1 Revision history for Perl extension Time::HiRes.
4 - add support for the POSIX clock_gettime() and clock_getres(),
5 if available, either as library calls or as syscalls
6 - be more defensive about missing functionality: break out
7 early (during 'use') if no e.g. clock_getres() is available,
8 and protect our back by trapping those cases also in HiRes.xs
9 - the test added in 1.76 could cause an endless loop e.g. in Solaris,
10 due to mixing of sleep() and alarm() (bad programmer, no cookie!)
13 - testing for nanosleep had wrong logic which caused nanosleep
14 to become undefined for e.g. Mac OS X
15 - added a test for a core dump that was introduced by Perl 5.8.0
16 safe signals and was fixed for the time of 5.8.1 (one report of
17 the core dump was [perl #20920]), the test skipped pre-5.8.1.
18 - *cough* s/unanosleep/nanosleep/g; *cough*
21 - installation patch from Gisle Aas: in Perls 5.8.x and later
22 use MakeMaker INSTALLDIRS value of 'perl' instead of 'site'.
25 - [cpan #14608] Solaris 8 perl 5.005_03 File::Spec module does not have method rel2abs
26 (the workaround is not to use rel2abs, should not be necessary)
27 - [cpan #14642] U2time wrongly exported on the C API
28 (patch supplied by the reporter, SALVA@cpan.org)
29 - add release dates to Changes
32 - Time::HiRes::nanosleep support for Solaris [PATCH]
33 (POSIX::uname() not available if building with core perl,
34 from Gisle Aas, via perl5-porters, perl change #25295)
37 - going back to the 1.68 loader setup (using DynaLoader)
38 since too many weird things starting breaking
39 - fix a typo in José Auguste-Etienne's name
42 - a thinko in the nanosleep() detection
43 - move more changes stuff from the README to Changes
44 - add -w to the Makefile.PL
47 - oops in 1.69 about @ISA (not affecting anything but silly)
48 - add copyright 2005 to HiRes.pm
49 - add copyright and license to HiRes.xs
50 - add copyrights 2003, 2004, 2005 to README
53 - actually run a test for nanosleep
54 (if there is no $Config{d_nanosleep}) since e.g. in AIX 4.2
55 it seems that one can link in nanosleep() but then calling
56 it fails instantly and sets errno to ENOSYS (Not implemented).
57 This may be fixable in the AIX case by figuring out the right
58 (realtime POSIX?) libs and whatnot, but in the general case
59 running a real test case is better. (Of course, this change
60 will no doubt run into portability problems because of the
61 execution step...) Note that because of hysterical raisins
62 most Perls do NOT have $Config{d_nanosleep} (scanning for
63 it by Configure would in many platforms require linking in
64 things like -lrt, which would in many platforms be a bad idea
66 (from José Auguste-Etienne)
67 - support XSLoader also since it's much faster
69 - add SEE ALSO (BSD::Resource and Time::TAI64)
72 - somehow 1.67 had a lot of doubled lines (a major cut-and-paste
73 error suspected), but miraculously it still worked since the
74 doubling took place below the __END__ token
75 - undef Pause() before defining it to avoid redefinition warnings
76 during compilation in case perl.h had already defined Pause()
77 (part of perl change #24271)
81 - (internal) don't ignore the return value of gettimeofday()
82 - (external) return undef or an empty if the C gettimeofday() fails
83 (affects Time::HiRes gettimeofday() and the hires time())
87 - fix the 'hierachy' typo in Makefile.PL [rt.cpan.org #8492]
88 - should now build in Solaris [rt.cpan.org #7165] (since 1.64)
89 - should now build in Cygwin [rt.cpan.org #7535] (since 1.64)
90 - close also [rt.cpan.org #5933] "Time::HiRes::time does not pick up time adjustments like ntp" since ever reproducing it in the same environment
91 has become rather unlikely
94 - one should not mix u?alarm and sleep (the tests modified
95 by 1.65, #12 and #13, hung in Solaris), now we just busy
96 loop executing an empty block
97 - in the documentation underline the unspecificity of mixing
99 - small spelling fixes
102 - regenerate ppport.h with Devel::PPPort 3.03,
103 now the MY_CXT_CLONE is defined in ppport.h,
104 we no more need to do that.
