 Mon Dec 30 08:00:00 1996
724 - HiRes.pm: update documentation to say what to do when missing
725 ualarm() and friends.
726 - README: update to warn that ualarm() and friends need to exist
728 1.01 Fri Oct 17 08:00:00 1996
729 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
730 - HiRes.pm: put blank line between __END__ and =head1 so that
733 1.00 Tue Sep 03 13:00:00 1996
734 - original version; created by h2xs 1.16