1 Revision history for Perl extension Time::HiRes.
4 - testing for nanosleep had wrong logic which caused nanosleep
5 to become undefined for e.g. Mac OS X
6 - added a test for a core dump that was introduced by Perl 5.8.0
7 safe signals and was fixed for the time of 5.8.1 (one report of
8 the core dump was [perl #20920]), the test skipped pre-5.8.1.
9 - *cough* s/unanosleep/nanosleep/g; *cough*
12 - installation patch from Gisle Aas: in Perls 5.8.x and later
13 use MakeMaker INSTALLDIRS value of 'perl' instead of 'site'.
16 - [cpan #14608] Solaris 8 perl 5.005_03 File::Spec module does not have method rel2abs
17 (the workaround is not to use rel2abs, should not be necessary)
18 - [cpan #14642] U2time wrongly exported on the C API
19 (patch supplied by the reporter, SALVA@cpan.org)
20 - add release dates to Changes
23 - Time::HiRes::nanosleep support for Solaris [PATCH]
24 (POSIX::uname() not available if building with core perl,
25 from Gisle Aas, via perl5-porters, perl change #25295)
28 - going back to the 1.68 loader setup (using DynaLoader)
29 since too many weird things starting breaking
30 - fix a typo in José Auguste-Etienne's name
33 - a thinko in the nanosleep() detection
34 - move more changes stuff from the README to Changes
35 - add -w to the Makefile.PL
38 - oops in 1.69 about @ISA (not affecting anything but silly)
39 - add copyright 2005 to HiRes.pm
40 - add copyright and license to HiRes.xs
41 - add copyrights 2003, 2004, 2005 to README
44 - actually run a test for nanosleep
45 (if there is no $Config{d_nanosleep}) since e.g. in AIX 4.2
46 it seems that one can link in nanosleep() but then calling
47 it fails instantly and sets errno to ENOSYS (Not implemented).
48 This may be fixable in the AIX case by figuring out the right
49 (realtime POSIX?) libs and whatnot, but in the general case
50 running a real test case is better. (Of course, this change
51 will no doubt run into portability problems because of the
52 execution step...) Note that because of hysterical raisins
53 most Perls do NOT have $Config{d_nanosleep} (scanning for
54 it by Configure would in many platforms require linking in
55 things like -lrt, which would in many platforms be a bad idea
57 (from José Auguste-Etienne)
58 - support XSLoader also since it's much faster
60 - add SEE ALSO (BSD::Resource and Time::TAI64)
63 - somehow 1.67 had a lot of doubled lines (a major cut-and-paste
64 error suspected), but miraculously it still worked since the
65 doubling took place below the __END__ token
66 - undef Pause() before defining it to avoid redefinition warnings
67 during compilation in case perl.h had already defined Pause()
68 (part of perl change #24271)
72 - (internal) don't ignore the return value of gettimeofday()
73 - (external) return undef or an empty if the C gettimeofday() fails
74 (affects Time::HiRes gettimeofday() and the hires time())
78 - fix the 'hierachy' typo in Makefile.PL [rt.cpan.org #8492]
79 - should now build in Solaris [rt.cpan.org #7165] (since 1.64)
80 - should now build in Cygwin [rt.cpan.org #7535] (since 1.64)
81 - 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
82 has become rather unlikely
85 - one should not mix u?alarm and sleep (the tests modified
86 by 1.65, #12 and #13, hung in Solaris), now we just busy
87 loop executing an empty block
88 - in the documentation underline the unspecificity of mixing
90 - small spelling fixes
93 - regenerate ppport.h with Devel::PPPort 3.03,
94 now the MY_CXT_CLONE is defined in ppport.h,
95 we no more need to do that.
97 - the test #12 would often hang in sigsuspend() (at least that's
98 where Mac OS X' ktrace shows it hanging). With the sleep()s
99 changed to sleep(1)s, the tests still pass but no hang after
100 a few hundred repeats.
