This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Upgrade to Time-HiRes-1.9703
[perl5.git] / ext / Time / HiRes / Changes
CommitLineData
75d5269b 1Revision history for the Perl extension Time::HiRes.
dcf686c9 2
f3370602
SP
31.9703 [2006-12-08]
4 - use int main(int argc, char **argv) consistently in Makefile.PL,
5 should help with
6 [rt.cpan.org #23868] nanosleep not detected under Mac OS 10.3.9 starting with Time::HiRes 1.96
7 - if someone still has the locale-broken Perl 5.8.0,
8 suggest that they upgrade their Perl
9
9e000d5b 101.9702 [2006-12-06]
f3370602 11 - restore the -DATLEASTFIVEOHOHFIVE, Win32 needed it still
9e000d5b
SH
12
131.9701 [2006-12-04]
14 - upgrade to ppport.h 3.10_02
15 - remove the -DATLEASTFIVEOHOHFIVE
16 - use the ppport.h PL_ppaddr, PL_statcache, PL_laststatval
17 - use the ppport.h aTHXR for calling Perl stat()
18 - switch into four-digit version since 2.0 is coming up
19 awfully fast but not feeling like a major rewrite
20
211.97 [2006-11-30]
22 - 1.95 broke building in Win32 (since pp_stat is not exported),
23 figured out how to call an op directly in 5.005 (use Perl_ppaddr
24 instead of PL_ppaddr)
25 - backport to Perl 5.004_05 (requires using statcache
26 and laststatval instead of PL_statcache and PL_laststatval)
27 (also checked to work in 5.005_04, 5.6.1, and 5.8.8 with threads)
28
291.96 [2006-11-30]
30 - 1.95 broke builds for threaded Perls, rt.cpan.org tickets:
31 [rt.cpan.org #23694] Time::HiRes fails tests on Solaris and Perl 5.6.1
32 [rt.cpan.org #23712] Time-HiRes 1.95 Fails make on AIX 5.2 with Perl 5.8.8
33 [rt.cpan.org #23730] Time::HiRes 1.95 fails make on MacOS X 10.3.9/perl 5.8.8
71ac5476
SP
34 - use main() prototype consistently in Makefile.PL
35
9e000d5b 361.95 [2006-11-29]
7b9ffbef
SP
37 - integrate core change #29180: Silence VC++ compiler warnings
38 from Steve Hay
39 - do not use PL_ppaddr in stat() because that is not available
40 in Perl 5.005_04
41 - regenerate fallback/*.inc for older Perls without
42 ExtUtils::Constant because of d_hires_stat, resolves
43 [rt.cpan.org #23694] Time::HiRes fails tests on Solaris and Perl 5.6.1
9e000d5b 44 - Make Makefile.PL more defensive against false PERL_CORE
7b9ffbef 45
c36568be
SP
461.94 [2006-10-16]
47 - file timestamps oddities seen: the atime and mtime
48 can be out of sync (modify first and read second can leave
49 atime < mtime) and mtime can be subsecond while atime is not.
50 So make the test more forgiving.
51
c09e847b
SP
521.93 [2006-10-15]
53 - the ualarm() tests (34-37) assumed that ualarm(N)
54 could never alarm in less than N seconds, widened
55 the acceptable relative range to 0.9..1.5. Addresses
56 [rt.cpan.org #22090] and [rt.cpan.org #22091].
57
58 - skip the stat() tests in cygwin and win32, because
59 if run on FAT the timestamp granularity is only 2 seconds.
60 Any good way to detect (cygwin or win32) whether we are
7b9ffbef 61 being run on NTFS or anywhere with better timestamps?
c09e847b
SP
62 Addresses [rt.cpan.org #22089] and [rt.cpan.org #22098].
63
75d5269b
SP
641.92 [2006-10-13]
65 - scan for subsecond resolution timestamps in struct stat,
66 some known possibilities:
67
68 (1) struct timespec st_atimespec;
69 st_atimespec.tv_nsec;
70 (2) time_t st_atime;
71 long st_atimensec;
72 (3) time_t st_atime;
73 int st_atime_n;
74 (4) timestruc_t st_atim;
75 st_atim.tv_nsec
76 (5) time_t st_atime;
77 int st_uatime;
78
79 If something like this is found, one can do
80
81 use Time::HiRes;
82 my @stat = Time::HiRes::stat();
83
84 or even override the standard stat():
85
86 use Time::HiRes qw(stat);
87
88 to get the stat() timestamps
89
90 my ($atime, $mtime, $ctime) = @stat[8, 9, 10];
91
92 with subsecond resolution (assuming both the operating
93 system and the filesystem support that kind of thing).
