Fix long standing memory leak with pop and shift!
[p5sagit/p5-mst-13.2.git] / ext / Time / HiRes / Changes
CommitLineData
dcf686c9 1Revision history for Perl extension Time::HiRes.
2
6a20eacc 31.39
4 - fix from Craig Berry for better building in VMS with PERL_CORE
5
3f2ee006 61.38
7 - no functional changes
8 - move lib/Time/HiRes.pm as Hires.pm
9 - libraries scanning was slightly broken (always scanned
10 for a library even when $Config{libs} already had it)
11
121.37
13 - Ray Zimmerman ran into a race condition in Mac OS X.
14 A 0.01-second alarm fired before the test expected.
15 The test first slept indefinitely (blocking for signals)
16 and only after that tested for the signal having been sent.
17 Since the signal had already been sent, the test #12 never
18 completed. The solution: test first, then block.
19 - default to being silent on all probing attempts, set the
20 environment variable VERBOSE to a true value to see the
21 details (the probing command and the possible errors)
22
231.36
24 - do not clear MAN3PODS in Makefile.PL (Radoslaw Zielinski)
25 - INSTALLDIRS => 'perl' missing which means that Time::HiRes
26 cannot be upgraded from CPAN to override the 5.8.0 version
27 (Guido A. Ostkamp)
28 - Time::HiRes 1.35 could not be dropped as-is to bleadperl
29 because the include directories did not adjust themselves
30 if $ENV{PERL_CORE} (Hugo van der Sanden)
31 - add documentation about the restart of select() under alarm()
32
331.35
34 - small documentation tweaks
35
36
371.34
38 - better VMS operation (Craig Berry)
39
401.33
41 - our time machine is accelerating: now works with Perl 5.004_01
42 (tried with 5.003_07 and 5.002 but I get segmentation faults
43 from running the Makefile.PL with those in Tru64 4.0D)
44
451.32
46 - backward compatibility (pre-5.6.0) tweaks:
47 - no XSLoader in 5.00503, use DynaLoader instead
48 - no SvPV_nolen, either
49 - no PerlProc_pause(), either
50 - now tested with 5.00404 and 5.00503
51 - Makefile.PL requires 5.00404 (no more 5.002)
52 - use nanosleep instead of usleep, if it is available (Wilson Snyder)
53 (this means that one can mix subsecond sleeps with alarms)
54 - because of nanosleep we probe for -lrt and -lposix4
55 - the existence of getitimer/nanosleep/setitimer/ualarm/usleep
56 is available by exportable constants Time::HiRes::d_func
57 (since older Perl do not have them in %Config, and even
58 5.8.0 does not probe for nanosleep)
59
601.31
61 - backward compatibility (pre-5.6.1) tweaks:
62 - define NV if no NVTYPE
63 - define IVdf if needed (note: the Devel::PPPort
64 in 5.8.0 does not try hard hard enough since
65 the IVSIZE might not be defined)
66 - define NVgf if needed
67 - grab the typemap from 5.8.0 for the NV stuff
681.30
69
70 - release 1.29_02 as 1.30
71
721.29_02
73
74 - fix a silly unclosed comment typo in HiRes.xs
75 - document and export REALTIME_REALPROF (Solaris)
76
771.29_01
78
79 - only getitimer(ITIMER_REAL) available in Cygwin and Win32
80 (need to patch this also in Perl 5.[89])
81 - remove CVS revision log from HiRes.xs
82
831.29_00
84
85 The following numbered patches refer to the Perl 5.7 changes,
86 you can browse them at http://public.activestate.com/cgi-bin/perlbrowse
87
88 - 17558: Add #!./perl to the .t
89 - 17201: linux + usemorebits fix, from Rafael Garcia-Suarez
