| 1 |
Revision history for Perl extension Time::HiRes. |
| 2 |
|
| 3 |
1.66 |
| 4 |
- add nanosleep() |
| 5 |
- fix the 'hierachy' typo in Makefile.PL [rt.cpan.org #8492] |
| 6 |
- should now build in Solaris [rt.cpan.org #7165] (since 1.64) |
| 7 |
- should now build in Cygwin [rt.cpan.org #7535] (since 1.64) |
| 8 |
- 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 |
| 9 |
has become rather unlikely |
| 10 |
|
| 11 |
1.65 |
| 12 |
- one should not mix u?alarm and sleep (the tests modified |
| 13 |
by 1.65, #12 and #13, hung in Solaris), now we just busy |
| 14 |
loop executing an empty block |
| 15 |
- in the documentation underline the unspecificity of mixing |
| 16 |
sleeps and alarms |
| 17 |
- small spelling fixes |
| 18 |
|
| 19 |
1.64 |
| 20 |
- regenerate ppport.h with Devel::PPPort 3.03, |
| 21 |
now the MY_CXT_CLONE is defined in ppport.h, |
| 22 |
we no more need to do that. |
| 23 |
|
| 24 |
- the test #12 would often hang in sigsuspend() (at least that's |
| 25 |
where Mac OS X' ktrace shows it hanging). With the sleep()s |
| 26 |
changed to sleep(1)s, the tests still pass but no hang after |
| 27 |
a few hundred repeats. |
| 28 |
|
| 29 |
1.63 |
| 30 |
- Win32 and any ithread build: ppport.h didn't define |
| 31 |
MY_CXT_CLONE, which seems to be a Time-HiResism. |
| 32 |
|
| 33 |
1.62 |
| 34 |
- Skip testing if under PERL_CORE and Time::HiRes has not |
| 35 |
been Configured (from Marcus Holland-Moritz, core change |
| 36 |
#23246) |
| 37 |
- Use ppport.h generated by Devel::PPPort 3.01, |
| 38 |
allowing cutting away our own portability code. |
| 39 |
- Don't use $ENV{PERL_CORE} for < 5.6.0. |
| 40 |
- Don't use "for my $i" for <= 5.003. |
| 41 |
- Don't use Pause() for <= 5.003. |
| 42 |
- Can't use newSVpvf for <= 5.003. |
| 43 |
(most of the changes from Marcus) |
| 44 |
|
| 45 |
1.61 |
| 46 |
- Win32: reset reading from the performance counters every |
| 47 |
five minutes to better track wall clock time (thanks to |
| 48 |
PC timers being often quite bad), should help long-running |
| 49 |
programs. |
| 50 |
|
| 51 |
1.60 |
| 52 |
- Win32: Patch from Steve Hay |
| 53 |
[PATCH] Re: [perl #30755] [Win32] Different results from Time::HiRes::gettimeofdayunder the debugger |
| 54 |
to [perl #30755] reported by Nigel Sandever |
| 55 |
|
| 56 |
- Cygwin: Use the Win32 recalibration code also in Cygwin if the |
| 57 |
<w32api/windows.h> APIs are available. Cygwin testing by |
| 58 |
Yitzchak Scott-Thoennes. |
| 59 |
|
| 60 |
- Solaris: use -lposix4 to get nanosleep for Solaris 2.6, |
| 61 |
after that keep using -lrt, patch from Alan Burlison, |
| 62 |
bug reported in [cpan #7165] |
| 63 |
|
| 64 |
1.59 |
| 65 |
- Change the Win32 recalibration limit to 0.5 seconds and tweak |
| 66 |
the documentation to blather less about the gory details of the |
| 67 |
Win32 implementation and more about the complications in general |
| 68 |
of meddling with the system clock. |
| 69 |
|
| 70 |
1.58 |
| 71 |
- Document the 1.57 change better. |
| 72 |
|
| 73 |
1.57 |
| 74 |
- Win32/Cygwin/MinGW: if the performance counter drifts by more |
| 75 |
than two seconds from the system clock (due to ntp adjustments, |
| 76 |
for example), recalibrate our internal counter: from Jan Dubois, |
| 77 |
based on [cpan #5933] by Jerry D. Hedden. |
| 78 |
|
| 79 |
1.56 |
| 80 |
