1 Revision history for Perl extension Time::HiRes.
4 - Give a clearer message if the tests timeout (perl change #22253)
5 - Don't use /tmp or its moral equivalents (perl bug #15036,
9 - Windows: ming32 patch from Mike Pomraning (use Perl's Const64()
10 instead of VC-specific i64 suffix)
13 - Solaris: like Tru64 (dec_osf) also Solaris need -lrt for nanosleep
16 - Windows: higher resolution time() by using the Windows
17 performance counter API, from Jan Dubois and Anton Shcherbinin.
18 The exact new higher resolution depends on the hardware,
19 but it should be quite a bit better than using the basic
23 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
24 the subtest 18. No known analysis nor fix, but added
25 an alarm (that requires fork() and alarm()) to the test.
28 - doc tweaks from mjd (perl change #20456)
29 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
32 - add a message (for non-core builds) to Makefile.PL about
33 the LC_ALL=C workaround
34 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
35 - the nanosleep() probe was broken (perl change #20061)
36 - use existence instead of definedness for feature probes
38 - MPE/iX tweak (perl change #20042)
39 - do not use HAS_NANOSLEEP (perl change #19898)
42 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
43 - OS/2 can always mix subsecond sleeps with signals
44 (part of perl change #19789)
47 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
51 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
52 - unnecessary (nanosleep is in libc anyway)
53 - harmful (-lrt slows down execution)
54 - incompatible (with many distributions' pthreads)
57 - do not create files in blib directories under core
58 (perl change #19160, from rgs)
59 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
62 - guarantee that $xdefine in HiRes.t is always defined
63 (perl change #19109, from IlyaZ)
64 - a cleaner way to detect PERL_CORE (perl change #19111,
68 - add hints/irix.pl to turn off overly POSIX flags that
69 cause hide struct timespec to be hidden (and compilation
70 to fail) (bleadperl change #19085)
71 - documentation tweaks
74 - add c:/temp to the list of temp directories to probe
75 so that cygwin (and win*?) builds are happy. This was
76 needed at least in my cygwin 1.3.20/w2k setup.
79 - modernize the constants code (from Nicholas Clark)
82 - At some point the ability to figure our the correct incdir
83 for EXTERN.h (either a core perl build, or an installed perl)
84 had broken (which lead into all test compiles failing with
85 a core perl build, but thanks to the robustness of Makefile.PL
86 nothing of this was visible). The brokenness seemed to be
87 caused by $ENV{PERL_CORE} not being on for core builds?
88 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
89 right, and both styles of build should work again.
92 - Nicholas Clark noticed that the my_catdir() emulation function
93 was broken (which means that we didn't really work for Perls
95 - inspired by fixing the above made the whole Makefile.PL -w
97 - tightened up the Makefile.PL output, less whitespace
100 - fix from Craig Berry for better building in VMS with PERL_CORE
103 - no functional changes
104 - move lib/Time/HiRes.pm as Hires.pm
105 - libraries scanning was slightly broken (always scanned
106 for a library even when $Config{libs} already had it)
109 - Ray Zimmerman ran into a race condition in Mac OS X.
110 A 0.01-second alarm fired before the test expected.
111 The test first slept indefinitely (blocking for signals)
112 and only after that tested for the signal having been sent.
113 Since the signal had already been sent, the test #12 never
114 completed. The solution: test first, then block.