94
95 Contributions for more systems (especially non-UNIX,
7b9ffbef 96 e.g. but not limited to: Win32, VMS, OS/2) gladly accepted.
f3370602 97 (also more UNIX variants welcome: HP-UX? IRIX?)
75d5269b
SP
98
99 Thanks to H.Merijn Brand, John Peacock, and Craig
100 Berry for brave beta testing.
101
1021.91 [2006-09-29]
34f69483 103 - ualarm() in SuSE 10.1 was overflowing after ~4.2 seconds,
f3370602
SP
104 possibly due to a glibc bug/feature (suspected overflow at
105 2**32 microseconds?), workaround by using the setitimer()
106 implementation of ualarm() if either useconds or
75d5269b
SP
107 interval > 999_999 (this case seems to vary between systems:
108 are useconds more than 999_999 for ualarm() defined or not)
109 Added more ualarm() tests to catch various overflow points,
110 hopefully no problems in various platforms.
111 (The problem report by Mark Seger and Jon Paul Sullivan of HP.)
34f69483 112
f445b110
SP
1131.90 [2006-08-22]
114 - tweak still needed for Const64(), from Jerry Hedden
115 - get a freshly generated ppport.h
116 - update Copyright years
117
b69b1b83
SP
1181.89 [2006-08-22]
119 - Const64() already appends an 'LL' (or i64), so provide LL and i64
f445b110
SP
120 forms for the IV_1E[679] (effects Win32 and Cygwin), reported by
121 Jerry Hedden.
75d5269b
SP
122 - the Changes entry for 1.88 talked about [IN]V_1[679],
123 missing the 'E'.
b69b1b83 124
47e797f6
SP
1251.88 [2006-08-21]
126 - clean up the g++ warnings in HiRes.xs, all of them
127 about mixing integer and floating point, introduce
b69b1b83 128 constants IV_1E[679] and NV_1E[679]
47e797f6 129
9b15102a
SP
1301.87 [2006-02-13]
131 - [rt.cpan.org #17442] 'make test' frequently fails under
132 Cygwin Perl v5.8.8, reported and patched by J. R. Hedden
133 (two race condition bugs in the END block in the case the
134 main process dies before the timer process, unearthed
135 by a bug in Cygwin ualarm)
136
ff7df920
SP
1371.86 [2005-12-17]
138 - HiRes.t:s/ok 32/ok 33/, from Dominic Dunlop
139 - tighten up the clock() test marginally by requiring non-negative
140 - clock_nanosleep() and clock() doc tweaks
141
170c5524
SP
1421.85 [2005-12-16]
143 - the interface to clock_nanosleep() is more natural
144 when it is like (hires) time() (instead of like nanosleep),
145 and the .xs implementation of clock_nanosleep() in 1.84
146 was broken anyway
147 - the semantics of clock() are not quite so silly as I thought,
148 but still somewhat odd, documented as such
149 - additional enhancements to the clock() documentation
150 - add test for clock_nanosleep() (I cannot test this
151 since none of my systems have the function)
152 - add test for clock()
153
1541.84 [2005-12-16]
155 - add clock() which returns the processor time in
156 (floating point) seconds since an arbitrary era
157 - add clock_nanosleep() which suspends the current
158 thread until either absolute time or for relative time
159 - [rt.cpan.org #16486] printf missing value in HiRes.t
160 - add constants CLOCKS_PER_SEC, CLOCK_SOFTTIME, TIMER_ABSTIME
161 - tiny typo fixes
162
a5929ff3
RGS
1631.83 [2005-11-19]
164 - has_symbol() was wrong since e.g. ITIMER_VIRTUAL is exported
165 via @EXPORT_OK even when it is not available. This is heinous.
166 @EXPORT_OK should be determined at Makefile.PL time.