- Give a clearer message if the tests timeout (perl change #22253) |
| 81 |
- Don't use /tmp or its moral equivalents (perl bug #15036, |
| 82 |
perl change #22258) |
| 83 |
|
| 84 |
1.55 |
| 85 |
- Windows: mingw32 patch from Mike Pomraning (use Perl's Const64() |
| 86 |
instead of VC-specific i64 suffix) |
| 87 |
|
| 88 |
1.54 |
| 89 |
- Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep |
| 90 |
|
| 91 |
1.53 |
| 92 |
- Windows: higher resolution time() by using the Windows |
| 93 |
performance counter API, from Jan Dubois and Anton Shcherbinin. |
| 94 |
The exact new higher resolution depends on the hardware, |
| 95 |
but it should be quite a bit better than using the basic |
| 96 |
Windows timers. |
| 97 |
|
| 98 |
1.52 |
| 99 |
- In AIX (v?) with perl 5.6.1 the HiRes.t can hang after |
| 100 |
the subtest 18. No known analysis nor fix, but added |
| 101 |
an alarm (that requires fork() and alarm()) to the test. |
| 102 |
|
| 103 |
1.51 |
| 104 |
- doc tweaks from mjd (perl change #20456) |
| 105 |
- NCR MP-RAS hints file added (svr4.pl) (perl change #21249) |
| 106 |
|
| 107 |
1.50 |
| 108 |
- add a message (for non-core builds) to Makefile.PL about |
| 109 |
the LC_ALL=C workaround |
| 110 |
- &Time::HiRes::d_nanosleep was broken (perl change #20131) |
| 111 |
- the nanosleep() probe was broken (perl change #20061) |
| 112 |
- use existence instead of definedness for feature probes |
| 113 |
(perl change #20043) |
| 114 |
- MPE/iX tweak (perl change #20042) |
| 115 |
- do not use HAS_NANOSLEEP (perl change #19898) |
| 116 |
|
| 117 |
1.49 |
| 118 |
- UVuf for non-IVSIZE platforms (from Keiichiro Nagano) |
| 119 |
- OS/2 can always mix subsecond sleeps with signals |
| 120 |
(part of perl change #19789) |
| 121 |
|
| 122 |
1.48 |
| 123 |
- workaround for buggy gcc 2.95.3 in openbsd/sparc64 |
| 124 |
(perl change #19592) |
| 125 |
|
| 126 |
1.47 |
| 127 |
- do not use -lrt in Linux (from March Lehmann, perl change #19449) |
| 128 |
- unnecessary (nanosleep is in libc anyway) |
| 129 |
- harmful (-lrt slows down execution) |
| 130 |
- incompatible (with many distributions' pthreads) |
| 131 |
|
| 132 |
1.46 |
| 133 |
- do not create files in blib directories under core |
| 134 |
(perl change #19160, from rgs) |
| 135 |
- detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd) |
| 136 |
|
| 137 |
1.45 |
| 138 |
- guarantee that $xdefine in HiRes.t is always defined |
| 139 |
(perl change #19109, from IlyaZ) |
| 140 |
- a cleaner way to detect PERL_CORE (perl change #19111, |
| 141 |
from IlyaZ) |
| 142 |
|
| 143 |
1.44 |
| 144 |
- add hints/irix.pl to turn off overly POSIX flags that |
| 145 |
cause hide struct timespec to be hidden (and compilation |
| 146 |
to fail) (bleadperl change #19085) |
| 147 |
- documentation tweaks |
| 148 |
|
| 149 |
1.43 |
| 150 |
- add c:/temp to the list of temp directories to probe |
| 151 |
so that cygwin (and win*?) builds are happy. This was |
| 152 |
needed at least in my cygwin 1.3.20/w2k setup. |
| 153 |
|
| 154 |
1.42 |
| 155 |
- modernize the constants code (from Nicholas Clark) |
| 156 |
|
| 157 |
1.41 |
| 158 |
- At some point the ability to figure our the correct incdir |
| 159 |
for EXTERN.h (either a core perl build, or an installed perl) |
| 160 |
had broken (which lead into all test compiles failing with |