103 - Win32 and any ithread build: ppport.h didn't define
104 MY_CXT_CLONE, which seems to be a Time-HiResism.
107 - Skip testing if under PERL_CORE and Time::HiRes has not
108 been Configured (from Marcus Holland-Moritz, core change
110 - Use ppport.h generated by Devel::PPPort 3.01,
111 allowing cutting away our own portability code.
112 - Don't use $ENV{PERL_CORE} for < 5.6.0.
113 - Don't use "for my $i" for <= 5.003.
114 - Don't use Pause() for <= 5.003.
115 - Can't use newSVpvf for <= 5.003.
116 (most of the changes from Marcus)
119 - Win32: reset reading from the performance counters every
120 five minutes to better track wall clock time (thanks to
121 PC timers being often quite bad), should help long-running
125 - Win32: Patch from Steve Hay
126 [PATCH] Re: [perl #30755] [Win32] Different results from Time::HiRes::gettimeofdayunder the debugger
127 to [perl #30755] reported by Nigel Sandever
129 - Cygwin: Use the Win32 recalibration code also in Cygwin if the
130 <w32api/windows.h> APIs are available. Cygwin testing by
131 Yitzchak Scott-Thoennes.
133 - Solaris: use -lposix4 to get nanosleep for Solaris 2.6,
134 after that keep using -lrt, patch from Alan Burlison,
135 bug reported in [cpan #7165]
138 - Change the Win32 recalibration limit to 0.5 seconds and tweak
139 the documentation to blather less about the gory details of the
140 Win32 implementation and more about the complications in general
141 of meddling with the system clock.
144 - Document the 1.57 change better.
147 - Win32/Cygwin/MinGW: if the performance counter drifts by more
148 than two seconds from the system clock (due to ntp adjustments,
149 for example), recalibrate our internal counter: from Jan Dubois,
150 based on [cpan #5933] by Jerry D. Hedden.
153 - Give a clearer message if the tests timeout (perl change #22253)
154 - Don't use /tmp or its moral equivalents (perl bug #15036,
158 - Windows: mingw32 patch from Mike Pomraning (use Perl's Const64()
159 instead of VC-specific i64 suffix)
162 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
165 - Windows: higher resolution time() by using the Windows
166 performance counter API, from Jan Dubois and Anton Shcherbinin.
167 The exact new higher resolution depends on the hardware,
168 but it should be quite a bit better than using the basic
172 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
173 the subtest 18. No known analysis nor fix, but added
174 an alarm (that requires fork() and alarm()) to the test.
177 - doc tweaks from mjd (perl change #20456)
178 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
181 - add a message (for non-core builds) to Makefile.PL about
182 the LC_ALL=C workaround
183 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
184 - the nanosleep() probe was broken (perl change #20061)
185 - use existence instead of definedness for feature probes
187 - MPE/iX tweak (perl change #20042)
188 - do not use HAS_NANOSLEEP (perl change #19898)
191 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
192 - OS/2 can always mix subsecond sleeps with signals
193 (part of perl change #19789)
196 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
200 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
201 - unnecessary (nanosleep is in libc anyway)
202 - harmful (-lrt slows down execution)
203 - incompatible (with many distributions' pthreads)
206 - do not create files in blib directories under core
207 (perl change #19160, from rgs)
208 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
211 - guarantee that $xdefine in HiRes.t is always defined
212 (perl change #19109, from IlyaZ)
213 - a cleaner way to detect PERL_CORE (perl change #19111,
217 - add hints/irix.pl to turn off overly POSIX flags that
218 cause hide struct timespec to be hidden (and compilation
219 to fail) (bleadperl change #19085)
220 - documentation tweaks
223 - add c:/temp to the list of temp directories to probe
224 so that cygwin (and win*?) builds are happy. This was
225 needed at least in my cygwin 1.3.20/w2k setup.
228 - modernize the constants code (from Nicholas Clark)
231 - At some point the ability to figure our the correct incdir
232 for EXTERN.h (either a core perl build, or an installed perl)
233 had broken (which lead into all test compiles failing with
234 a core perl build, but thanks to the robustness of Makefile.PL
235 nothing of this was visible). The brokenness seemed to be
236 caused by $ENV{PERL_CORE} not being on for core builds?