167 - be more lenient is testing clock_gettime(): allow more slop,
168 and retry up to three times, sleeping a random nap between
169 the retries
170 - human months are one-based (noticed by Anton Berezin)
171
82cbdcc3
SP
1721.82 [2005-10-06]
173 - CLOCK_REALTIME is an enum value (of the clockid_t enum)
174 in HP-UX (and might be so elsewhere, too), debugged by
175 H. Merijn Brand
176 - include const-c.inc as late as possible (from Randy Kobes,
177 [rt.cpan.org #15552] to avoid undefined usleep() on Win32
178
a5929ff3 1791.81 [2005-11-05]
b311af62
SP
180 - try to be more robust and consistent in the detection of
181 CLOCK_REALTIME and ITIMER_VIRTUAL in HiRes.t: the proper
182 way is
183
184 sub has_symbol {
185 my $symbol = shift;
186 eval 'import Time::HiRes qw($symbol)';
187 return 0 unless $@ eq '';
188 return exists ${"Time::HiRes::$symbol"};
189 }
190
191 and then use
192
193 &FOO_BAR
194
195 in the test. All these moves are needed because
196
197 1) one cannot directly do eval 'Time::HiRes::FOO_BAR'
198 because FOO_BAR might have a true value of zero
199 (or in the general case an empty string or even undef)
200
201 2) In case FOO_BAR is not available in this platform,
202 &FOO_BAR avoids the bareword warning
203
204 - wait more (1.5 seconds instead of 0.1) for the CLOCK_REALTIME test
205 but expect the 'customary' slop of 0.20 instead of 0.25
206 - fixed inside a comment HAS_POLL -> TIME_HIRES_NANOSLEEP
207 - at the end of HiRest.t tell how close we were to termination
208
a5929ff3 2091.80 [2005-11-04]
70cf0185
SP
210 - Gisle noticed a mistake (using HAS_NANOSLEEP) in 1.79
211
a5929ff3 2121.79 [2005-11-03]
e5433ad8
SP
213 - try nanosleep for emulating usleep -- may help in some weird
214 embedded realtime places which have nanosleep but neither usleep
215 nor select nor poll (doesn't have to be weird embedded realtime
216 place, though -- in many places usleep is nanosleep anyway)
217 - try poll for emulating usleep -- this may help some obscure/old
218 SVR4 places that have neither usleep nor select
219 - a redundant test guard in HiRes.t
220
a5929ff3 2211.78 [2005-11-03]
1a7d3a53 222 - ITIMER_VIRTUAL detection in HiRes.t had problems (that we cannot
e5433ad8 223 in the general case fail already at 'use' phase is suboptimal)
1a7d3a53
SP
224 - fixes to the documentation of clock_gettime() and clock_getres()
225
a5929ff3 2261.77 [2005-11-03]
ced84e60
SP
227 - add support for the POSIX clock_gettime() and clock_getres(),
228 if available, either as library calls or as syscalls
229 - be more defensive about missing functionality: break out
230 early (during 'use') if no e.g. clock_getres() is available,
231 and protect our back by trapping those cases also in HiRes.xs
232 - the test added in 1.76 could cause an endless loop e.g. in Solaris,
233 due to mixing of sleep() and alarm() (bad programmer, no cookie!)
234
3d0346a5
SP
2351.76 [2005-10-22]
236 - testing for nanosleep had wrong logic which caused nanosleep
237 to become undefined for e.g. Mac OS X
238 - added a test for a core dump that was introduced by Perl 5.8.0
239 safe signals and was fixed for the time of 5.8.1 (one report of
240 the core dump was [perl #20920]), the test skipped pre-5.8.1.
241 - *cough* s/unanosleep/nanosleep/g; *cough*
242
9649b81b
SP
2431.75 [2005-10-18]
244 - installation patch from Gisle Aas: in Perls 5.8.x and later
245 use MakeMaker INSTALLDIRS value of 'perl' instead of 'site'.
246
06252d99
RGS
2471.74 [2005-09-19]
248 - [cpan #14608] Solaris 8 perl 5.005_03 File::Spec module does not have method rel2abs
249 (the workaround is not to use rel2abs, should not be necessary)
250 - [cpan #14642] U2time wrongly exported on the C API
251 (patch supplied by the reporter, SALVA@cpan.org)
252 - add release dates to Changes
253
2541.73 [2005-08-16]
255 - Time::HiRes::nanosleep support for Solaris [PATCH]
937b804e
RGS
256 (POSIX::uname() not available if building with core perl,
257 from Gisle Aas, via perl5-porters, perl change #25295)
258
06252d99
RGS
2591.72 [2005-07-01]
260 - going back to the 1.68 loader setup (using DynaLoader)
0cf8ddea
RGS
261 since too many weird things starting breaking
262 - fix a typo in José Auguste-Etienne's name
263
06252d99
RGS
2641.71 [2005-06-28]
265 - a thinko in the nanosleep() detection
56c1b3bd
RGS
266 - move more changes stuff from the README to Changes
267 - add -w to the Makefile.PL
268
06252d99
RGS
2691.70 [2005-06-26]
270 - oops in 1.69 about @ISA (not affecting anything but silly)
f8daf111
RGS
271 - add copyright 2005 to HiRes.pm
272 - add copyright and license to HiRes.xs
56c1b3bd 273 - add copyrights 2003, 2004, 2005 to README
f8daf111 274
06252d99
RGS
2751.69 [2005-06-25]
276 - actually run a test for nanosleep
56c1b3bd 277 (if there is no $Config{d_nanosleep}) since e.g. in AIX 4.2
26e22fd9
RGS
278 it seems that one can link in nanosleep() but then calling
279 it fails instantly and sets errno to ENOSYS (Not implemented).