| 161 |
a core perl build, but thanks to the robustness of Makefile.PL |
| 162 |
nothing of this was visible). The brokenness seemed to be |
| 163 |
caused by $ENV{PERL_CORE} not being on for core builds? |
| 164 |
Now stole a trick from the Encode that sets $ENV{PERL_CORE} |
| 165 |
right, and both styles of build should work again. |
| 166 |
|
| 167 |
1.40 |
| 168 |
- Nicholas Clark noticed that the my_catdir() emulation function |
| 169 |
was broken (which means that we didn't really work for Perls |
| 170 |
5.002 and 5.003) |
| 171 |
- inspired by fixing the above made the whole Makefile.PL -w |
| 172 |
and strict clean |
| 173 |
- tightened up the Makefile.PL output, less whitespace |
| 174 |
|
| 175 |
1.39 |
| 176 |
- fix from Craig Berry for better building in VMS with PERL_CORE |
| 177 |
|
| 178 |
1.38 |
| 179 |
- no functional changes |
| 180 |
- move lib/Time/HiRes.pm as Hires.pm |
| 181 |
- libraries scanning was slightly broken (always scanned |
| 182 |
for a library even when $Config{libs} already had it) |
| 183 |
|
| 184 |
1.37 |
| 185 |
- Ray Zimmerman ran into a race condition in Mac OS X. |
| 186 |
A 0.01-second alarm fired before the test expected. |
| 187 |
The test first slept indefinitely (blocking for signals) |
| 188 |
and only after that tested for the signal having been sent. |
| 189 |
Since the signal had already been sent, the test #12 never |
| 190 |
completed. The solution: test first, then block. |
| 191 |
- default to being silent on all probing attempts, set the |
| 192 |
environment variable VERBOSE to a true value to see the |
| 193 |
details (the probing command and the possible errors) |
| 194 |
|
| 195 |
1.36 |
| 196 |
- do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski) |
| 197 |
- INSTALLDIRS => 'perl' missing which means that Time::HiRes |
| 198 |
cannot be upgraded from CPAN to override the 5.8.0 version |
| 199 |
(Guido A. Ostkamp) |
| 200 |
- Time::HiRes 1.35 could not be dropped as-is to bleadperl |
| 201 |
because the include directories did not adjust themselves |
| 202 |
if $ENV{PERL_CORE} (Hugo van der Sanden) |
| 203 |
- add documentation about the restart of select() under alarm() |
| 204 |
|
| 205 |
1.35 |
| 206 |
- small documentation tweaks |
| 207 |
|
| 208 |
|
| 209 |
1.34 |
| 210 |
- better VMS operation (Craig Berry) |
| 211 |
|
| 212 |
1.33 |
| 213 |
- our time machine is accelerating: now works with Perl 5.004_01 |
| 214 |
(tried with 5.003_07 and 5.002 but I get segmentation faults |
| 215 |
from running the Makefile.PL with those in Tru64 4.0D) |
| 216 |
|
| 217 |
1.32 |
| 218 |
- backward compatibility (pre-5.6.0) tweaks: |
| 219 |
- no XSLoader in 5.00503, use DynaLoader instead |
| 220 |
- no SvPV_nolen, either |
| 221 |
- no PerlProc_pause(), either |
| 222 |
- now tested with 5.00404 and 5.00503 |
| 223 |
- Makefile.PL requires 5.00404 (no more 5.002) |
| 224 |
- use nanosleep instead of usleep, if it is available (Wilson Snyder) |
| 225 |
(this means that one can mix subsecond sleeps with alarms) |
| 226 |
- because of nanosleep we probe for -lrt and -lposix4 |
| 227 |
- the existence of getitimer/nanosleep/setitimer/ualarm/usleep |
| 228 |
is available by exportable constants Time::HiRes::d_func |
| 229 |
(since older Perl do not have them in %Config, and even |
| 230 |
