1 Revision history for Perl extension Time::HiRes.
4 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
7 - Windows: higher resolution time() by using the Windows
8 performance counter API, from Jan Dubois and Anton Shcherbinin.
9 The exact new higher resolution depends on the hardware,
10 but it should be quite a bit better than using the basic
14 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
15 the subtest 18. No known analysis nor fix, but added
16 an alarm (that requires fork() and alarm()) to the test.
19 - doc tweaks from mjd (perl change #20456)
20 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
23 - add a message (for non-core builds) to Makefile.PL about
24 the LC_ALL=C workaround
25 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
26 - the nanosleep() probe was broken (perl change #20061)
27 - use existence instead of definedness for feature probes
29 - MPE/iX tweak (perl change #20042)
30 - do not use HAS_NANOSLEEP (perl change #19898)
33 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
34 - OS/2 can always mix subsecond sleeps with signals
35 (part of perl change #19789)
38 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
42 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
43 - unnecessary (nanosleep is in libc anyway)
44 - harmful (-lrt slows down execution)
45 - incompatible (with many distributions' pthreads)
48 - do not create files in blib directories under core
49 (perl change #19160, from rgs)
50 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
53 - guarantee that $xdefine in HiRes.t is always defined
54 (perl change #19109, from IlyaZ)
55 - a cleaner way to detect PERL_CORE (perl change #19111,
59 - add hints/irix.pl to turn off overly POSIX flags that
60 cause hide struct timespec to be hidden (and compilation
61 to fail) (bleadperl change #19085)
62 - documentation tweaks
65 - add c:/temp to the list of temp directories to probe
66 so that cygwin (and win*?) builds are happy. This was
67 needed at least in my cygwin 1.3.20/w2k setup.
70 - modernize the constants code (from Nicholas Clark)
73 - At some point the ability to figure our the correct incdir
74 for EXTERN.h (either a core perl build, or an installed perl)
75 had broken (which lead into all test compiles failing with
76 a core perl build, but thanks to the robustness of Makefile.PL
77 nothing of this was visible). The brokenness seemed to be
78 caused by $ENV{PERL_CORE} not being on for core builds?
79 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
80 right, and both styles of build should work again.
83 - Nicholas Clark noticed that the my_catdir() emulation function
84 was broken (which means that we didn't really work for Perls
86 - inspired by fixing the above made the whole Makefile.PL -w
88 - tightened up the Makefile.PL output, less whitespace
91 - fix from Craig Berry for better building in VMS with PERL_CORE
94 - no functional changes
95 - move lib/Time/HiRes.pm as Hires.pm
96 - libraries scanning was slightly broken (always scanned
97 for a library even when $Config{libs} already had it)
100 - Ray Zimmerman ran into a race condition in Mac OS X.
101 A 0.01-second alarm fired before the test expected.
102 The test first slept indefinitely (blocking for signals)
103 and only after that tested for the signal having been sent.
104 Since the signal had already been sent, the test #12 never
105 completed. The solution: test first, then block.