115 - default to being silent on all probing attempts, set the
116 environment variable VERBOSE to a true value to see the
117 details (the probing command and the possible errors)
120 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
121 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
122 cannot be upgraded from CPAN to override the 5.8.0 version
124 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
125 because the include directories did not adjust themselves
126 if $ENV{PERL_CORE} (Hugo van der Sanden)
127 - add documentation about the restart of select() under alarm()
130 - small documentation tweaks
134 - better VMS operation (Craig Berry)
137 - our time machine is accelerating: now works with Perl 5.004_01
138 (tried with 5.003_07 and 5.002 but I get segmentation faults
139 from running the Makefile.PL with those in Tru64 4.0D)
142 - backward compatibility (pre-5.6.0) tweaks:
143 - no XSLoader in 5.00503, use DynaLoader instead
144 - no SvPV_nolen, either
145 - no PerlProc_pause(), either
146 - now tested with 5.00404 and 5.00503
147 - Makefile.PL requires 5.00404 (no more 5.002)
148 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
149 (this means that one can mix subsecond sleeps with alarms)
150 - because of nanosleep we probe for -lrt and -lposix4
151 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
152 is available by exportable constants Time::HiRes::d_func
153 (since older Perl do not have them in %Config, and even
154 5.8.0 does not probe for nanosleep)
157 - backward compatibility (pre-5.6.1) tweaks:
158 - define NV if no NVTYPE
159 - define IVdf if needed (note: the Devel::PPPort
160 in 5.8.0 does not try hard hard enough since
161 the IVSIZE might not be defined)
162 - define NVgf if needed
163 - grab the typemap from 5.8.0 for the NV stuff
166 - release 1.29_02 as 1.30
170 - fix a silly unclosed comment typo in HiRes.xs
171 - document and export REALTIME_REALPROF (Solaris)
175 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
176 (need to patch this also in Perl 5.[89])
177 - remove CVS revision log from HiRes.xs
181 The following numbered patches refer to the Perl 5.7 changes,
182 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
184 - 17558: Add #!./perl to the .t
185 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
186 - 16198: political correctness, from Simon Cozens
187 - 15857: doc tweaks, from Jarkko Hietaniemi
188 - 15593: optimization in .xs, from Paul Green
189 - 14892: pod fixes, from Robin Barker
190 - 14100: VOS fixes, from Paul Green
191 - 13422: XS segfault, from Marc Lehmann
192 - 13378: whether select() gets restarted on signals, depends
193 - 13354: timing constraints, again, from Andy Dougherty
194 - 13278: can't do subecond alarms with ualarm;
195 break out early if alarms do not seem to be working
196 - 13266: test relaxation (cygwin gets lower hires
197 times than lores ones)
198 - 12846: protect against high load, from Jarkko Hietaniemi
199 - 12837: HiRes.t VMS tweak, from Craig A. Berry
200 - 12797: HiRes.t VMS tweak, from Charles Lane
201 - 12769: HiRes.t VMS tweak, from Craig A. Berry
202 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
203 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
204 - 12692: alarm() ain't gonna work if ualarm() ain't,
205 from Gurusamy Sarathy
206 - 12680: minor VMS tweak, from Charles Lane
207 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
208 - 12609: croak on negative time, from Jarkko Hietaniemi
209 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
210 - 12594: MacOS Classic timeofday, from Chris Nandor
211 - 12473: allow for more than one second for sleep() and usleep()
212 - 12458: test tuning, relax timing constraints,
213 from Jarkko Hietaniemi
214 - 12449: make sleep() and usleep() to return the number
215 of seconds and microseconds actually slept (analogously
216 with the builtin sleep()), also make usleep() croak if
217 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
218 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
219 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
220 - 12196: use ftime() on Win32, from Artur Bergman
221 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
222 - 12105: use GetSystemTime() on Win32, from Artur Bergman
223 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
224 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
225 - 11797: problem in HiRes.t, from John P. Linderman
226 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
227 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
228 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
229 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
230 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
231 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
233 - 10942: MPE/IX test tweaks, from Mark Bixby
234 - 10784: unnecessary pod2man calls, from Andy Dougherty
235 - 10354: ext/ + -Wall, from Doug MacEachern
236 - 10320: fix the BOOT section to call myU2time correctly
237 - 10317: correct casting for AIX< from H. Merijn Brand
238 - 10119: document that the core time() may be rounding, not truncating
239 - 10118: test fix, from John Peacock
240 - 9988: long =item, from Robin Barker
241 - 9714: correct test output
242 - 9708: test also the scalar aspect of getitimer()
243 - 9705: Add interval timers (setitimer, getitimer)
244 - 9692: do not require at least 5.005 using XS
246 The following changes were made on top of the changes
247 made for Time::HiRes during the Perl 5.7 development
248 cycle that culminated in the release of Perl 5.8.0.