5.8.0 does not probe for nanosleep) |
| 231 |
|
| 232 |
1.31 |
| 233 |
- backward compatibility (pre-5.6.1) tweaks: |
| 234 |
- define NV if no NVTYPE |
| 235 |
- define IVdf if needed (note: the Devel::PPPort |
| 236 |
in 5.8.0 does not try hard hard enough since |
| 237 |
the IVSIZE might not be defined) |
| 238 |
- define NVgf if needed |
| 239 |
- grab the typemap from 5.8.0 for the NV stuff |
| 240 |
1.30 |
| 241 |
|
| 242 |
- release 1.29_02 as 1.30 |
| 243 |
|
| 244 |
1.29_02 |
| 245 |
|
| 246 |
- fix a silly unclosed comment typo in HiRes.xs |
| 247 |
- document and export REALTIME_REALPROF (Solaris) |
| 248 |
|
| 249 |
1.29_01 |
| 250 |
|
| 251 |
- only getitimer(ITIMER_REAL) available in Cygwin and Win32 |
| 252 |
(need to patch this also in Perl 5.[89]) |
| 253 |
- remove CVS revision log from HiRes.xs |
| 254 |
|
| 255 |
1.29_00 |
| 256 |
|
| 257 |
The following numbered patches refer to the Perl 5.7 changes, |
| 258 |
you can browse them at http://public.activestate.com/cgi-bin/perlbrowse |
| 259 |
|
| 260 |
- 17558: Add #!./perl to the .t |
| 261 |
- 17201: linux + usemorebits fix, from Rafael Garcia-Suarez |
| 262 |
- 16198: political correctness, from Simon Cozens |
| 263 |
- 15857: doc tweaks, from Jarkko Hietaniemi |
| 264 |
- 15593: optimization in .xs, from Paul Green |
| 265 |
- 14892: pod fixes, from Robin Barker |
| 266 |
- 14100: VOS fixes, from Paul Green |
| 267 |
- 13422: XS segfault, from Marc Lehmann |
| 268 |
- 13378: whether select() gets restarted on signals, depends |
| 269 |
- 13354: timing constraints, again, from Andy Dougherty |
| 270 |
- 13278: can't do subecond alarms with ualarm; |
| 271 |
break out early if alarms do not seem to be working |
| 272 |
- 13266: test relaxation (cygwin gets lower hires |
| 273 |
times than lores ones) |
| 274 |
- 12846: protect against high load, from Jarkko Hietaniemi |
| 275 |
- 12837: HiRes.t VMS tweak, from Craig A. Berry |
| 276 |
- 12797: HiRes.t VMS tweak, from Charles Lane |
| 277 |
- 12769: HiRes.t VMS tweak, from Craig A. Berry |
| 278 |
- 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons |
| 279 |
- 12722: VMS ualarm for VMS without ualarm, from Charles Lane |
| 280 |
- 12692: alarm() ain't gonna work if ualarm() ain't, |
| 281 |
from Gurusamy Sarathy |
| 282 |
- 12680: minor VMS tweak, from Charles Lane |
| 283 |
- 12617: don't try to print ints as IVs, from Jarkko Hietaniemi |
| 284 |
- 12609: croak on negative time, from Jarkko Hietaniemi |
| 285 |
- 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi |
| 286 |
- 12594: MacOS Classic timeofday, from Chris Nandor |
| 287 |
- 12473: allow for more than one second for sleep() and usleep() |
| 288 |
- 12458: test tuning, relax timing constraints, |
| 289 |
from Jarkko Hietaniemi |
| 290 |
- 12449: make sleep() and usleep() to return the number |
| 291 |
of seconds and microseconds actually slept (analogously |
| 292 |
with the builtin sleep()), also make usleep() croak if |
| 293 |
asked for more than 1_000_000 useconds, from Jarkko Hietaniemi |
| 294 |
- 12366: Time::HiRes for VMS pre-7.0, from Charles Lane |
| 295 |
- 12199: do not use ftime on Win32, from Gurusamy Sarathy |
| 296 |
- 12196: use ftime() on Win32, from Artur Bergman |
| 297 |
- 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy |
| 298 |
- 12105: use GetSystemTime() on Win32, from Artur Bergman |