106 - the test #12 would often hang in sigsuspend() (at least that's
107 where Mac OS X' ktrace shows it hanging). With the sleep()s
108 changed to sleep(1)s, the tests still pass but no hang after
109 a few hundred repeats.
112 - Win32 and any ithread build: ppport.h didn't define
113 MY_CXT_CLONE, which seems to be a Time-HiResism.
116 - Skip testing if under PERL_CORE and Time::HiRes has not
117 been Configured (from Marcus Holland-Moritz, core change
119 - Use ppport.h generated by Devel::PPPort 3.01,
120 allowing cutting away our own portability code.
121 - Don't use $ENV{PERL_CORE} for < 5.6.0.
122 - Don't use "for my $i" for <= 5.003.
123 - Don't use Pause() for <= 5.003.
124 - Can't use newSVpvf for <= 5.003.
125 (most of the changes from Marcus)
128 - Win32: reset reading from the performance counters every
129 five minutes to better track wall clock time (thanks to
130 PC timers being often quite bad), should help long-running
134 - Win32: Patch from Steve Hay
135 [PATCH] Re: [perl #30755] [Win32] Different results from Time::HiRes::gettimeofdayunder the debugger
136 to [perl #30755] reported by Nigel Sandever
138 - Cygwin: Use the Win32 recalibration code also in Cygwin if the
139 <w32api/windows.h> APIs are available. Cygwin testing by
140 Yitzchak Scott-Thoennes.
142 - Solaris: use -lposix4 to get nanosleep for Solaris 2.6,
143 after that keep using -lrt, patch from Alan Burlison,
144 bug reported in [cpan #7165]
147 - Change the Win32 recalibration limit to 0.5 seconds and tweak
148 the documentation to blather less about the gory details of the
149 Win32 implementation and more about the complications in general
150 of meddling with the system clock.
153 - Document the 1.57 change better.
156 - Win32/Cygwin/MinGW: if the performance counter drifts by more
157 than two seconds from the system clock (due to ntp adjustments,
158 for example), recalibrate our internal counter: from Jan Dubois,
159 based on [cpan #5933] by Jerry D. Hedden.
162 - Give a clearer message if the tests timeout (perl change #22253)
163 - Don't use /tmp or its moral equivalents (perl bug #15036,
167 - Windows: mingw32 patch from Mike Pomraning (use Perl's Const64()
168 instead of VC-specific i64 suffix)
171 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
174 - Windows: higher resolution time() by using the Windows
175 performance counter API, from Jan Dubois and Anton Shcherbinin.
176 The exact new higher resolution depends on the hardware,
177 but it should be quite a bit better than using the basic
181 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
182 the subtest 18. No known analysis nor fix, but added
183 an alarm (that requires fork() and alarm()) to the test.
186 - doc tweaks from mjd (perl change #20456)
187 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
190 - add a message (for non-core builds) to Makefile.PL about
191 the LC_ALL=C workaround
192 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
193 - the nanosleep() probe was broken (perl change #20061)
194 - use existence instead of definedness for feature probes
196 - MPE/iX tweak (perl change #20042)
197 - do not use HAS_NANOSLEEP (perl change #19898)
200 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
201 - OS/2 can always mix subsecond sleeps with signals
202 (part of perl change #19789)
205 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
209 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
210 - unnecessary (nanosleep is in libc anyway)
211 - harmful (-lrt slows down execution)
212 - incompatible (with many distributions' pthreads)
215 - do not create files in blib directories under core
216 (perl change #19160, from rgs)
217 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
220 - guarantee that $xdefine in HiRes.t is always defined
221 (perl change #19109, from IlyaZ)
222 - a cleaner way to detect PERL_CORE (perl change #19111,
226 - add hints/irix.pl to turn off overly POSIX flags that
227 cause hide struct timespec to be hidden (and compilation
228 to fail) (bleadperl change #19085)
229 - documentation tweaks
232 - add c:/temp to the list of temp directories to probe
233 so that cygwin (and win*?) builds are happy. This was
234 needed at least in my cygwin 1.3.20/w2k setup.