250 - add "require 5.005" to the Makefile.PL
251 - remove the REVISION section (CVS log) from HiRes.pm
252 - add jhi's copyright alongside Douglas'
253 - move HiRes.pm to lib/Time/
255 - modify HiRes.t to use $ENV{PERL_CORE}
256 - modify the original Time::HiRes version 1.20 Makefile.PL
257 to work both with Perl 5.8.0 and the new code with pre-5.8.0
258 Perls (tried with 5.6.1)
259 - tiny tweaks and updates in README and TODO
260 - bump the VERSION to 1.29
262 1.20 Wed Feb 24 21:30 1999
263 - make our usleep and ualarm substitutes into hrt_usleep
264 and hrt_ualarm. This helps static links of Perl with other
265 packages that also have usleep, etc. From
266 Ilya Zakharevich <ilya@math.ohio-state.edu>
267 - add C API stuff. From Joshua Pritikin
268 <joshua.pritikin@db.com>
269 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
270 - hopefully correct "-lc" fix for SCO.
273 1.19 Tue Sep 29 22:30 1998
274 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
275 <seb@stian.demon.co.uk>
276 - change GIMME_V to GIMME to help people with older versions of
278 - fix Win32 version of gettimeofday(). It didn't affect anything,
279 but it confuses people reading the code when the return value
280 is backwards (0 is success).
281 - fix Makefile.PL (more) so that detection of gettimeofday is
284 1.18 Mon Jul 6 22:40 1998
285 - add usleep() for Win32.
286 - fix Makefile.PL to fix reported HP/UX feature where unresolved
287 externals still cause an executable to be generated (though no
288 x bit set). Thanks to David Kozinn for report and explanation.
289 Problems with the fix are mine :)
291 1.17 Wed Jul 1 20:10 1998
292 - fix setitimer calls so microseconds is not more than 1000000.
293 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
294 - make Win32. We only get gettimeofday (the select hack doesn't
295 seem to work on my Win95 system).
296 - fix test 4 on 01test.t. add test to see if time() and
297 Time::HiRes::time() are close.
299 1.16 Wed Nov 12 21:05 1997
300 - add missing EXTEND in new gettimeofday scalar code.
302 1.15 Mon Nov 10 21:30 1997
303 - HiRes.pm: update pod. Provided by Gisle Aas.
304 - HiRes.xs: if gettimeofday() called in scalar context, do
305 something more useful than before. Provided by Gisle Aas.
306 - README: tell of xsubpp '-nolinenumber' woes. thanks to
307 Edward Henigin <ed@texas.net> for pointing out the problem.
309 1.14 Wed Nov 5 9:40 1997
310 - Makefile.PL: look for setitimer
311 - HiRes.xs: if missing ualarm, but we have setitimer, make up
312 our own setitimer. These were provided by Gisle Aas.
314 1.13 Tue Nov 4 23:30 1997
315 - Makefile.PL: fix autodetect mechanism to do try linking in addition
316 to just compiling; should fix Linux build problem. Fix was provided
319 1.12 Sun Oct 12 12:00:00 1997
320 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
321 you may need to comment this back out if you have an older xsubpp.
322 - HiRes.xs: set PROTOTYPES: DISABLE
324 1.11 Fri Sep 05 16:00:00 1997
326 Had some line commented out that shouldn't have been (testing
329 Previous version was corrupted.
331 1.10 Thu May 22 20:20:00 1997
332 - HiRes.xs, HiRes.pm, t/*:
333 - only compile what we have OS support for (or can
335 - only test what we compiled
336 - gross improvement to the test suite
338 This work was all done by Roderick Schertler
339 <roderick@argon.org>. If you run Linux or
340 one of the other ualarm-less platoforms, and you like this
341 module, let Roderick know; without him, it still wouldn't
342 be working on those boxes...
343 - Makefile.PL: figure out what routines the OS has and
344 only build what we need. These bits were written by Jarkko
345 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
347 1.02 Mon Dec 30 08:00:00 1996
348 - HiRes.pm: update documentation to say what to do when missing
349 ualarm() and friends.
350 - README: update to warn that ualarm() and friends need to exist
352 1.01 Fri Oct 17 08:00:00 1996
353 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
354 - HiRes.pm: put blank line between __END__ and =head1 so that
357 1.00 Tue Sep 03 13:00:00 1996
358 - original version; created by h2xs 1.16