1 Revision history for Perl extension Time::HiRes.
4 - Windows: ming32 patch from Mike Pomraning (use Perl's Const64()
5 instead of VC-specific i64 suffix)
8 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
11 - Windows: higher resolution time() by using the Windows
12 performance counter API, from Jan Dubois and Anton Shcherbinin.
13 The exact new higher resolution depends on the hardware,
14 but it should be quite a bit better than using the basic
18 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
19 the subtest 18. No known analysis nor fix, but added
20 an alarm (that requires fork() and alarm()) to the test.
23 - doc tweaks from mjd (perl change #20456)
24 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
27 - add a message (for non-core builds) to Makefile.PL about
28 the LC_ALL=C workaround
29 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
30 - the nanosleep() probe was broken (perl change #20061)
31 - use existence instead of definedness for feature probes
33 - MPE/iX tweak (perl change #20042)
34 - do not use HAS_NANOSLEEP (perl change #19898)
37 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
38 - OS/2 can always mix subsecond sleeps with signals
39 (part of perl change #19789)
42 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
46 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
47 - unnecessary (nanosleep is in libc anyway)
48 - harmful (-lrt slows down execution)
49 - incompatible (with many distributions' pthreads)
52 - do not create files in blib directories under core
53 (perl change #19160, from rgs)
54 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
57 - guarantee that $xdefine in HiRes.t is always defined
58 (perl change #19109, from IlyaZ)
59 - a cleaner way to detect PERL_CORE (perl change #19111,
63 - add hints/irix.pl to turn off overly POSIX flags that
64 cause hide struct timespec to be hidden (and compilation
65 to fail) (bleadperl change #19085)
66 - documentation tweaks
69 - add c:/temp to the list of temp directories to probe
70 so that cygwin (and win*?) builds are happy. This was
71 needed at least in my cygwin 1.3.20/w2k setup.
74 - modernize the constants code (from Nicholas Clark)
77 - At some point the ability to figure our the correct incdir
78 for EXTERN.h (either a core perl build, or an installed perl)
79 had broken (which lead into all test compiles failing with
80 a core perl build, but thanks to the robustness of Makefile.PL
81 nothing of this was visible). The brokenness seemed to be
82 caused by $ENV{PERL_CORE} not being on for core builds?
83 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
84 right, and both styles of build should work again.
87 - Nicholas Clark noticed that the my_catdir() emulation function
88 was broken (which means that we didn't really work for Perls
90 - inspired by fixing the above made the whole Makefile.PL -w
92 - tightened up the Makefile.PL output, less whitespace
95 - fix from Craig Berry for better building in VMS with PERL_CORE
98 - no functional changes
99 - move lib/Time/HiRes.pm as Hires.pm
100 - libraries scanning was slightly broken (always scanned
101 for a library even when $Config{libs} already had it)
104 - Ray Zimmerman ran into a race condition in Mac OS X.
105 A 0.01-second alarm fired before the test expected.
106 The test first slept indefinitely (blocking for signals)
107 and only after that tested for the signal having been sent.
108 Since the signal had already been sent, the test #12 never
109 completed. The solution: test first, then block.