237 - modernize the constants code (from Nicholas Clark)
240 - At some point the ability to figure our the correct incdir
241 for EXTERN.h (either a core perl build, or an installed perl)
242 had broken (which lead into all test compiles failing with
243 a core perl build, but thanks to the robustness of Makefile.PL
244 nothing of this was visible). The brokenness seemed to be
245 caused by $ENV{PERL_CORE} not being on for core builds?
246 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
247 right, and both styles of build should work again.
250 - Nicholas Clark noticed that the my_catdir() emulation function
251 was broken (which means that we didn't really work for Perls
253 - inspired by fixing the above made the whole Makefile.PL -w
255 - tightened up the Makefile.PL output, less whitespace
258 - fix from Craig Berry for better building in VMS with PERL_CORE
261 - no functional changes
262 - move lib/Time/HiRes.pm as Hires.pm
263 - libraries scanning was slightly broken (always scanned
264 for a library even when $Config{libs} already had it)
267 - Ray Zimmerman ran into a race condition in Mac OS X.
268 A 0.01-second alarm fired before the test expected.
269 The test first slept indefinitely (blocking for signals)
270 and only after that tested for the signal having been sent.
271 Since the signal had already been sent, the test #12 never
272 completed. The solution: test first, then block.
273 - default to being silent on all probing attempts, set the
274 environment variable VERBOSE to a true value to see the
275 details (the probing command and the possible errors)
278 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
279 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
280 cannot be upgraded from CPAN to override the 5.8.0 version
282 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
283 because the include directories did not adjust themselves
284 if $ENV{PERL_CORE} (Hugo van der Sanden)
285 - add documentation about the restart of select() under alarm()
288 - small documentation tweaks
292 - better VMS operation (Craig Berry)
295 - our time machine is accelerating: now works with Perl 5.004_01
296 (tried with 5.003_07 and 5.002 but I get segmentation faults
297 from running the Makefile.PL with those in Tru64 4.0D)
300 - backward compatibility (pre-5.6.0) tweaks:
301 - no XSLoader in 5.00503, use DynaLoader instead
302 - no SvPV_nolen, either
303 - no PerlProc_pause(), either
304 - now tested with 5.00404 and 5.00503
305 - Makefile.PL requires 5.00404 (no more 5.002)
306 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
307 (this means that one can mix subsecond sleeps with alarms)
308 - because of nanosleep we probe for -lrt and -lposix4
309 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
310 is available by exportable constants Time::HiRes::d_func
311 (since older Perl do not have them in %Config, and even
312 5.8.0 does not probe for nanosleep)
315 - backward compatibility (pre-5.6.1) tweaks:
316 - define NV if no NVTYPE
317 - define IVdf if needed (note: the Devel::PPPort
318 in 5.8.0 does not try hard hard enough since
319 the IVSIZE might not be defined)
320 - define NVgf if needed
321 - grab the typemap from 5.8.0 for the NV stuff
323 1.31 and 1.32 add more backward compatibility (now all the way
324 back to Perl 5.00404), and using nanosleep() (if available) for
329 - release 1.29_02 as 1.30
331 1.30 adds all the changes made during the Perl 5.6->5.7->5.8
332 development cycle. Most notably portability across platforms has been
333 enhanced, and the interval timers (setitimer, getitimer) have been
334 added. Note that the version of Time::HiRes that is included in Perl
335 5.8.0 calls itself 1.20_00, but it is equivalent to this Time::HiRes
336 version. Note also that in 1.30 Wegscheid turns over the maintenance
337 to Jarkko Hietaniemi.