237 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
238 right, and both styles of build should work again.
241 - Nicholas Clark noticed that the my_catdir() emulation function
242 was broken (which means that we didn't really work for Perls
244 - inspired by fixing the above made the whole Makefile.PL -w
246 - tightened up the Makefile.PL output, less whitespace
249 - fix from Craig Berry for better building in VMS with PERL_CORE
252 - no functional changes
253 - move lib/Time/HiRes.pm as Hires.pm
254 - libraries scanning was slightly broken (always scanned
255 for a library even when $Config{libs} already had it)
258 - Ray Zimmerman ran into a race condition in Mac OS X.
259 A 0.01-second alarm fired before the test expected.
260 The test first slept indefinitely (blocking for signals)
261 and only after that tested for the signal having been sent.
262 Since the signal had already been sent, the test #12 never
263 completed. The solution: test first, then block.
264 - default to being silent on all probing attempts, set the
265 environment variable VERBOSE to a true value to see the
266 details (the probing command and the possible errors)
269 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
270 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
271 cannot be upgraded from CPAN to override the 5.8.0 version
273 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
274 because the include directories did not adjust themselves
275 if $ENV{PERL_CORE} (Hugo van der Sanden)
276 - add documentation about the restart of select() under alarm()
279 - small documentation tweaks
283 - better VMS operation (Craig Berry)
286 - our time machine is accelerating: now works with Perl 5.004_01
287 (tried with 5.003_07 and 5.002 but I get segmentation faults
288 from running the Makefile.PL with those in Tru64 4.0D)
291 - backward compatibility (pre-5.6.0) tweaks:
292 - no XSLoader in 5.00503, use DynaLoader instead
293 - no SvPV_nolen, either
294 - no PerlProc_pause(), either
295 - now tested with 5.00404 and 5.00503
296 - Makefile.PL requires 5.00404 (no more 5.002)
297 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
298 (this means that one can mix subsecond sleeps with alarms)
299 - because of nanosleep we probe for -lrt and -lposix4
300 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
301 is available by exportable constants Time::HiRes::d_func
302 (since older Perl do not have them in %Config, and even
303 5.8.0 does not probe for nanosleep)
306 - backward compatibility (pre-5.6.1) tweaks:
307 - define NV if no NVTYPE
308 - define IVdf if needed (note: the Devel::PPPort
309 in 5.8.0 does not try hard hard enough since
310 the IVSIZE might not be defined)
311 - define NVgf if needed
312 - grab the typemap from 5.8.0 for the NV stuff
314 1.31 and 1.32 add more backward compatibility (now all the way
315 back to Perl 5.00404), and using nanosleep() (if available) for
320 - release 1.29_02 as 1.30
322 1.30 adds all the changes made during the Perl 5.6->5.7->5.8
323 development cycle. Most notably portability across platforms has been
324 enhanced, and the interval timers (setitimer, getitimer) have been
325 added. Note that the version of Time::HiRes that is included in Perl
326 5.8.0 calls itself 1.20_00, but it is equivalent to this Time::HiRes
327 version. Note also that in 1.30 Wegscheid turns over the maintenance
328 to Jarkko Hietaniemi.