| 299 |
- 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi |
| 300 |
- 11901: UNICOS sloppy division, from Jarkko Hietaniemi |
| 301 |
- 11797: problem in HiRes.t, from John P. Linderman |
| 302 |
- 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen |
| 303 |
- 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen |
| 304 |
- 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer |
| 305 |
- 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer |
| 306 |
- 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer |
| 307 |
- 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(), |
| 308 |
from Jonathan Stowe |
| 309 |
- 10942: MPE/IX test tweaks, from Mark Bixby |
| 310 |
- 10784: unnecessary pod2man calls, from Andy Dougherty |
| 311 |
- 10354: ext/ + -Wall, from Doug MacEachern |
| 312 |
- 10320: fix the BOOT section to call myU2time correctly |
| 313 |
- 10317: correct casting for AIX< from H. Merijn Brand |
| 314 |
- 10119: document that the core time() may be rounding, not truncating |
| 315 |
- 10118: test fix, from John Peacock |
| 316 |
- 9988: long =item, from Robin Barker |
| 317 |
- 9714: correct test output |
| 318 |
- 9708: test also the scalar aspect of getitimer() |
| 319 |
- 9705: Add interval timers (setitimer, getitimer) |
| 320 |
- 9692: do not require at least 5.005 using XS |
| 321 |
|
| 322 |
The following changes were made on top of the changes |
| 323 |
made for Time::HiRes during the Perl 5.7 development |
| 324 |
cycle that culminated in the release of Perl 5.8.0. |
| 325 |
|
| 326 |
- add "require 5.005" to the Makefile.PL |
| 327 |
- remove the REVISION section (CVS log) from HiRes.pm |
| 328 |
- add jhi's copyright alongside Douglas' |
| 329 |
- move HiRes.pm to lib/Time/ |
| 330 |
- move HiRes.t to t/ |
| 331 |
- modify HiRes.t to use $ENV{PERL_CORE} |
| 332 |
- modify the original Time::HiRes version 1.20 Makefile.PL |
| 333 |
to work both with Perl 5.8.0 and the new code with pre-5.8.0 |
| 334 |
Perls (tried with 5.6.1) |
| 335 |
- tiny tweaks and updates in README and TODO |
| 336 |
- bump the VERSION to 1.29 |
| 337 |
|
| 338 |
1.20 Wed Feb 24 21:30 1999 |
| 339 |
- make our usleep and ualarm substitutes into hrt_usleep |
| 340 |
and hrt_ualarm. This helps static links of Perl with other |
| 341 |
packages that also have usleep, etc. From |
| 342 |
Ilya Zakharevich <ilya@math.ohio-state.edu> |
| 343 |
- add C API stuff. From Joshua Pritikin |
| 344 |
<joshua.pritikin@db.com> |
| 345 |
- VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer) |
| 346 |
- hopefully correct "-lc" fix for SCO. |
| 347 |
- add PPD stuff |
| 348 |
|
| 349 |
1.19 Tue Sep 29 22:30 1998 |
| 350 |
- put VMS gettimeofday() in. Patch is from Sebastian Bazley |
| 351 |
<seb@stian.demon.co.uk> |
| 352 |
- change GIMME_V to GIMME to help people with older versions of |
| 353 |
Perl. |
| 354 |
- fix Win32 version of gettimeofday(). It didn't affect anything, |
| 355 |
but it confuses people reading the code when the return value |
| 356 |
is backwards (0 is success). |
| 357 |
- fix Makefile.PL (more) so that detection of gettimeofday is |
| 358 |
more correct. |
| 359 |
|
| 360 |
1.18 Mon Jul 6 22:40 1998 |
| 361 |
- add usleep() for Win32. |
| 362 |
- fix Makefile.PL to fix reported HP/UX feature where unresolved |
