1 Revision history for Perl extension Time::HiRes.
4 - doc tweaks from mjd (perl change #20456)
5 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
8 - add a message (for non-core builds) to Makefile.PL about
9 the LC_ALL=C workaround
10 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
11 - the nanosleep() probe was broken (perl change #20061)
12 - use existence instead of definedness for feature probes
14 - MPE/iX tweak (perl change #20042)
15 - do not use HAS_NANOSLEEP (perl change #19898)
18 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
19 - OS/2 can always mix subsecond sleeps with signals
20 (part of perl change #19789)
23 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
27 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
28 - unnecessary (nanosleep is in libc anyway)
29 - harmful (-lrt slows down execution)
30 - incompatible (with many distributions' pthreads)
33 - do not create files in blib directories under core
34 (perl change #19160, from rgs)
35 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
38 - guarantee that $xdefine in HiRes.t is always defined
39 (perl change #19109, from IlyaZ)
40 - a cleaner way to detect PERL_CORE (perl change #19111,
44 - add hints/irix.pl to turn off overly POSIX flags that
45 cause hide struct timespec to be hidden (and compilation
46 to fail) (bleadperl change #19085)
47 - documentation tweaks
50 - add c:/temp to the list of temp directories to probe
51 so that cygwin (and win*?) builds are happy. This was
52 needed at least in my cygwin 1.3.20/w2k setup.
55 - modernize the constants code (from Nicholas Clark)
58 - At some point the ability to figure our the correct incdir
59 for EXTERN.h (either a core perl build, or an installed perl)
60 had broken (which lead into all test compiles failing with
61 a core perl build, but thanks to the robustness of Makefile.PL
62 nothing of this was visible). The brokenness seemed to be
63 caused by $ENV{PERL_CORE} not being on for core builds?
64 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
65 right, and both styles of build should work again.
68 - Nicholas Clark noticed that the my_catdir() emulation function
69 was broken (which means that we didn't really work for Perls
71 - inspired by fixing the above made the whole Makefile.PL -w
73 - tightened up the Makefile.PL output, less whitespace
76 - fix from Craig Berry for better building in VMS with PERL_CORE
79 - no functional changes
80 - move lib/Time/HiRes.pm as Hires.pm
81 - libraries scanning was slightly broken (always scanned
82 for a library even when $Config{libs} already had it)
85 - Ray Zimmerman ran into a race condition in Mac OS X.
86 A 0.01-second alarm fired before the test expected.
87 The test first slept indefinitely (blocking for signals)
88 and only after that tested for the signal having been sent.
89 Since the signal had already been sent, the test #12 never
90 completed. The solution: test first, then block.
91 - default to being silent on all probing attempts, set the
92 environment variable VERBOSE to a true value to see the
93 details (the probing command and the possible errors)
96 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
97 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
98 cannot be upgraded from CPAN to override the 5.8.0 version
100 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
101 because the include directories did not adjust themselves
102 if $ENV{PERL_CORE} (Hugo van der Sanden)
103 - add documentation about the restart of select() under alarm()
106 - small documentation tweaks
110 - better VMS operation (Craig Berry)
113 - our time machine is accelerating: now works with Perl 5.004_01
114 (tried with 5.003_07 and 5.002 but I get segmentation faults
115 from running the Makefile.PL with those in Tru64 4.0D)
118 - backward compatibility (pre-5.6.0) tweaks:
119 - no XSLoader in 5.00503, use DynaLoader instead
120 - no SvPV_nolen, either
121 - no PerlProc_pause(), either
122 - now tested with 5.00404 and 5.00503
123 - Makefile.PL requires 5.00404 (no more 5.002)
124 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
125 (this means that one can mix subsecond sleeps with alarms)
126 - because of nanosleep we probe for -lrt and -lposix4