280 This may be fixable in the AIX case by figuring out the right
281 (realtime POSIX?) libs and whatnot, but in the general case
282 running a real test case is better. (Of course, this change
283 will no doubt run into portability problems because of the
0cf8ddea
RGS
284 execution step...) Note that because of hysterical raisins
285 most Perls do NOT have $Config{d_nanosleep} (scanning for
286 it by Configure would in many platforms require linking in
287 things like -lrt, which would in many platforms be a bad idea
288 for Perl itself).
289 (from José Auguste-Etienne)
26e22fd9
RGS
290 - support XSLoader also since it's much faster
291 (from Alexey Tourbin)
292 - add SEE ALSO (BSD::Resource and Time::TAI64)
293
06252d99 2941.68 [2005-05-14]
993164ab
RGS
295 - somehow 1.67 had a lot of doubled lines (a major cut-and-paste
296 error suspected), but miraculously it still worked since the
297 doubling took place below the __END__ token
298 - undef Pause() before defining it to avoid redefinition warnings
299 during compilation in case perl.h had already defined Pause()
300 (part of perl change #24271)
301 - minor doc tweaks
302
06252d99 3031.67 [2005-05-04]
993164ab
RGS
304 - (internal) don't ignore the return value of gettimeofday()
305 - (external) return undef or an empty if the C gettimeofday() fails
306 (affects Time::HiRes gettimeofday() and the hires time())
307
06252d99 3081.66 [2004-12-19]
44d3ce20
RGS
309 - add nanosleep()
310 - fix the 'hierachy' typo in Makefile.PL [rt.cpan.org #8492]
311 - should now build in Solaris [rt.cpan.org #7165] (since 1.64)
312 - should now build in Cygwin [rt.cpan.org #7535] (since 1.64)
82cbdcc3
SP
313 - close also [rt.cpan.org #5933] "Time::HiRes::time does not
314 pick up time adjustments like ntp" since ever reproducing it
315 (and therefore verifying a possible fix) in the same environment
44d3ce20
RGS
316 has become rather unlikely
317
06252d99 3181.65 [2004-09-18]
64a7a97c
RGS
319 - one should not mix u?alarm and sleep (the tests modified
320 by 1.65, #12 and #13, hung in Solaris), now we just busy
321 loop executing an empty block
322 - in the documentation underline the unspecificity of mixing
323 sleeps and alarms
324 - small spelling fixes
325
06252d99 3261.64 [2004-09-16]
64a7a97c
RGS
327 - regenerate ppport.h with Devel::PPPort 3.03,
328 now the MY_CXT_CLONE is defined in ppport.h,
329 we no more need to do that.
330
331 - the test #12 would often hang in sigsuspend() (at least that's
332 where Mac OS X' ktrace shows it hanging). With the sleep()s
333 changed to sleep(1)s, the tests still pass but no hang after
334 a few hundred repeats.
335
06252d99 3361.63 [2004-09-01]
1fbb4de4 337 - Win32 and any ithread build: ppport.h didn't define
170c5524 338 MY_CXT_CLONE, which seems to be a Time-HiRes-ism.
1fbb4de4 339
06252d99 3401.62 [2004-08-31]
1fbb4de4
RGS
341 - Skip testing if under PERL_CORE and Time::HiRes has not
342 been Configured (from Marcus Holland-Moritz, core change
343 #23246)
344 - Use ppport.h generated by Devel::PPPort 3.01,
345 allowing cutting away our own portability code.
346 - Don't use $ENV{PERL_CORE} for < 5.6.0.
347 - Don't use "for my $i" for <= 5.003.
348 - Don't use Pause() for <= 5.003.
349 - Can't use newSVpvf for <= 5.003.
350 (most of the changes from Marcus)
351
06252d99 3521.61 [2004-08-21]
4ed0e2d4
RGS
353 - Win32: reset reading from the performance counters every
354 five minutes to better track wall clock time (thanks to
355 PC timers being often quite bad), should help long-running
356 programs.
357
06252d99 3581.60 [2004-08-15]
4ed0e2d4
RGS
359 - Win32: Patch from Steve Hay
360 [PATCH] Re: [perl #30755] [Win32] Different results from Time::HiRes::gettimeofdayunder the debugger
361 to [perl #30755] reported by Nigel Sandever
362
363 - Cygwin: Use the Win32 recalibration code also in Cygwin if the
364 <w32api/windows.h> APIs are available. Cygwin testing by
365 Yitzchak Scott-Thoennes.
366
367 - Solaris: use -lposix4 to get nanosleep for Solaris 2.6,
368 after that keep using -lrt, patch from Alan Burlison,
369 bug reported in [cpan #7165]
370
06252d99 3711.59 [2004-04-08]
d8cb5b61
RGS
372 - Change the Win32 recalibration limit to 0.5 seconds and tweak
373 the documentation to blather less about the gory details of the
374 Win32 implementation and more about the complications in general
375 of meddling with the system clock.