106 - default to being silent on all probing attempts, set the
107 environment variable VERBOSE to a true value to see the
108 details (the probing command and the possible errors)
111 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
112 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
113 cannot be upgraded from CPAN to override the 5.8.0 version
115 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
116 because the include directories did not adjust themselves
117 if $ENV{PERL_CORE} (Hugo van der Sanden)
118 - add documentation about the restart of select() under alarm()
121 - small documentation tweaks
125 - better VMS operation (Craig Berry)
128 - our time machine is accelerating: now works with Perl 5.004_01
129 (tried with 5.003_07 and 5.002 but I get segmentation faults
130 from running the Makefile.PL with those in Tru64 4.0D)
133 - backward compatibility (pre-5.6.0) tweaks:
134 - no XSLoader in 5.00503, use DynaLoader instead
135 - no SvPV_nolen, either
136 - no PerlProc_pause(), either
137 - now tested with 5.00404 and 5.00503
138 - Makefile.PL requires 5.00404 (no more 5.002)
139 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
140 (this means that one can mix subsecond sleeps with alarms)
141 - because of nanosleep we probe for -lrt and -lposix4
142 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
143 is available by exportable constants Time::HiRes::d_func
144 (since older Perl do not have them in %Config, and even
145 5.8.0 does not probe for nanosleep)
148 - backward compatibility (pre-5.6.1) tweaks:
149 - define NV if no NVTYPE
150 - define IVdf if needed (note: the Devel::PPPort
151 in 5.8.0 does not try hard hard enough since
152 the IVSIZE might not be defined)
153 - define NVgf if needed
154 - grab the typemap from 5.8.0 for the NV stuff
157 - release 1.29_02 as 1.30
161 - fix a silly unclosed comment typo in HiRes.xs
162 - document and export REALTIME_REALPROF (Solaris)
166 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
167 (need to patch this also in Perl 5.[89])
168 - remove CVS revision log from HiRes.xs
172 The following numbered patches refer to the Perl 5.7 changes,
173 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
175 - 17558: Add #!./perl to the .t
176 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
177 - 16198: political correctness, from Simon Cozens
178 - 15857: doc tweaks, from Jarkko Hietaniemi
179 - 15593: optimization in .xs, from Paul Green
180 - 14892: pod fixes, from Robin Barker
181 - 14100: VOS fixes, from Paul Green
182 - 13422: XS segfault, from Marc Lehmann
183 - 13378: whether select() gets restarted on signals, depends
184 - 13354: timing constraints, again, from Andy Dougherty
185 - 13278: can't do subecond alarms with ualarm;
186 break out early if alarms do not seem to be working
187 - 13266: test relaxation (cygwin gets lower hires
188 times than lores ones)
189 - 12846: protect against high load, from Jarkko Hietaniemi
190 - 12837: HiRes.t VMS tweak, from Craig A. Berry
191 - 12797: HiRes.t VMS tweak, from Charles Lane
192 - 12769: HiRes.t VMS tweak, from Craig A. Berry
193 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
194 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
195 - 12692: alarm() ain't gonna work if ualarm() ain't,
196 from Gurusamy Sarathy
197 - 12680: minor VMS tweak, from Charles Lane
198 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
199 - 12609: croak on negative time, from Jarkko Hietaniemi
200 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
201 - 12594: MacOS Classic timeofday, from Chris Nandor
202 - 12473: allow for more than one second for sleep() and usleep()
203 - 12458: test tuning, relax timing constraints,
204 from Jarkko Hietaniemi
205 - 12449: make sleep() and usleep() to return the number
206 of seconds and microseconds actually slept (analogously
207 with the builtin sleep()), also make usleep() croak if
208 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
209 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
210 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
211 - 12196: use ftime() on Win32, from Artur Bergman
212 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
213 - 12105: use GetSystemTime() on Win32, from Artur Bergman
214 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
215 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
216 - 11797: problem in HiRes.t, from John P. Linderman
217 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
218 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
219 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
220 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
221 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
222 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
224 - 10942: MPE/IX test tweaks, from Mark Bixby
225 - 10784: unnecessary pod2man calls, from Andy Dougherty
226 - 10354: ext/ + -Wall, from Doug MacEachern
227 - 10320: fix the BOOT section to call myU2time correctly
228 - 10317: correct casting for AIX< from H. Merijn Brand
229 - 10119: document that the core time() may be rounding, not truncating
230 - 10118: test fix, from John Peacock
231 - 9988: long =item, from Robin Barker
232 - 9714: correct test output
233 - 9708: test also the scalar aspect of getitimer()
234 - 9705: Add interval timers (setitimer, getitimer)
235 - 9692: do not require at least 5.005 using XS
237 The following changes were made on top of the changes
238 made for Time::HiRes during the Perl 5.7 development
239 cycle that culminated in the release of Perl 5.8.0.
