1 Revision history for Perl extension Time::HiRes.
4 - In AIX (v?) with perl 5.6.1 the HiRes.t can hang after
5 the subtest 18. No known analysis nor fix, but added
6 an alarm (that requires fork() and alarm()) to the test.
9 - doc tweaks from mjd (perl change #20456)
10 - NCR MP-RAS hints file added (svr4.pl) (perl change #21249)
13 - add a message (for non-core builds) to Makefile.PL about
14 the LC_ALL=C workaround
15 - &Time::HiRes::d_nanosleep was broken (perl change #20131)
16 - the nanosleep() probe was broken (perl change #20061)
17 - use existence instead of definedness for feature probes
19 - MPE/iX tweak (perl change #20042)
20 - do not use HAS_NANOSLEEP (perl change #19898)
23 - UVuf for non-IVSIZE platforms (from Keiichiro Nagano)
24 - OS/2 can always mix subsecond sleeps with signals
25 (part of perl change #19789)
28 - workaround for buggy gcc 2.95.3 in openbsd/sparc64
32 - do not use -lrt in Linux (from March Lehmann, perl change #19449)
33 - unnecessary (nanosleep is in libc anyway)
34 - harmful (-lrt slows down execution)
35 - incompatible (with many distributions' pthreads)
38 - do not create files in blib directories under core
39 (perl change #19160, from rgs)
40 - detypo s/VTLARM/VTARLM/ (perl change #19328, from mjd)
43 - guarantee that $xdefine in HiRes.t is always defined
44 (perl change #19109, from IlyaZ)
45 - a cleaner way to detect PERL_CORE (perl change #19111,
49 - add hints/irix.pl to turn off overly POSIX flags that
50 cause hide struct timespec to be hidden (and compilation
51 to fail) (bleadperl change #19085)
52 - documentation tweaks
55 - add c:/temp to the list of temp directories to probe
56 so that cygwin (and win*?) builds are happy. This was
57 needed at least in my cygwin 1.3.20/w2k setup.
60 - modernize the constants code (from Nicholas Clark)
63 - At some point the ability to figure our the correct incdir
64 for EXTERN.h (either a core perl build, or an installed perl)
65 had broken (which lead into all test compiles failing with
66 a core perl build, but thanks to the robustness of Makefile.PL
67 nothing of this was visible). The brokenness seemed to be
68 caused by $ENV{PERL_CORE} not being on for core builds?
69 Now stole a trick from the Encode that sets $ENV{PERL_CORE}
70 right, and both styles of build should work again.
73 - Nicholas Clark noticed that the my_catdir() emulation function
74 was broken (which means that we didn't really work for Perls
76 - inspired by fixing the above made the whole Makefile.PL -w
78 - tightened up the Makefile.PL output, less whitespace
81 - fix from Craig Berry for better building in VMS with PERL_CORE
84 - no functional changes
85 - move lib/Time/HiRes.pm as Hires.pm
86 - libraries scanning was slightly broken (always scanned
87 for a library even when $Config{libs} already had it)
90 - Ray Zimmerman ran into a race condition in Mac OS X.
91 A 0.01-second alarm fired before the test expected.
92 The test first slept indefinitely (blocking for signals)
93 and only after that tested for the signal having been sent.
94 Since the signal had already been sent, the test #12 never
95 completed. The solution: test first, then block.