376
06252d99 3771.58 [2004-04-08]
d8cb5b61
RGS
378 - Document the 1.57 change better.
379
06252d99 3801.57 [2004-07-04]
d8cb5b61
RGS
381 - Win32/Cygwin/MinGW: if the performance counter drifts by more
382 than two seconds from the system clock (due to ntp adjustments,
c1dc6e7c
RGS
383 for example), recalibrate our internal counter: from Jan Dubois,
384 based on [cpan #5933] by Jerry D. Hedden.
385
06252d99 3861.56 [2004-29-02]
1caec985
DM
387 - Give a clearer message if the tests timeout (perl change #22253)
388 - Don't use /tmp or its moral equivalents (perl bug #15036,
389 perl change #22258)
390
06252d99 3911.55 [2004-01-14]
4ed0e2d4 392 - Windows: mingw32 patch from Mike Pomraning (use Perl's Const64()
b6136f41
RGS
393 instead of VC-specific i64 suffix)
394
06252d99 3951.54 [2003-12-31]
0225372c
RGS
396 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
397
06252d99 3981.53 [2003-12-30]
0225372c
RGS
399 - Windows: higher resolution time() by using the Windows
400 performance counter API, from Jan Dubois and Anton Shcherbinin.
401 The exact new higher resolution depends on the hardware,
402 but it should be quite a bit better than using the basic
403 Windows timers.
404
06252d99 4051.52 [2003-10-28]
690f7c5f
RGS
406 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
407 the subtest 18. No known analysis nor fix, but added
408 an alarm (that requires fork() and alarm()) to the test.
409
06252d99 4101.51 [2003-09-22]
b85af263
JH
411 - doc tweaks from mjd (perl change #20456)
412 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
413
06252d99 4141.50 [2003-08-02]
d7358e6a
JH
415 - add a message (for non-core builds) to Makefile.PL about
416 the LC_ALL=C workaround
417 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
418 - the nanosleep() probe was broken (perl change #20061)
419 - use existence instead of definedness for feature probes
420 (perl change #20043)
421 - MPE/iX tweak (perl change #20042)
422 - do not use HAS_NANOSLEEP (perl change #19898)
423
06252d99 4241.49 [2003-06-23]
df16a331
JH
425 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
426 - OS/2 can always mix subsecond sleeps with signals
427 (part of perl change #19789)
428
06252d99 4291.48 [2003-06-04]
2d9815af
JH
430 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
431 (perl change #19592)
432
06252d99 4331.47 [2003-05-03]
2d9815af 434 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
1a08a6ab
ML
435 - unnecessary (nanosleep is in libc anyway)
436 - harmful (-lrt slows down execution)
437 - incompatible (with many distributions' pthreads)
438
06252d99 4391.46 [2003-04-25]
25ca0970
JH
440 - do not create files in blib directories under core
441 (perl change #19160, from rgs)
442 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
443
06252d99 4441.45 [2003-04-01]
25ca0970
JH
445 - guarantee that $xdefine in HiRes.t is always defined
446 (perl change #19109, from IlyaZ)
447 - a cleaner way to detect PERL_CORE (perl change #19111,
448 from IlyaZ)
449
06252d99 4501.44 [2003-03-30]
0be47ac6
JH
451 - add hints/irix.pl to turn off overly POSIX flags that
452 cause hide struct timespec to be hidden (and compilation
25ca0970 453 to fail) (bleadperl change #19085)
0be47ac6
JH
454 - documentation tweaks
455
06252d99 4561.43 [2003-03-11]
a0c8e3cf
JH
457 - add c:/temp to the list of temp directories to probe
458 so that cygwin (and win*?) builds are happy. This was
459 needed at least in my cygwin 1.3.20/w2k setup.
460
06252d99 4611.42 [2003-01-07]
98b50af3
JH
462 - modernize the constants code (from Nicholas Clark)
463
06252d99 4641.41 [2003-01-03]
98b50af3
JH
465 - At some point the ability to figure our the correct incdir
466 for EXTERN.h (either a core perl build, or an installed perl)
467 had broken (which lead into all test compiles failing with
468 a core perl build, but thanks to the robustness of Makefile.PL
a0c8e3cf
JH
469 nothing of this was visible). The brokenness seemed to be
470 caused by $ENV{PERL_CORE} not being on for core builds?