110 - default to being silent on all probing attempts, set the
111 environment variable VERBOSE to a true value to see the
112 details (the probing command and the possible errors)
115 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
116 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
117 cannot be upgraded from CPAN to override the 5.8.0 version
119 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
120 because the include directories did not adjust themselves
121 if $ENV{PERL_CORE} (Hugo van der Sanden)
122 - add documentation about the restart of select() under alarm()
125 - small documentation tweaks
129 - better VMS operation (Craig Berry)
132 - our time machine is accelerating: now works with Perl 5.004_01
133 (tried with 5.003_07 and 5.002 but I get segmentation faults
134 from running the Makefile.PL with those in Tru64 4.0D)
137 - backward compatibility (pre-5.6.0) tweaks:
138 - no XSLoader in 5.00503, use DynaLoader instead
139 - no SvPV_nolen, either
140 - no PerlProc_pause(), either
141 - now tested with 5.00404 and 5.00503
142 - Makefile.PL requires 5.00404 (no more 5.002)
143 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
144 (this means that one can mix subsecond sleeps with alarms)
145 - because of nanosleep we probe for -lrt and -lposix4
146 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
147 is available by exportable constants Time::HiRes::d_func
148 (since older Perl do not have them in %Config, and even
149 5.8.0 does not probe for nanosleep)
152 - backward compatibility (pre-5.6.1) tweaks:
153 - define NV if no NVTYPE
154 - define IVdf if needed (note: the Devel::PPPort
155 in 5.8.0 does not try hard hard enough since
156 the IVSIZE might not be defined)
157 - define NVgf if needed
158 - grab the typemap from 5.8.0 for the NV stuff
161 - release 1.29_02 as 1.30
165 - fix a silly unclosed comment typo in HiRes.xs
166 - document and export REALTIME_REALPROF (Solaris)
170 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
171 (need to patch this also in Perl 5.[89])
172 - remove CVS revision log from HiRes.xs
176 The following numbered patches refer to the Perl 5.7 changes,
177 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
179 - 17558: Add #!./perl to the .t
180 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
181 - 16198: political correctness, from Simon Cozens
182 - 15857: doc tweaks, from Jarkko Hietaniemi
183 - 15593: optimization in .xs, from Paul Green
184 - 14892: pod fixes, from Robin Barker
185 - 14100: VOS fixes, from Paul Green
186 - 13422: XS segfault, from Marc Lehmann
187 - 13378: whether select() gets restarted on signals, depends
188 - 13354: timing constraints, again, from Andy Dougherty
189 - 13278: can't do subecond alarms with ualarm;
190 break out early if alarms do not seem to be working
191 - 13266: test relaxation (cygwin gets lower hires
192 times than lores ones)
193 - 12846: protect against high load, from Jarkko Hietaniemi
194 - 12837: HiRes.t VMS tweak, from Craig A. Berry
195 - 12797: HiRes.t VMS tweak, from Charles Lane
196 - 12769: HiRes.t VMS tweak, from Craig A. Berry
197 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
198 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
199 - 12692: alarm() ain't gonna work if ualarm() ain't,
200 from Gurusamy Sarathy
201 - 12680: minor VMS tweak, from Charles Lane
202 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
203 - 12609: croak on negative time, from Jarkko Hietaniemi
204 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
205 - 12594: MacOS Classic timeofday, from Chris Nandor
206 - 12473: allow for more than one second for sleep() and usleep()
207 - 12458: test tuning, relax timing constraints,
208 from Jarkko Hietaniemi
209 - 12449: make sleep() and usleep() to return the number
210 of seconds and microseconds actually slept (analogously
211 with the builtin sleep()), also make usleep() croak if
212 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
213 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
214 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
215 - 12196: use ftime() on Win32, from Artur Bergman
216 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
217 - 12105: use GetSystemTime() on Win32, from Artur Bergman
218 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
219 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
220 - 11797: problem in HiRes.t, from John P. Linderman
221 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
222 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
223 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
224 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
225 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
226 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
228 - 10942: MPE/IX test tweaks, from Mark Bixby
229 - 10784: unnecessary pod2man calls, from Andy Dougherty
230 - 10354: ext/ + -Wall, from Doug MacEachern
231 - 10320: fix the BOOT section to call myU2time correctly
232 - 10317: correct casting for AIX< from H. Merijn Brand
233 - 10119: document that the core time() may be rounding, not truncating
234 - 10118: test fix, from John Peacock
235 - 9988: long =item, from Robin Barker
236 - 9714: correct test output
237 - 9708: test also the scalar aspect of getitimer()
238 - 9705: Add interval timers (setitimer, getitimer)
239 - 9692: do not require at least 5.005 using XS
241 The following changes were made on top of the changes
242 made for Time::HiRes during the Perl 5.7 development
243 cycle that culminated in the release of Perl 5.8.0.