96 - default to being silent on all probing attempts, set the
97 environment variable VERBOSE to a true value to see the
98 details (the probing command and the possible errors)
101 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
102 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
103 cannot be upgraded from CPAN to override the 5.8.0 version
105 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
106 because the include directories did not adjust themselves
107 if $ENV{PERL_CORE} (Hugo van der Sanden)
108 - add documentation about the restart of select() under alarm()
111 - small documentation tweaks
115 - better VMS operation (Craig Berry)
118 - our time machine is accelerating: now works with Perl 5.004_01
119 (tried with 5.003_07 and 5.002 but I get segmentation faults
120 from running the Makefile.PL with those in Tru64 4.0D)
123 - backward compatibility (pre-5.6.0) tweaks:
124 - no XSLoader in 5.00503, use DynaLoader instead
125 - no SvPV_nolen, either
126 - no PerlProc_pause(), either
127 - now tested with 5.00404 and 5.00503
128 - Makefile.PL requires 5.00404 (no more 5.002)
129 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
130 (this means that one can mix subsecond sleeps with alarms)
131 - because of nanosleep we probe for -lrt and -lposix4
132 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
133 is available by exportable constants Time::HiRes::d_func
134 (since older Perl do not have them in %Config, and even
135 5.8.0 does not probe for nanosleep)
138 - backward compatibility (pre-5.6.1) tweaks:
139 - define NV if no NVTYPE
140 - define IVdf if needed (note: the Devel::PPPort
141 in 5.8.0 does not try hard hard enough since
142 the IVSIZE might not be defined)
143 - define NVgf if needed
144 - grab the typemap from 5.8.0 for the NV stuff
147 - release 1.29_02 as 1.30
151 - fix a silly unclosed comment typo in HiRes.xs
152 - document and export REALTIME_REALPROF (Solaris)
156 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
157 (need to patch this also in Perl 5.[89])
158 - remove CVS revision log from HiRes.xs
162 The following numbered patches refer to the Perl 5.7 changes,
163 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
165 - 17558: Add #!./perl to the .t
166 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
167 - 16198: political correctness, from Simon Cozens
168 - 15857: doc tweaks, from Jarkko Hietaniemi
169 - 15593: optimization in .xs, from Paul Green
170 - 14892: pod fixes, from Robin Barker
171 - 14100: VOS fixes, from Paul Green
172 - 13422: XS segfault, from Marc Lehmann
173 - 13378: whether select() gets restarted on signals, depends
174 - 13354: timing constraints, again, from Andy Dougherty
175 - 13278: can't do subecond alarms with ualarm;
176 break out early if alarms do not seem to be working
177 - 13266: test relaxation (cygwin gets lower hires
178 times than lores ones)
179 - 12846: protect against high load, from Jarkko Hietaniemi
180 - 12837: HiRes.t VMS tweak, from Craig A. Berry
181 - 12797: HiRes.t VMS tweak, from Charles Lane
182 - 12769: HiRes.t VMS tweak, from Craig A. Berry
183 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
184 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
185 - 12692: alarm() ain't gonna work if ualarm() ain't,
186 from Gurusamy Sarathy
187 - 12680: minor VMS tweak, from Charles Lane
188 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
189 - 12609: croak on negative time, from Jarkko Hietaniemi
190 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
191 - 12594: MacOS Classic timeofday, from Chris Nandor
192 - 12473: allow for more than one second for sleep() and usleep()
193 - 12458: test tuning, relax timing constraints,
194 from Jarkko Hietaniemi
195 - 12449: make sleep() and usleep() to return the number
196 of seconds and microseconds actually slept (analogously
197 with the builtin sleep()), also make usleep() croak if
198 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
199 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
200 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
201 - 12196: use ftime() on Win32, from Artur Bergman
202 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
203 - 12105: use GetSystemTime() on Win32, from Artur Bergman
204 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
205 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
206 - 11797: problem in HiRes.t, from John P. Linderman
207 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
208 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
209 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
210 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
211 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
212 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
214 - 10942: MPE/IX test tweaks, from Mark Bixby
215 - 10784: unnecessary pod2man calls, from Andy Dougherty
216 - 10354: ext/ + -Wall, from Doug MacEachern
217 - 10320: fix the BOOT section to call myU2time correctly
218 - 10317: correct casting for AIX< from H. Merijn Brand
219 - 10119: document that the core time() may be rounding, not truncating
220 - 10118: test fix, from John Peacock
221 - 9988: long =item, from Robin Barker
222 - 9714: correct test output
223 - 9708: test also the scalar aspect of getitimer()
224 - 9705: Add interval timers (setitimer, getitimer)
225 - 9692: do not require at least 5.005 using XS
227 The following changes were made on top of the changes
228 made for Time::HiRes during the Perl 5.7 development
229 cycle that culminated in the release of Perl 5.8.0.