332 - fix a silly unclosed comment typo in HiRes.xs
333 - document and export REALTIME_REALPROF (Solaris)
337 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
338 (need to patch this also in Perl 5.[89])
339 - remove CVS revision log from HiRes.xs
343 The following numbered patches refer to the Perl 5.7 changes,
344 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
346 - 17558: Add #!./perl to the .t
347 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
348 - 16198: political correctness, from Simon Cozens
349 - 15857: doc tweaks, from Jarkko Hietaniemi
350 - 15593: optimization in .xs, from Paul Green
351 - 14892: pod fixes, from Robin Barker
352 - 14100: VOS fixes, from Paul Green
353 - 13422: XS segfault, from Marc Lehmann
354 - 13378: whether select() gets restarted on signals, depends
355 - 13354: timing constraints, again, from Andy Dougherty
356 - 13278: can't do subecond alarms with ualarm;
357 break out early if alarms do not seem to be working
358 - 13266: test relaxation (cygwin gets lower hires
359 times than lores ones)
360 - 12846: protect against high load, from Jarkko Hietaniemi
361 - 12837: HiRes.t VMS tweak, from Craig A. Berry
362 - 12797: HiRes.t VMS tweak, from Charles Lane
363 - 12769: HiRes.t VMS tweak, from Craig A. Berry
364 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
365 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
366 - 12692: alarm() ain't gonna work if ualarm() ain't,
367 from Gurusamy Sarathy
368 - 12680: minor VMS tweak, from Charles Lane
369 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
370 - 12609: croak on negative time, from Jarkko Hietaniemi
371 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
372 - 12594: MacOS Classic timeofday, from Chris Nandor
373 - 12473: allow for more than one second for sleep() and usleep()
374 - 12458: test tuning, relax timing constraints,
375 from Jarkko Hietaniemi
376 - 12449: make sleep() and usleep() to return the number
377 of seconds and microseconds actually slept (analogously
378 with the builtin sleep()), also make usleep() croak if
379 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
380 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
381 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
382 - 12196: use ftime() on Win32, from Artur Bergman
383 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
384 - 12105: use GetSystemTime() on Win32, from Artur Bergman
385 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
386 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
387 - 11797: problem in HiRes.t, from John P. Linderman
388 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
389 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
390 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
391 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
392 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
393 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
395 - 10942: MPE/IX test tweaks, from Mark Bixby
396 - 10784: unnecessary pod2man calls, from Andy Dougherty
397 - 10354: ext/ + -Wall, from Doug MacEachern
398 - 10320: fix the BOOT section to call myU2time correctly
399 - 10317: correct casting for AIX< from H. Merijn Brand
400 - 10119: document that the core time() may be rounding, not truncating
401 - 10118: test fix, from John Peacock
402 - 9988: long =item, from Robin Barker
403 - 9714: correct test output
404 - 9708: test also the scalar aspect of getitimer()
405 - 9705: Add interval timers (setitimer, getitimer)
406 - 9692: do not require at least 5.005 using XS
408 The following changes were made on top of the changes
409 made for Time::HiRes during the Perl 5.7 development
410 cycle that culminated in the release of Perl 5.8.0.
412 - add "require 5.005" to the Makefile.PL
413 - remove the REVISION section (CVS log) from HiRes.pm
414 - add jhi's copyright alongside Douglas'
415 - move HiRes.pm to lib/Time/
417 - modify HiRes.t to use $ENV{PERL_CORE}
418 - modify the original Time::HiRes version 1.20 Makefile.PL
419 to work both with Perl 5.8.0 and the new code with pre-5.8.0
420 Perls (tried with 5.6.1)
421 - tiny tweaks and updates in README and TODO
422 - bump the VERSION to 1.29
424 1.20 Wed Feb 24 21:30 1999
425 - make our usleep and ualarm substitutes into hrt_usleep
426 and hrt_ualarm. This helps static links of Perl with other
427 packages that also have usleep, etc. From
428 Ilya Zakharevich <ilya@math.ohio-state.edu>
429 - add C API stuff. From Joshua Pritikin
430 <joshua.pritikin@db.com>
431 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
432 - hopefully correct "-lc" fix for SCO.
435 1.20 adds a platform neutral set of C accessible routines if you are
436 running 5.005+. All other changes are packaging changes and build
437 fixes(?) for statically linked Perl, SCO, and VMS.
439 1.19 Tue Sep 29 22:30 1998
440 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
441 <seb@stian.demon.co.uk>
442 - change GIMME_V to GIMME to help people with older versions of
444 - fix Win32 version of gettimeofday(). It didn't affect anything,
445 but it confuses people reading the code when the return value
446 is backwards (0 is success).