471 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
472 right, and both styles of build should work again.
98b50af3 473
06252d99 4741.40 [2003-01-03]
98b50af3
JH
475 - Nicholas Clark noticed that the my_catdir() emulation function
476 was broken (which means that we didn't really work for Perls
477 5.002 and 5.003)
478 - inspired by fixing the above made the whole Makefile.PL -w
479 and strict clean
480 - tightened up the Makefile.PL output, less whitespace
481
06252d99 4821.39 [2003-10-20]
6a20eacc
HS
483 - fix from Craig Berry for better building in VMS with PERL_CORE
484
06252d99 4851.38 [2003-10-13]
3f2ee006
HS
486 - no functional changes
487 - move lib/Time/HiRes.pm as Hires.pm
488 - libraries scanning was slightly broken (always scanned
489 for a library even when $Config{libs} already had it)
490
06252d99 4911.37 [2003-09-23]
3f2ee006
HS
492 - Ray Zimmerman ran into a race condition in Mac OS X.
493 A 0.01-second alarm fired before the test expected.
494 The test first slept indefinitely (blocking for signals)
495 and only after that tested for the signal having been sent.
496 Since the signal had already been sent, the test #12 never
497 completed. The solution: test first, then block.
498 - default to being silent on all probing attempts, set the
499 environment variable VERBOSE to a true value to see the
500 details (the probing command and the possible errors)
501
06252d99 5021.36 [2003-09-12]
3f2ee006
HS
503 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
504 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
505 cannot be upgraded from CPAN to override the 5.8.0 version
506 (Guido A. Ostkamp)
507 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
508 because the include directories did not adjust themselves
509 if $ENV{PERL_CORE} (Hugo van der Sanden)
510 - add documentation about the restart of select() under alarm()
511
06252d99 5121.35 [2003-08-24]
3f2ee006
HS
513 - small documentation tweaks
514
515
06252d99 5161.34 [2003-08-22]
3f2ee006
HS
517 - better VMS operation (Craig Berry)
518
06252d99 5191.33 [2003-08-20]
3f2ee006
HS
520 - our time machine is accelerating: now works with Perl 5.004_01
521 (tried with 5.003_07 and 5.002 but I get segmentation faults
522 from running the Makefile.PL with those in Tru64 4.0D)
523
06252d99 5241.32 [2003-08-20]
3f2ee006
HS
525 - backward compatibility (pre-5.6.0) tweaks:
526 - no XSLoader in 5.00503, use DynaLoader instead
527 - no SvPV_nolen, either
528 - no PerlProc_pause(), either
529 - now tested with 5.00404 and 5.00503
530 - Makefile.PL requires 5.00404 (no more 5.002)
531 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
532 (this means that one can mix subsecond sleeps with alarms)
533 - because of nanosleep we probe for -lrt and -lposix4
534 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
535 is available by exportable constants Time::HiRes::d_func
536 (since older Perl do not have them in %Config, and even
537 5.8.0 does not probe for nanosleep)
538
06252d99 5391.31 [2003-08-19]
3f2ee006
HS
540 - backward compatibility (pre-5.6.1) tweaks:
541 - define NV if no NVTYPE
542 - define IVdf if needed (note: the Devel::PPPort
543 in 5.8.0 does not try hard hard enough since
544 the IVSIZE might not be defined)
545 - define NVgf if needed
546 - grab the typemap from 5.8.0 for the NV stuff
56c1b3bd
RGS
547
548 1.31 and 1.32 add more backward compatibility (now all the way
549 back to Perl 5.00404), and using nanosleep() (if available) for
550 subsecond sleeps.
551
06252d99 5521.30 [2003-08-16]
3f2ee006
HS
553
554 - release 1.29_02 as 1.30
555
56c1b3bd
RGS
556 1.30 adds all the changes made during the Perl 5.6->5.7->5.8
557 development cycle. Most notably portability across platforms has been
558 enhanced, and the interval timers (setitimer, getitimer) have been
559 added. Note that the version of Time::HiRes that is included in Perl
560 5.8.0 calls itself 1.20_00, but it is equivalent to this Time::HiRes
561 version. Note also that in 1.30 Wegscheid turns over the maintenance
562 to Jarkko Hietaniemi.
