1 Revision history for Perl extension Time::HiRes.
4 - do not create files in blib directories under core
5 (perl change #19160, from rgs)
6 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
9 - guarantee that $xdefine in HiRes.t is always defined
10 (perl change #19109, from IlyaZ)
11 - a cleaner way to detect PERL_CORE (perl change #19111,
15 - add hints/irix.pl to turn off overly POSIX flags that
16 cause hide struct timespec to be hidden (and compilation
17 to fail) (bleadperl change #19085)
18 - documentation tweaks
21 - add c:/temp to the list of temp directories to probe
22 so that cygwin (and win*?) builds are happy. This was
23 needed at least in my cygwin 1.3.20/w2k setup.
26 - modernize the constants code (from Nicholas Clark)
29 - At some point the ability to figure our the correct incdir
30 for EXTERN.h (either a core perl build, or an installed perl)
31 had broken (which lead into all test compiles failing with
32 a core perl build, but thanks to the robustness of Makefile.PL
33 nothing of this was visible). The brokenness seemed to be
34 caused by $ENV{PERL_CORE} not being on for core builds?
35 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
36 right, and both styles of build should work again.
39 - Nicholas Clark noticed that the my_catdir() emulation function
40 was broken (which means that we didn't really work for Perls
42 - inspired by fixing the above made the whole Makefile.PL -w
44 - tightened up the Makefile.PL output, less whitespace
47 - fix from Craig Berry for better building in VMS with PERL_CORE
50 - no functional changes
51 - move lib/Time/HiRes.pm as Hires.pm
52 - libraries scanning was slightly broken (always scanned
53 for a library even when $Config{libs} already had it)
56 - Ray Zimmerman ran into a race condition in Mac OS X.
57 A 0.01-second alarm fired before the test expected.
58 The test first slept indefinitely (blocking for signals)
59 and only after that tested for the signal having been sent.
60 Since the signal had already been sent, the test #12 never
61 completed. The solution: test first, then block.
62 - default to being silent on all probing attempts, set the
63 environment variable VERBOSE to a true value to see the
64 details (the probing command and the possible errors)
67 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
68 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
69 cannot be upgraded from CPAN to override the 5.8.0 version
71 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
72 because the include directories did not adjust themselves
73 if $ENV{PERL_CORE} (Hugo van der Sanden)
74 - add documentation about the restart of select() under alarm()
77 - small documentation tweaks
81 - better VMS operation (Craig Berry)
84 - our time machine is accelerating: now works with Perl 5.004_01
85 (tried with 5.003_07 and 5.002 but I get segmentation faults
86 from running the Makefile.PL with those in Tru64 4.0D)
89 - backward compatibility (pre-5.6.0) tweaks:
90 - no XSLoader in 5.00503, use DynaLoader instead
91 - no SvPV_nolen, either
92 - no PerlProc_pause(), either
93 - now tested with 5.00404 and 5.00503
94 - Makefile.PL requires 5.00404 (no more 5.002)
95 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
96 (this means that one can mix subsecond sleeps with alarms)
97 - because of nanosleep we probe for -lrt and -lposix4
98 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
99 is available by exportable constants Time::HiRes::d_func
100 (since older Perl do not have them in %Config, and even
101 5.8.0 does not probe for nanosleep)
104 - backward compatibility (pre-5.6.1) tweaks:
105 - define NV if no NVTYPE
106 - define IVdf if needed (note: the Devel::PPPort
107 in 5.8.0 does not try hard hard enough since
108 the IVSIZE might not be defined)
109 - define NVgf if needed
110 - grab the typemap from 5.8.0 for the NV stuff
113 - release 1.29_02 as 1.30
117 - fix a silly unclosed comment typo in HiRes.xs
118 - document and export REALTIME_REALPROF (Solaris)
122 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
123 (need to patch this also in Perl 5.[89])
124 - remove CVS revision log from HiRes.xs
128 The following numbered patches refer to the Perl 5.7 changes,
129 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
131 - 17558: Add #!./perl to the .t
132 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
133 - 16198: political correctness, from Simon Cozens
134 - 15857: doc tweaks, from Jarkko Hietaniemi
135 - 15593: optimization in .xs, from Paul Green
136 - 14892: pod fixes, from Robin Barker
137 - 14100: VOS fixes, from Paul Green
138 - 13422: XS segfault, from Marc Lehmann
139 - 13378: whether select() gets restarted on signals, depends
140 - 13354: timing constraints, again, from Andy Dougherty
141 - 13278: can't do subecond alarms with ualarm;
142 break out early if alarms do not seem to be working
143 - 13266: test relaxation (cygwin gets lower hires
144 times than lores ones)
145 - 12846: protect against high load, from Jarkko Hietaniemi
146 - 12837: HiRes.t VMS tweak, from Craig A. Berry
147 - 12797: HiRes.t VMS tweak, from Charles Lane
148 - 12769: HiRes.t VMS tweak, from Craig A. Berry
149 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
150 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
151 - 12692: alarm() ain't gonna work if ualarm() ain't,
152 from Gurusamy Sarathy
153 - 12680: minor VMS tweak, from Charles Lane
154 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
155 - 12609: croak on negative time, from Jarkko Hietaniemi
156 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
157 - 12594: MacOS Classic timeofday, from Chris Nandor
158 - 12473: allow for more than one second for sleep() and usleep()
159 - 12458: test tuning, relax timing constraints,
160 from Jarkko Hietaniemi
161 - 12449: make sleep() and usleep() to return the number
162 of seconds and microseconds actually slept (analogously
163 with the builtin sleep()), also make usleep() croak if
164 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
165 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
166 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
167 - 12196: use ftime() on Win32, from Artur Bergman
168 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
169 - 12105: use GetSystemTime() on Win32, from Artur Bergman
170 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
171 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
172 - 11797: problem in HiRes.t, from John P. Linderman
173 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
174 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
175 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
176 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
177 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
178 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
180 - 10942: MPE/IX test tweaks, from Mark Bixby
181 - 10784: unnecessary pod2man calls, from Andy Dougherty
182 - 10354: ext/ + -Wall, from Doug MacEachern
183 - 10320: fix the BOOT section to call myU2time correctly
184 - 10317: correct casting for AIX< from H. Merijn Brand
185 - 10119: document that the core time() may be rounding, not truncating
186 - 10118: test fix, from John Peacock
187 - 9988: long =item, from Robin Barker
188 - 9714: correct test output
189 - 9708: test also the scalar aspect of getitimer()
190 - 9705: Add interval timers (setitimer, getitimer)
191 - 9692: do not require at least 5.005 using XS
193 The following changes were made on top of the changes
194 made for Time::HiRes during the Perl 5.7 development
195 cycle that culminated in the release of Perl 5.8.0.