447 - fix Makefile.PL (more) so that detection of gettimeofday is
450 1.19 has better VMS support.
452 1.18 Mon Jul 6 22:40 1998
453 - add usleep() for Win32.
454 - fix Makefile.PL to fix reported HP/UX feature where unresolved
455 externals still cause an executable to be generated (though no
456 x bit set). Thanks to David Kozinn for report and explanation.
457 Problems with the fix are mine :)
459 1.18 has limited Win32 support (no ualarm). Added usleep for Win32.
460 Probably buggy. I'm sure I'll hear.
462 1.17 Wed Jul 1 20:10 1998
463 - fix setitimer calls so microseconds is not more than 1000000.
464 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
465 - make Win32. We only get gettimeofday (the select hack doesn't
466 seem to work on my Win95 system).
467 - fix test 4 on 01test.t. add test to see if time() and
468 Time::HiRes::time() are close.
470 1.16 Wed Nov 12 21:05 1997
471 - add missing EXTEND in new gettimeofday scalar code.
473 1.16+ should be closer to building out of the box on Linux. Thanks
474 to Gisle Aas for patches, and the ualarm equivalent using setitimer.
476 If your underlying operating system doesn't implement ualarm(), then
477 a fake using setitimer() will be made. If the OS is missing usleep(),
478 a fake one using select() will be made. If a fake can't be made for
479 either ualarm() or usleep(), then the corresponding Perl function will
480 not be available. If the OS is missing gettimeofday(), you will get
481 unresolved externals, either at link- or run-time.
483 This is an improvement; the package used to not even build if
484 you were missing any of these bits. Roderick Schertler
486 <roderick@argon.org> did all the conditional compilation stuff,
487 look at HiRes.pm and the test suites; it's good educational reading.
489 1.15 Mon Nov 10 21:30 1997
490 - HiRes.pm: update pod. Provided by Gisle Aas.
491 - HiRes.xs: if gettimeofday() called in scalar context, do
492 something more useful than before. Provided by Gisle Aas.
493 - README: tell of xsubpp '-nolinenumber' woes. thanks to
494 Edward Henigin <ed@texas.net> for pointing out the problem.
496 1.14 Wed Nov 5 9:40 1997
497 - Makefile.PL: look for setitimer
498 - HiRes.xs: if missing ualarm, but we have setitimer, make up
499 our own setitimer. These were provided by Gisle Aas.
501 1.13 Tue Nov 4 23:30 1997
502 - Makefile.PL: fix autodetect mechanism to do try linking in addition
503 to just compiling; should fix Linux build problem. Fix was provided
506 1.12 Sun Oct 12 12:00:00 1997
507 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
508 you may need to comment this back out if you have an older xsubpp.
509 - HiRes.xs: set PROTOTYPES: DISABLE
511 1.11 Fri Sep 05 16:00:00 1997
513 Had some line commented out that shouldn't have been (testing
516 Previous version was corrupted.
518 1.10 Thu May 22 20:20:00 1997
519 - HiRes.xs, HiRes.pm, t/*:
520 - only compile what we have OS support for (or can
522 - only test what we compiled
523 - gross improvement to the test suite
525 This work was all done by Roderick Schertler
526 <roderick@argon.org>. If you run Linux or
527 one of the other ualarm-less platoforms, and you like this
528 module, let Roderick know; without him, it still wouldn't
529 be working on those boxes...
530 - Makefile.PL: figure out what routines the OS has and
531 only build what we need. These bits were written by Jarkko
532 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
534 1.02 Mon Dec 30 08:00:00 1996
535 - HiRes.pm: update documentation to say what to do when missing
536 ualarm() and friends.
537 - README: update to warn that ualarm() and friends need to exist
539 1.01 Fri Oct 17 08:00:00 1996
540 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
541 - HiRes.pm: put blank line between __END__ and =head1 so that
544 1.00 Tue Sep 03 13:00:00 1996
545 - original version; created by h2xs 1.16