| 363 |
externals still cause an executable to be generated (though no |
| 364 |
x bit set). Thanks to David Kozinn for report and explanation. |
| 365 |
Problems with the fix are mine :) |
| 366 |
|
| 367 |
1.17 Wed Jul 1 20:10 1998 |
| 368 |
- fix setitimer calls so microseconds is not more than 1000000. |
| 369 |
Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD. |
| 370 |
- make Win32. We only get gettimeofday (the select hack doesn't |
| 371 |
seem to work on my Win95 system). |
| 372 |
- fix test 4 on 01test.t. add test to see if time() and |
| 373 |
Time::HiRes::time() are close. |
| 374 |
|
| 375 |
1.16 Wed Nov 12 21:05 1997 |
| 376 |
- add missing EXTEND in new gettimeofday scalar code. |
| 377 |
|
| 378 |
1.15 Mon Nov 10 21:30 1997 |
| 379 |
- HiRes.pm: update pod. Provided by Gisle Aas. |
| 380 |
- HiRes.xs: if gettimeofday() called in scalar context, do |
| 381 |
something more useful than before. Provided by Gisle Aas. |
| 382 |
- README: tell of xsubpp '-nolinenumber' woes. thanks to |
| 383 |
Edward Henigin <ed@texas.net> for pointing out the problem. |
| 384 |
|
| 385 |
1.14 Wed Nov 5 9:40 1997 |
| 386 |
- Makefile.PL: look for setitimer |
| 387 |
- HiRes.xs: if missing ualarm, but we have setitimer, make up |
| 388 |
our own setitimer. These were provided by Gisle Aas. |
| 389 |
|
| 390 |
1.13 Tue Nov 4 23:30 1997 |
| 391 |
- Makefile.PL: fix autodetect mechanism to do try linking in addition |
| 392 |
to just compiling; should fix Linux build problem. Fix was provided |
| 393 |
by Gisle Aas. |
| 394 |
|
| 395 |
1.12 Sun Oct 12 12:00:00 1997 |
| 396 |
- Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug; |
| 397 |
you may need to comment this back out if you have an older xsubpp. |
| 398 |
- HiRes.xs: set PROTOTYPES: DISABLE |
| 399 |
|
| 400 |
1.11 Fri Sep 05 16:00:00 1997 |
| 401 |
- Makefile.PL: |
| 402 |
Had some line commented out that shouldn't have been (testing |
| 403 |
remnants) |
| 404 |
- README: |
| 405 |
Previous version was corrupted. |
| 406 |
|
| 407 |
1.10 Thu May 22 20:20:00 1997 |
| 408 |
- HiRes.xs, HiRes.pm, t/*: |
| 409 |
- only compile what we have OS support for (or can |
| 410 |
fake with select()) |
| 411 |
- only test what we compiled |
| 412 |
- gross improvement to the test suite |
| 413 |
- fix EXPORT_FAIL. |
| 414 |
This work was all done by Roderick Schertler |
| 415 |
<roderick@argon.org>. If you run Linux or |
| 416 |
one of the other ualarm-less platoforms, and you like this |
| 417 |
module, let Roderick know; without him, it still wouldn't |
| 418 |
be working on those boxes... |
| 419 |
- Makefile.PL: figure out what routines the OS has and |
| 420 |
only build what we need. These bits were written by Jarkko |
| 421 |
Hietaniemi <jhi@iki.fi>. Again, gratitude is due... |
| 422 |
|
| 423 |
1.02 Mon Dec 30 08:00:00 1996 |
| 424 |
- HiRes.pm: update documentation to say what to do when missing |
| 425 |
ualarm() and friends. |
| 426 |
- README: update to warn that ualarm() and friends need to exist |
| 427 |
|
| 428 |
1.01 Fri Oct 17 08:00:00 1996 |
| 429 |
- Makefile.PL: make XSPROTOARGS => '-noprototyopes' |
| 430 |
- HiRes.pm: put blank line between __END__ and =head1 so that |
| 431 |
pod2man works. |
| 432 |
|
| 433 |
1.00 Tue Sep 03 13:00:00 1996 |
| 434 |
- original version; created by h2xs 1.16 |