341 - fix a silly unclosed comment typo in HiRes.xs
342 - document and export REALTIME_REALPROF (Solaris)
346 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
347 (need to patch this also in Perl 5.[89])
348 - remove CVS revision log from HiRes.xs
352 The following numbered patches refer to the Perl 5.7 changes,
353 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
355 - 17558: Add #!./perl to the .t
356 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
357 - 16198: political correctness, from Simon Cozens
358 - 15857: doc tweaks, from Jarkko Hietaniemi
359 - 15593: optimization in .xs, from Paul Green
360 - 14892: pod fixes, from Robin Barker
361 - 14100: VOS fixes, from Paul Green
362 - 13422: XS segfault, from Marc Lehmann
363 - 13378: whether select() gets restarted on signals, depends
364 - 13354: timing constraints, again, from Andy Dougherty
365 - 13278: can't do subecond alarms with ualarm;
366 break out early if alarms do not seem to be working
367 - 13266: test relaxation (cygwin gets lower hires
368 times than lores ones)
369 - 12846: protect against high load, from Jarkko Hietaniemi
370 - 12837: HiRes.t VMS tweak, from Craig A. Berry
371 - 12797: HiRes.t VMS tweak, from Charles Lane
372 - 12769: HiRes.t VMS tweak, from Craig A. Berry
373 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
374 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
375 - 12692: alarm() ain't gonna work if ualarm() ain't,
376 from Gurusamy Sarathy
377 - 12680: minor VMS tweak, from Charles Lane
378 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
379 - 12609: croak on negative time, from Jarkko Hietaniemi
380 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
381 - 12594: MacOS Classic timeofday, from Chris Nandor
382 - 12473: allow for more than one second for sleep() and usleep()
383 - 12458: test tuning, relax timing constraints,
384 from Jarkko Hietaniemi
385 - 12449: make sleep() and usleep() to return the number
386 of seconds and microseconds actually slept (analogously
387 with the builtin sleep()), also make usleep() croak if
388 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
389 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
390 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
391 - 12196: use ftime() on Win32, from Artur Bergman
392 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
393 - 12105: use GetSystemTime() on Win32, from Artur Bergman
394 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
395 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
396 - 11797: problem in HiRes.t, from John P. Linderman
397 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
398 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
399 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
400 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
401 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
402 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
404 - 10942: MPE/IX test tweaks, from Mark Bixby
405 - 10784: unnecessary pod2man calls, from Andy Dougherty
406 - 10354: ext/ + -Wall, from Doug MacEachern
407 - 10320: fix the BOOT section to call myU2time correctly
408 - 10317: correct casting for AIX< from H. Merijn Brand
409 - 10119: document that the core time() may be rounding, not truncating
410 - 10118: test fix, from John Peacock
411 - 9988: long =item, from Robin Barker
412 - 9714: correct test output
413 - 9708: test also the scalar aspect of getitimer()
414 - 9705: Add interval timers (setitimer, getitimer)
415 - 9692: do not require at least 5.005 using XS
417 The following changes were made on top of the changes
418 made for Time::HiRes during the Perl 5.7 development
419 cycle that culminated in the release of Perl 5.8.0.
421 - add "require 5.005" to the Makefile.PL
422 - remove the REVISION section (CVS log) from HiRes.pm
423 - add jhi's copyright alongside Douglas'
424 - move HiRes.pm to lib/Time/
426 - modify HiRes.t to use $ENV{PERL_CORE}
427 - modify the original Time::HiRes version 1.20 Makefile.PL
428 to work both with Perl 5.8.0 and the new code with pre-5.8.0
429 Perls (tried with 5.6.1)
430 - tiny tweaks and updates in README and TODO
431 - bump the VERSION to 1.29
433 1.20 Wed Feb 24 21:30 1999
434 - make our usleep and ualarm substitutes into hrt_usleep
435 and hrt_ualarm. This helps static links of Perl with other
436 packages that also have usleep, etc. From
437 Ilya Zakharevich <ilya@math.ohio-state.edu>
438 - add C API stuff. From Joshua Pritikin
439 <joshua.pritikin@db.com>
440 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
441 - hopefully correct "-lc" fix for SCO.
444 1.20 adds a platform neutral set of C accessible routines if you are
445 running 5.005+. All other changes are packaging changes and build
446 fixes(?) for statically linked Perl, SCO, and VMS.
448 1.19 Tue Sep 29 22:30 1998
449 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
450 <seb@stian.demon.co.uk>
451 - change GIMME_V to GIMME to help people with older versions of
453 - fix Win32 version of gettimeofday(). It didn't affect anything,
454 but it confuses people reading the code when the return value
455 is backwards (0 is success).