245 - add "require 5.005" to the Makefile.PL
246 - remove the REVISION section (CVS log) from HiRes.pm
247 - add jhi's copyright alongside Douglas'
248 - move HiRes.pm to lib/Time/
250 - modify HiRes.t to use $ENV{PERL_CORE}
251 - modify the original Time::HiRes version 1.20 Makefile.PL
252 to work both with Perl 5.8.0 and the new code with pre-5.8.0
253 Perls (tried with 5.6.1)
254 - tiny tweaks and updates in README and TODO
255 - bump the VERSION to 1.29
257 1.20 Wed Feb 24 21:30 1999
258 - make our usleep and ualarm substitutes into hrt_usleep
259 and hrt_ualarm. This helps static links of Perl with other
260 packages that also have usleep, etc. From
261 Ilya Zakharevich <ilya@math.ohio-state.edu>
262 - add C API stuff. From Joshua Pritikin
263 <joshua.pritikin@db.com>
264 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
265 - hopefully correct "-lc" fix for SCO.
268 1.19 Tue Sep 29 22:30 1998
269 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
270 <seb@stian.demon.co.uk>
271 - change GIMME_V to GIMME to help people with older versions of
273 - fix Win32 version of gettimeofday(). It didn't affect anything,
274 but it confuses people reading the code when the return value
275 is backwards (0 is success).
276 - fix Makefile.PL (more) so that detection of gettimeofday is
279 1.18 Mon Jul 6 22:40 1998
280 - add usleep() for Win32.
281 - fix Makefile.PL to fix reported HP/UX feature where unresolved
282 externals still cause an executable to be generated (though no
283 x bit set). Thanks to David Kozinn for report and explanation.
284 Problems with the fix are mine :)
286 1.17 Wed Jul 1 20:10 1998
287 - fix setitimer calls so microseconds is not more than 1000000.
288 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
289 - make Win32. We only get gettimeofday (the select hack doesn't
290 seem to work on my Win95 system).
291 - fix test 4 on 01test.t. add test to see if time() and
292 Time::HiRes::time() are close.
294 1.16 Wed Nov 12 21:05 1997
295 - add missing EXTEND in new gettimeofday scalar code.
297 1.15 Mon Nov 10 21:30 1997
298 - HiRes.pm: update pod. Provided by Gisle Aas.
299 - HiRes.xs: if gettimeofday() called in scalar context, do
300 something more useful than before. Provided by Gisle Aas.
301 - README: tell of xsubpp '-nolinenumber' woes. thanks to
302 Edward Henigin <ed@texas.net> for pointing out the problem.
304 1.14 Wed Nov 5 9:40 1997
305 - Makefile.PL: look for setitimer
306 - HiRes.xs: if missing ualarm, but we have setitimer, make up
307 our own setitimer. These were provided by Gisle Aas.
309 1.13 Tue Nov 4 23:30 1997
310 - Makefile.PL: fix autodetect mechanism to do try linking in addition
311 to just compiling; should fix Linux build problem. Fix was provided
314 1.12 Sun Oct 12 12:00:00 1997
315 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
316 you may need to comment this back out if you have an older xsubpp.
317 - HiRes.xs: set PROTOTYPES: DISABLE
319 1.11 Fri Sep 05 16:00:00 1997
321 Had some line commented out that shouldn't have been (testing
324 Previous version was corrupted.
326 1.10 Thu May 22 20:20:00 1997
327 - HiRes.xs, HiRes.pm, t/*:
328 - only compile what we have OS support for (or can
330 - only test what we compiled
331 - gross improvement to the test suite
333 This work was all done by Roderick Schertler
334 <roderick@argon.org>. If you run Linux or
335 one of the other ualarm-less platoforms, and you like this
336 module, let Roderick know; without him, it still wouldn't
337 be working on those boxes...
338 - Makefile.PL: figure out what routines the OS has and
339 only build what we need. These bits were written by Jarkko
340 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
342 1.02 Mon Dec 30 08:00:00 1996
343 - HiRes.pm: update documentation to say what to do when missing
344 ualarm() and friends.
345 - README: update to warn that ualarm() and friends need to exist
347 1.01 Fri Oct 17 08:00:00 1996
348 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
349 - HiRes.pm: put blank line between __END__ and =head1 so that
352 1.00 Tue Sep 03 13:00:00 1996
353 - original version; created by h2xs 1.16