127 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
128 is available by exportable constants Time::HiRes::d_func
129 (since older Perl do not have them in %Config, and even
130 5.8.0 does not probe for nanosleep)
133 - backward compatibility (pre-5.6.1) tweaks:
134 - define NV if no NVTYPE
135 - define IVdf if needed (note: the Devel::PPPort
136 in 5.8.0 does not try hard hard enough since
137 the IVSIZE might not be defined)
138 - define NVgf if needed
139 - grab the typemap from 5.8.0 for the NV stuff
142 - release 1.29_02 as 1.30
146 - fix a silly unclosed comment typo in HiRes.xs
147 - document and export REALTIME_REALPROF (Solaris)
151 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
152 (need to patch this also in Perl 5.[89])
153 - remove CVS revision log from HiRes.xs
157 The following numbered patches refer to the Perl 5.7 changes,
158 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
160 - 17558: Add #!./perl to the .t
161 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
162 - 16198: political correctness, from Simon Cozens
163 - 15857: doc tweaks, from Jarkko Hietaniemi
164 - 15593: optimization in .xs, from Paul Green
165 - 14892: pod fixes, from Robin Barker
166 - 14100: VOS fixes, from Paul Green
167 - 13422: XS segfault, from Marc Lehmann
168 - 13378: whether select() gets restarted on signals, depends
169 - 13354: timing constraints, again, from Andy Dougherty
170 - 13278: can't do subecond alarms with ualarm;
171 break out early if alarms do not seem to be working
172 - 13266: test relaxation (cygwin gets lower hires
173 times than lores ones)
174 - 12846: protect against high load, from Jarkko Hietaniemi
175 - 12837: HiRes.t VMS tweak, from Craig A. Berry
176 - 12797: HiRes.t VMS tweak, from Charles Lane
177 - 12769: HiRes.t VMS tweak, from Craig A. Berry
178 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
179 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
180 - 12692: alarm() ain't gonna work if ualarm() ain't,
181 from Gurusamy Sarathy
182 - 12680: minor VMS tweak, from Charles Lane
183 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
184 - 12609: croak on negative time, from Jarkko Hietaniemi
185 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
186 - 12594: MacOS Classic timeofday, from Chris Nandor
187 - 12473: allow for more than one second for sleep() and usleep()
188 - 12458: test tuning, relax timing constraints,
189 from Jarkko Hietaniemi
190 - 12449: make sleep() and usleep() to return the number
191 of seconds and microseconds actually slept (analogously
192 with the builtin sleep()), also make usleep() croak if
193 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
194 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
195 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
196 - 12196: use ftime() on Win32, from Artur Bergman
197 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
198 - 12105: use GetSystemTime() on Win32, from Artur Bergman
199 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
200 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
201 - 11797: problem in HiRes.t, from John P. Linderman
202 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
203 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
204 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
205 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
206 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
207 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
209 - 10942: MPE/IX test tweaks, from Mark Bixby
210 - 10784: unnecessary pod2man calls, from Andy Dougherty
211 - 10354: ext/ + -Wall, from Doug MacEachern
212 - 10320: fix the BOOT section to call myU2time correctly
213 - 10317: correct casting for AIX< from H. Merijn Brand
214 - 10119: document that the core time() may be rounding, not truncating
215 - 10118: test fix, from John Peacock
216 - 9988: long =item, from Robin Barker
217 - 9714: correct test output
218 - 9708: test also the scalar aspect of getitimer()
219 - 9705: Add interval timers (setitimer, getitimer)
220 - 9692: do not require at least 5.005 using XS
222 The following changes were made on top of the changes
223 made for Time::HiRes during the Perl 5.7 development
224 cycle that culminated in the release of Perl 5.8.0.