456 - fix Makefile.PL (more) so that detection of gettimeofday is
459 1.19 has better VMS support.
461 1.18 Mon Jul 6 22:40 1998
462 - add usleep() for Win32.
463 - fix Makefile.PL to fix reported HP/UX feature where unresolved
464 externals still cause an executable to be generated (though no
465 x bit set). Thanks to David Kozinn for report and explanation.
466 Problems with the fix are mine :)
468 1.18 has limited Win32 support (no ualarm). Added usleep for Win32.
469 Probably buggy. I'm sure I'll hear.
471 1.17 Wed Jul 1 20:10 1998
472 - fix setitimer calls so microseconds is not more than 1000000.
473 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
474 - make Win32. We only get gettimeofday (the select hack doesn't
475 seem to work on my Win95 system).
476 - fix test 4 on 01test.t. add test to see if time() and
477 Time::HiRes::time() are close.
479 1.16 Wed Nov 12 21:05 1997
480 - add missing EXTEND in new gettimeofday scalar code.
482 1.16+ should be closer to building out of the box on Linux. Thanks
483 to Gisle Aas for patches, and the ualarm equivalent using setitimer.
485 If your underlying operating system doesn't implement ualarm(), then
486 a fake using setitimer() will be made. If the OS is missing usleep(),
487 a fake one using select() will be made. If a fake can't be made for
488 either ualarm() or usleep(), then the corresponding Perl function will
489 not be available. If the OS is missing gettimeofday(), you will get
490 unresolved externals, either at link- or run-time.
492 This is an improvement; the package used to not even build if
493 you were missing any of these bits. Roderick Schertler
495 <roderick@argon.org> did all the conditional compilation stuff,
496 look at HiRes.pm and the test suites; it's good educational reading.
498 1.15 Mon Nov 10 21:30 1997
499 - HiRes.pm: update pod. Provided by Gisle Aas.
500 - HiRes.xs: if gettimeofday() called in scalar context, do
501 something more useful than before. Provided by Gisle Aas.
502 - README: tell of xsubpp '-nolinenumber' woes. thanks to
503 Edward Henigin <ed@texas.net> for pointing out the problem.
505 1.14 Wed Nov 5 9:40 1997
506 - Makefile.PL: look for setitimer
507 - HiRes.xs: if missing ualarm, but we have setitimer, make up
508 our own setitimer. These were provided by Gisle Aas.
510 1.13 Tue Nov 4 23:30 1997
511 - Makefile.PL: fix autodetect mechanism to do try linking in addition
512 to just compiling; should fix Linux build problem. Fix was provided
515 1.12 Sun Oct 12 12:00:00 1997
516 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
517 you may need to comment this back out if you have an older xsubpp.
518 - HiRes.xs: set PROTOTYPES: DISABLE
520 1.11 Fri Sep 05 16:00:00 1997
522 Had some line commented out that shouldn't have been (testing
525 Previous version was corrupted.
527 1.10 Thu May 22 20:20:00 1997
528 - HiRes.xs, HiRes.pm, t/*:
529 - only compile what we have OS support for (or can
531 - only test what we compiled
532 - gross improvement to the test suite
534 This work was all done by Roderick Schertler
535 <roderick@argon.org>. If you run Linux or
536 one of the other ualarm-less platoforms, and you like this
537 module, let Roderick know; without him, it still wouldn't
538 be working on those boxes...
539 - Makefile.PL: figure out what routines the OS has and
540 only build what we need. These bits were written by Jarkko
541 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
543 1.02 Mon Dec 30 08:00:00 1996
544 - HiRes.pm: update documentation to say what to do when missing
545 ualarm() and friends.
546 - README: update to warn that ualarm() and friends need to exist
548 1.01 Fri Oct 17 08:00:00 1996
549 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
550 - HiRes.pm: put blank line between __END__ and =head1 so that
553 1.00 Tue Sep 03 13:00:00 1996
554 - original version; created by h2xs 1.16