241 - add "require 5.005" to the Makefile.PL
242 - remove the REVISION section (CVS log) from HiRes.pm
243 - add jhi's copyright alongside Douglas'
244 - move HiRes.pm to lib/Time/
246 - modify HiRes.t to use $ENV{PERL_CORE}
247 - modify the original Time::HiRes version 1.20 Makefile.PL
248 to work both with Perl 5.8.0 and the new code with pre-5.8.0
249 Perls (tried with 5.6.1)
250 - tiny tweaks and updates in README and TODO
251 - bump the VERSION to 1.29
253 1.20 Wed Feb 24 21:30 1999
254 - make our usleep and ualarm substitutes into hrt_usleep
255 and hrt_ualarm. This helps static links of Perl with other
256 packages that also have usleep, etc. From
257 Ilya Zakharevich <ilya@math.ohio-state.edu>
258 - add C API stuff. From Joshua Pritikin
259 <joshua.pritikin@db.com>
260 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
261 - hopefully correct "-lc" fix for SCO.
264 1.19 Tue Sep 29 22:30 1998
265 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
266 <seb@stian.demon.co.uk>
267 - change GIMME_V to GIMME to help people with older versions of
269 - fix Win32 version of gettimeofday(). It didn't affect anything,
270 but it confuses people reading the code when the return value
271 is backwards (0 is success).
272 - fix Makefile.PL (more) so that detection of gettimeofday is
275 1.18 Mon Jul 6 22:40 1998
276 - add usleep() for Win32.
277 - fix Makefile.PL to fix reported HP/UX feature where unresolved
278 externals still cause an executable to be generated (though no
279 x bit set). Thanks to David Kozinn for report and explanation.
280 Problems with the fix are mine :)
282 1.17 Wed Jul 1 20:10 1998
283 - fix setitimer calls so microseconds is not more than 1000000.
284 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
285 - make Win32. We only get gettimeofday (the select hack doesn't
286 seem to work on my Win95 system).
287 - fix test 4 on 01test.t. add test to see if time() and
288 Time::HiRes::time() are close.
290 1.16 Wed Nov 12 21:05 1997
291 - add missing EXTEND in new gettimeofday scalar code.
293 1.15 Mon Nov 10 21:30 1997
294 - HiRes.pm: update pod. Provided by Gisle Aas.
295 - HiRes.xs: if gettimeofday() called in scalar context, do
296 something more useful than before. Provided by Gisle Aas.
297 - README: tell of xsubpp '-nolinenumber' woes. thanks to
298 Edward Henigin <ed@texas.net> for pointing out the problem.
300 1.14 Wed Nov 5 9:40 1997
301 - Makefile.PL: look for setitimer
302 - HiRes.xs: if missing ualarm, but we have setitimer, make up
303 our own setitimer. These were provided by Gisle Aas.
305 1.13 Tue Nov 4 23:30 1997
306 - Makefile.PL: fix autodetect mechanism to do try linking in addition
307 to just compiling; should fix Linux build problem. Fix was provided
310 1.12 Sun Oct 12 12:00:00 1997
311 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
312 you may need to comment this back out if you have an older xsubpp.
313 - HiRes.xs: set PROTOTYPES: DISABLE
315 1.11 Fri Sep 05 16:00:00 1997
317 Had some line commented out that shouldn't have been (testing
320 Previous version was corrupted.
322 1.10 Thu May 22 20:20:00 1997
323 - HiRes.xs, HiRes.pm, t/*:
324 - only compile what we have OS support for (or can
326 - only test what we compiled
327 - gross improvement to the test suite
329 This work was all done by Roderick Schertler
330 <roderick@argon.org>. If you run Linux or
331 one of the other ualarm-less platoforms, and you like this
332 module, let Roderick know; without him, it still wouldn't
333 be working on those boxes...
334 - Makefile.PL: figure out what routines the OS has and
335 only build what we need. These bits were written by Jarkko
336 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
338 1.02 Mon Dec 30 08:00:00 1996
339 - HiRes.pm: update documentation to say what to do when missing
340 ualarm() and friends.
341 - README: update to warn that ualarm() and friends need to exist
343 1.01 Fri Oct 17 08:00:00 1996
344 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
345 - HiRes.pm: put blank line between __END__ and =head1 so that
348 1.00 Tue Sep 03 13:00:00 1996
349 - original version; created by h2xs 1.16