231 - add "require 5.005" to the Makefile.PL
232 - remove the REVISION section (CVS log) from HiRes.pm
233 - add jhi's copyright alongside Douglas'
234 - move HiRes.pm to lib/Time/
236 - modify HiRes.t to use $ENV{PERL_CORE}
237 - modify the original Time::HiRes version 1.20 Makefile.PL
238 to work both with Perl 5.8.0 and the new code with pre-5.8.0
239 Perls (tried with 5.6.1)
240 - tiny tweaks and updates in README and TODO
241 - bump the VERSION to 1.29
243 1.20 Wed Feb 24 21:30 1999
244 - make our usleep and ualarm substitutes into hrt_usleep
245 and hrt_ualarm. This helps static links of Perl with other
246 packages that also have usleep, etc. From
247 Ilya Zakharevich <ilya@math.ohio-state.edu>
248 - add C API stuff. From Joshua Pritikin
249 <joshua.pritikin@db.com>
250 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
251 - hopefully correct "-lc" fix for SCO.
254 1.19 Tue Sep 29 22:30 1998
255 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
256 <seb@stian.demon.co.uk>
257 - change GIMME_V to GIMME to help people with older versions of
259 - fix Win32 version of gettimeofday(). It didn't affect anything,
260 but it confuses people reading the code when the return value
261 is backwards (0 is success).
262 - fix Makefile.PL (more) so that detection of gettimeofday is
265 1.18 Mon Jul 6 22:40 1998
266 - add usleep() for Win32.
267 - fix Makefile.PL to fix reported HP/UX feature where unresolved
268 externals still cause an executable to be generated (though no
269 x bit set). Thanks to David Kozinn for report and explanation.
270 Problems with the fix are mine :)
272 1.17 Wed Jul 1 20:10 1998
273 - fix setitimer calls so microseconds is not more than 1000000.
274 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
275 - make Win32. We only get gettimeofday (the select hack doesn't
276 seem to work on my Win95 system).
277 - fix test 4 on 01test.t. add test to see if time() and
278 Time::HiRes::time() are close.
280 1.16 Wed Nov 12 21:05 1997
281 - add missing EXTEND in new gettimeofday scalar code.
283 1.15 Mon Nov 10 21:30 1997
284 - HiRes.pm: update pod. Provided by Gisle Aas.
285 - HiRes.xs: if gettimeofday() called in scalar context, do
286 something more useful than before. Provided by Gisle Aas.
287 - README: tell of xsubpp '-nolinenumber' woes. thanks to
288 Edward Henigin <ed@texas.net> for pointing out the problem.
290 1.14 Wed Nov 5 9:40 1997
291 - Makefile.PL: look for setitimer
292 - HiRes.xs: if missing ualarm, but we have setitimer, make up
293 our own setitimer. These were provided by Gisle Aas.
295 1.13 Tue Nov 4 23:30 1997
296 - Makefile.PL: fix autodetect mechanism to do try linking in addition
297 to just compiling; should fix Linux build problem. Fix was provided
300 1.12 Sun Oct 12 12:00:00 1997
301 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
302 you may need to comment this back out if you have an older xsubpp.
303 - HiRes.xs: set PROTOTYPES: DISABLE
305 1.11 Fri Sep 05 16:00:00 1997
307 Had some line commented out that shouldn't have been (testing
310 Previous version was corrupted.
312 1.10 Thu May 22 20:20:00 1997
313 - HiRes.xs, HiRes.pm, t/*:
314 - only compile what we have OS support for (or can
316 - only test what we compiled
317 - gross improvement to the test suite
319 This work was all done by Roderick Schertler
320 <roderick@argon.org>. If you run Linux or
321 one of the other ualarm-less platoforms, and you like this
322 module, let Roderick know; without him, it still wouldn't
323 be working on those boxes...
324 - Makefile.PL: figure out what routines the OS has and
325 only build what we need. These bits were written by Jarkko
326 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
328 1.02 Mon Dec 30 08:00:00 1996
329 - HiRes.pm: update documentation to say what to do when missing
330 ualarm() and friends.
331 - README: update to warn that ualarm() and friends need to exist
333 1.01 Fri Oct 17 08:00:00 1996
334 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
335 - HiRes.pm: put blank line between __END__ and =head1 so that
338 1.00 Tue Sep 03 13:00:00 1996
339 - original version; created by h2xs 1.16