563
06252d99 5641.29_02 [2003-08-16]
3f2ee006
HS
565
566 - fix a silly unclosed comment typo in HiRes.xs
567 - document and export REALTIME_REALPROF (Solaris)
568
06252d99 5691.29_01 [2003-08-16]
3f2ee006
HS
570
571 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
572 (need to patch this also in Perl 5.[89])
573 - remove CVS revision log from HiRes.xs
574
06252d99 5751.29_00 [2003-08-14]
3f2ee006
HS
576
577 The following numbered patches refer to the Perl 5.7 changes,
578 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
579
580 - 17558: Add #!./perl to the .t
581 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
582 - 16198: political correctness, from Simon Cozens
583 - 15857: doc tweaks, from Jarkko Hietaniemi
584 - 15593: optimization in .xs, from Paul Green
585 - 14892: pod fixes, from Robin Barker
586 - 14100: VOS fixes, from Paul Green
587 - 13422: XS segfault, from Marc Lehmann
588 - 13378: whether select() gets restarted on signals, depends
589 - 13354: timing constraints, again, from Andy Dougherty
170c5524 590 - 13278: can't do subsecond alarms with ualarm;
3f2ee006
HS
591 break out early if alarms do not seem to be working
592 - 13266: test relaxation (cygwin gets lower hires
593 times than lores ones)
594 - 12846: protect against high load, from Jarkko Hietaniemi
595 - 12837: HiRes.t VMS tweak, from Craig A. Berry
596 - 12797: HiRes.t VMS tweak, from Charles Lane
597 - 12769: HiRes.t VMS tweak, from Craig A. Berry
598 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
599 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
600 - 12692: alarm() ain't gonna work if ualarm() ain't,
601 from Gurusamy Sarathy
602 - 12680: minor VMS tweak, from Charles Lane
603 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
604 - 12609: croak on negative time, from Jarkko Hietaniemi
605 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
606 - 12594: MacOS Classic timeofday, from Chris Nandor
607 - 12473: allow for more than one second for sleep() and usleep()
608 - 12458: test tuning, relax timing constraints,
609 from Jarkko Hietaniemi
610 - 12449: make sleep() and usleep() to return the number
611 of seconds and microseconds actually slept (analogously
612 with the builtin sleep()), also make usleep() croak if
613 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
614 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
615 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
616 - 12196: use ftime() on Win32, from Artur Bergman
617 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
618 - 12105: use GetSystemTime() on Win32, from Artur Bergman
619 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
620 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
621 - 11797: problem in HiRes.t, from John P. Linderman
622 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
623 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
624 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
625 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
626 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
627 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
628 from Jonathan Stowe
629 - 10942: MPE/IX test tweaks, from Mark Bixby
630 - 10784: unnecessary pod2man calls, from Andy Dougherty
631 - 10354: ext/ + -Wall, from Doug MacEachern
632 - 10320: fix the BOOT section to call myU2time correctly
633 - 10317: correct casting for AIX< from H. Merijn Brand
634 - 10119: document that the core time() may be rounding, not truncating
635 - 10118: test fix, from John Peacock
636 - 9988: long =item, from Robin Barker
637 - 9714: correct test output
638 - 9708: test also the scalar aspect of getitimer()
639 - 9705: Add interval timers (setitimer, getitimer)
640 - 9692: do not require at least 5.005 using XS
641
642 The following changes were made on top of the changes
643 made for Time::HiRes during the Perl 5.7 development
644 cycle that culminated in the release of Perl 5.8.0.
645
646 - add "require 5.005" to the Makefile.PL
647 - remove the REVISION section (CVS log) from HiRes.pm
648 - add jhi's copyright alongside Douglas'
649 - move HiRes.pm to lib/Time/
650 - move HiRes.t to t/
651 - modify HiRes.t to use $ENV{PERL_CORE}
652 - modify the original Time::HiRes version 1.20 Makefile.PL
653 to work both with Perl 5.8.0 and the new code with pre-5.8.0
654 Perls (tried with 5.6.1)
655 - tiny tweaks and updates in README and TODO
656 - bump the VERSION to 1.29
657
dcf686c9
JH
6581.20 Wed Feb 24 21:30 1999
659 - make our usleep and ualarm substitutes into hrt_usleep
660 and hrt_ualarm. This helps static links of Perl with other
661 packages that also have usleep, etc. From
662 Ilya Zakharevich <ilya@math.ohio-state.edu>
663 - add C API stuff. From Joshua Pritikin
664 <joshua.pritikin@db.com>
3f2ee006 665 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
dcf686c9
JH
666 - hopefully correct "-lc" fix for SCO.
667 - add PPD stuff
668
56c1b3bd
RGS
669 1.20 adds a platform neutral set of C accessible routines if you are
670 running 5.005+. All other changes are packaging changes and build
671 fixes(?) for statically linked Perl, SCO, and VMS.