197 - add "require 5.005" to the Makefile.PL
198 - remove the REVISION section (CVS log) from HiRes.pm
199 - add jhi's copyright alongside Douglas'
200 - move HiRes.pm to lib/Time/
202 - modify HiRes.t to use $ENV{PERL_CORE}
203 - modify the original Time::HiRes version 1.20 Makefile.PL
204 to work both with Perl 5.8.0 and the new code with pre-5.8.0
205 Perls (tried with 5.6.1)
206 - tiny tweaks and updates in README and TODO
207 - bump the VERSION to 1.29
209 1.20 Wed Feb 24 21:30 1999
210 - make our usleep and ualarm substitutes into hrt_usleep
211 and hrt_ualarm. This helps static links of Perl with other
212 packages that also have usleep, etc. From
213 Ilya Zakharevich <ilya@math.ohio-state.edu>
214 - add C API stuff. From Joshua Pritikin
215 <joshua.pritikin@db.com>
216 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
217 - hopefully correct "-lc" fix for SCO.
220 1.19 Tue Sep 29 22:30 1998
221 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
222 <seb@stian.demon.co.uk>
223 - change GIMME_V to GIMME to help people with older versions of
225 - fix Win32 version of gettimeofday(). It didn't affect anything,
226 but it confuses people reading the code when the return value
227 is backwards (0 is success).
228 - fix Makefile.PL (more) so that detection of gettimeofday is
231 1.18 Mon Jul 6 22:40 1998
232 - add usleep() for Win32.
233 - fix Makefile.PL to fix reported HP/UX feature where unresolved
234 externals still cause an executable to be generated (though no
235 x bit set). Thanks to David Kozinn for report and explanation.
236 Problems with the fix are mine :)
238 1.17 Wed Jul 1 20:10 1998
239 - fix setitimer calls so microseconds is not more than 1000000.
240 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
241 - make Win32. We only get gettimeofday (the select hack doesn't
242 seem to work on my Win95 system).
243 - fix test 4 on 01test.t. add test to see if time() and
244 Time::HiRes::time() are close.
246 1.16 Wed Nov 12 21:05 1997
247 - add missing EXTEND in new gettimeofday scalar code.
249 1.15 Mon Nov 10 21:30 1997
250 - HiRes.pm: update pod. Provided by Gisle Aas.
251 - HiRes.xs: if gettimeofday() called in scalar context, do
252 something more useful than before. Provided by Gisle Aas.
253 - README: tell of xsubpp '-nolinenumber' woes. thanks to
254 Edward Henigin <ed@texas.net> for pointing out the problem.
256 1.14 Wed Nov 5 9:40 1997
257 - Makefile.PL: look for setitimer
258 - HiRes.xs: if missing ualarm, but we have setitimer, make up
259 our own setitimer. These were provided by Gisle Aas.
261 1.13 Tue Nov 4 23:30 1997
262 - Makefile.PL: fix autodetect mechanism to do try linking in addition
263 to just compiling; should fix Linux build problem. Fix was provided
266 1.12 Sun Oct 12 12:00:00 1997
267 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
268 you may need to comment this back out if you have an older xsubpp.
269 - HiRes.xs: set PROTOTYPES: DISABLE
271 1.11 Fri Sep 05 16:00:00 1997
273 Had some line commented out that shouldn't have been (testing
276 Previous version was corrupted.
278 1.10 Thu May 22 20:20:00 1997
279 - HiRes.xs, HiRes.pm, t/*:
280 - only compile what we have OS support for (or can
282 - only test what we compiled
283 - gross improvement to the test suite
285 This work was all done by Roderick Schertler
286 <roderick@argon.org>. If you run Linux or
287 one of the other ualarm-less platoforms, and you like this
288 module, let Roderick know; without him, it still wouldn't
289 be working on those boxes...
290 - Makefile.PL: figure out what routines the OS has and
291 only build what we need. These bits were written by Jarkko
292 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
294 1.02 Mon Dec 30 08:00:00 1996
295 - HiRes.pm: update documentation to say what to do when missing
296 ualarm() and friends.
297 - README: update to warn that ualarm() and friends need to exist
299 1.01 Fri Oct 17 08:00:00 1996
300 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
301 - HiRes.pm: put blank line between __END__ and =head1 so that
304 1.00 Tue Sep 03 13:00:00 1996
305 - original version; created by h2xs 1.16