226 - add "require 5.005" to the Makefile.PL
227 - remove the REVISION section (CVS log) from HiRes.pm
228 - add jhi's copyright alongside Douglas'
229 - move HiRes.pm to lib/Time/
231 - modify HiRes.t to use $ENV{PERL_CORE}
232 - modify the original Time::HiRes version 1.20 Makefile.PL
233 to work both with Perl 5.8.0 and the new code with pre-5.8.0
234 Perls (tried with 5.6.1)
235 - tiny tweaks and updates in README and TODO
236 - bump the VERSION to 1.29
238 1.20 Wed Feb 24 21:30 1999
239 - make our usleep and ualarm substitutes into hrt_usleep
240 and hrt_ualarm. This helps static links of Perl with other
241 packages that also have usleep, etc. From
242 Ilya Zakharevich <ilya@math.ohio-state.edu>
243 - add C API stuff. From Joshua Pritikin
244 <joshua.pritikin@db.com>
245 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
246 - hopefully correct "-lc" fix for SCO.
249 1.19 Tue Sep 29 22:30 1998
250 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
251 <seb@stian.demon.co.uk>
252 - change GIMME_V to GIMME to help people with older versions of
254 - fix Win32 version of gettimeofday(). It didn't affect anything,
255 but it confuses people reading the code when the return value
256 is backwards (0 is success).
257 - fix Makefile.PL (more) so that detection of gettimeofday is
260 1.18 Mon Jul 6 22:40 1998
261 - add usleep() for Win32.
262 - fix Makefile.PL to fix reported HP/UX feature where unresolved
263 externals still cause an executable to be generated (though no
264 x bit set). Thanks to David Kozinn for report and explanation.
265 Problems with the fix are mine :)
267 1.17 Wed Jul 1 20:10 1998
268 - fix setitimer calls so microseconds is not more than 1000000.
269 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
270 - make Win32. We only get gettimeofday (the select hack doesn't
271 seem to work on my Win95 system).
272 - fix test 4 on 01test.t. add test to see if time() and
273 Time::HiRes::time() are close.
275 1.16 Wed Nov 12 21:05 1997
276 - add missing EXTEND in new gettimeofday scalar code.
278 1.15 Mon Nov 10 21:30 1997
279 - HiRes.pm: update pod. Provided by Gisle Aas.
280 - HiRes.xs: if gettimeofday() called in scalar context, do
281 something more useful than before. Provided by Gisle Aas.
282 - README: tell of xsubpp '-nolinenumber' woes. thanks to
283 Edward Henigin <ed@texas.net> for pointing out the problem.
285 1.14 Wed Nov 5 9:40 1997
286 - Makefile.PL: look for setitimer
287 - HiRes.xs: if missing ualarm, but we have setitimer, make up
288 our own setitimer. These were provided by Gisle Aas.
290 1.13 Tue Nov 4 23:30 1997
291 - Makefile.PL: fix autodetect mechanism to do try linking in addition
292 to just compiling; should fix Linux build problem. Fix was provided
295 1.12 Sun Oct 12 12:00:00 1997
296 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
297 you may need to comment this back out if you have an older xsubpp.
298 - HiRes.xs: set PROTOTYPES: DISABLE
300 1.11 Fri Sep 05 16:00:00 1997
302 Had some line commented out that shouldn't have been (testing
305 Previous version was corrupted.
307 1.10 Thu May 22 20:20:00 1997
308 - HiRes.xs, HiRes.pm, t/*:
309 - only compile what we have OS support for (or can
311 - only test what we compiled
312 - gross improvement to the test suite
314 This work was all done by Roderick Schertler
315 <roderick@argon.org>. If you run Linux or
316 one of the other ualarm-less platoforms, and you like this
317 module, let Roderick know; without him, it still wouldn't
318 be working on those boxes...
319 - Makefile.PL: figure out what routines the OS has and
320 only build what we need. These bits were written by Jarkko
321 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
323 1.02 Mon Dec 30 08:00:00 1996
324 - HiRes.pm: update documentation to say what to do when missing
325 ualarm() and friends.
326 - README: update to warn that ualarm() and friends need to exist
328 1.01 Fri Oct 17 08:00:00 1996
329 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
330 - HiRes.pm: put blank line between __END__ and =head1 so that
333 1.00 Tue Sep 03 13:00:00 1996
334 - original version; created by h2xs 1.16