672
dcf686c9
JH
6731.19 Tue Sep 29 22:30 1998
674 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
675 <seb@stian.demon.co.uk>
676 - change GIMME_V to GIMME to help people with older versions of
677 Perl.
678 - fix Win32 version of gettimeofday(). It didn't affect anything,
679 but it confuses people reading the code when the return value
680 is backwards (0 is success).
681 - fix Makefile.PL (more) so that detection of gettimeofday is
682 more correct.
683
56c1b3bd
RGS
684 1.19 has better VMS support.
685
dcf686c9
JH
6861.18 Mon Jul 6 22:40 1998
687 - add usleep() for Win32.
688 - fix Makefile.PL to fix reported HP/UX feature where unresolved
689 externals still cause an executable to be generated (though no
690 x bit set). Thanks to David Kozinn for report and explanation.
691 Problems with the fix are mine :)
692
56c1b3bd
RGS
693 1.18 has limited Win32 support (no ualarm). Added usleep for Win32.
694 Probably buggy. I'm sure I'll hear.
695
dcf686c9
JH
6961.17 Wed Jul 1 20:10 1998
697 - fix setitimer calls so microseconds is not more than 1000000.
698 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
3f2ee006 699 - make Win32. We only get gettimeofday (the select hack doesn't
dcf686c9 700 seem to work on my Win95 system).
3f2ee006 701 - fix test 4 on 01test.t. add test to see if time() and
dcf686c9
JH
702 Time::HiRes::time() are close.
703
7041.16 Wed Nov 12 21:05 1997
705 - add missing EXTEND in new gettimeofday scalar code.
706
56c1b3bd
RGS
707 1.16+ should be closer to building out of the box on Linux. Thanks
708 to Gisle Aas for patches, and the ualarm equivalent using setitimer.
709
710 If your underlying operating system doesn't implement ualarm(), then
711 a fake using setitimer() will be made. If the OS is missing usleep(),
712 a fake one using select() will be made. If a fake can't be made for
713 either ualarm() or usleep(), then the corresponding Perl function will
714 not be available. If the OS is missing gettimeofday(), you will get
715 unresolved externals, either at link- or run-time.
716
717 This is an improvement; the package used to not even build if
718 you were missing any of these bits. Roderick Schertler
719
720 <roderick@argon.org> did all the conditional compilation stuff,
721 look at HiRes.pm and the test suites; it's good educational reading.
722
dcf686c9
JH
7231.15 Mon Nov 10 21:30 1997
724 - HiRes.pm: update pod. Provided by Gisle Aas.
725 - HiRes.xs: if gettimeofday() called in scalar context, do
726 something more useful than before. Provided by Gisle Aas.
727 - README: tell of xsubpp '-nolinenumber' woes. thanks to
728 Edward Henigin <ed@texas.net> for pointing out the problem.
729
7301.14 Wed Nov 5 9:40 1997
731 - Makefile.PL: look for setitimer
732 - HiRes.xs: if missing ualarm, but we have setitimer, make up
733 our own setitimer. These were provided by Gisle Aas.
734
7351.13 Tue Nov 4 23:30 1997
736 - Makefile.PL: fix autodetect mechanism to do try linking in addition
737 to just compiling; should fix Linux build problem. Fix was provided
738 by Gisle Aas.
739
7401.12 Sun Oct 12 12:00:00 1997
741 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
742 you may need to comment this back out if you have an older xsubpp.
743 - HiRes.xs: set PROTOTYPES: DISABLE
744
7451.11 Fri Sep 05 16:00:00 1997
746 - Makefile.PL:
747 Had some line commented out that shouldn't have been (testing
748 remnants)
749 - README:
750 Previous version was corrupted.
751
7521.10 Thu May 22 20:20:00 1997
753 - HiRes.xs, HiRes.pm, t/*:
754 - only compile what we have OS support for (or can
755 fake with select())
756 - only test what we compiled
757 - gross improvement to the test suite
758 - fix EXPORT_FAIL.
759 This work was all done by Roderick Schertler
760 <roderick@argon.org>. If you run Linux or
170c5524 761 one of the other ualarm-less platforms, and you like this
dcf686c9
JH
762 module, let Roderick know; without him, it still wouldn't
763 be working on those boxes...
764 - Makefile.PL: figure out what routines the OS has and
765 only build what we need. These bits were written by Jarkko
766 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
767
7681.02 Mon Dec 30 08:00:00 1996
769 - HiRes.pm: update documentation to say what to do when missing
770 ualarm() and friends.
771 - README: update to warn that ualarm() and friends need to exist
772
7731.01 Fri Oct 17 08:00:00 1996
774 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
775 - HiRes.pm: put blank line between __END__ and =head1 so that
776 pod2man works.
777
7781.00 Tue Sep 03 13:00:00 1996
779 - original version; created by h2xs 1.16