90 - 16198: political correctness, from Simon Cozens
91 - 15857: doc tweaks, from Jarkko Hietaniemi
92 - 15593: optimization in .xs, from Paul Green
93 - 14892: pod fixes, from Robin Barker
94 - 14100: VOS fixes, from Paul Green
95 - 13422: XS segfault, from Marc Lehmann
96 - 13378: whether select() gets restarted on signals, depends
97 - 13354: timing constraints, again, from Andy Dougherty
98 - 13278: can't do subecond alarms with ualarm;
99 break out early if alarms do not seem to be working
100 - 13266: test relaxation (cygwin gets lower hires
101 times than lores ones)
102 - 12846: protect against high load, from Jarkko Hietaniemi
103 - 12837: HiRes.t VMS tweak, from Craig A. Berry
104 - 12797: HiRes.t VMS tweak, from Charles Lane
105 - 12769: HiRes.t VMS tweak, from Craig A. Berry
106 - 12744: gcc vs MS 64-bit constant syntax, from Nick Ing-Simmons
107 - 12722: VMS ualarm for VMS without ualarm, from Charles Lane
108 - 12692: alarm() ain't gonna work if ualarm() ain't,
109 from Gurusamy Sarathy
110 - 12680: minor VMS tweak, from Charles Lane
111 - 12617: don't try to print ints as IVs, from Jarkko Hietaniemi
112 - 12609: croak on negative time, from Jarkko Hietaniemi
113 - 12595: Cygwin rounds up for time(), from Jarkko Hietaniemi
114 - 12594: MacOS Classic timeofday, from Chris Nandor
115 - 12473: allow for more than one second for sleep() and usleep()
116 - 12458: test tuning, relax timing constraints,
117 from Jarkko Hietaniemi
118 - 12449: make sleep() and usleep() to return the number
119 of seconds and microseconds actually slept (analogously
120 with the builtin sleep()), also make usleep() croak if
121 asked for more than 1_000_000 useconds, from Jarkko Hietaniemi
122 - 12366: Time::HiRes for VMS pre-7.0, from Charles Lane
123 - 12199: do not use ftime on Win32, from Gurusamy Sarathy
124 - 12196: use ftime() on Win32, from Artur Bergman
125 - 12184: fix Time::HiRes gettimeofday() on Win32, from Gurusamy Sarathy
126 - 12105: use GetSystemTime() on Win32, from Artur Bergman
127 - 12060: explain the 1e9 seconds problem, from Jarkko Hietaniemi
128 - 11901: UNICOS sloppy division, from Jarkko Hietaniemi
129 - 11797: problem in HiRes.t, from John P. Linderman
130 - 11414: prototype from Time::HiRes::sleep(), from Abhijit Menon-Sen
131 - 11409: Time::HiRes qw(sleep) failed, from Abhijit Menon-Sen
132 - 11270: dynix/ptx 4.5.2 hints fix, from Peter Prymmer
133 - 11032: VAX VMS s/div/lib\$ediv/ fix, from Peter Prymmer
134 - 11011: VAX VMS s/qdiv/div/ fix, from Peter Prymmer
135 - 10953: SCO OpenServer 5.0.5 requires an explicit -lc for usleep(),
136 from Jonathan Stowe
137 - 10942: MPE/IX test tweaks, from Mark Bixby
138 - 10784: unnecessary pod2man calls, from Andy Dougherty
139 - 10354: ext/ + -Wall, from Doug MacEachern
140 - 10320: fix the BOOT section to call myU2time correctly
141 - 10317: correct casting for AIX< from H. Merijn Brand
142 - 10119: document that the core time() may be rounding, not truncating
143 - 10118: test fix, from John Peacock
144 - 9988: long =item, from Robin Barker
145 - 9714: correct test output
146 - 9708: test also the scalar aspect of getitimer()
147 - 9705: Add interval timers (setitimer, getitimer)
148 - 9692: do not require at least 5.005 using XS
149
150 The following changes were made on top of the changes
151 made for Time::HiRes during the Perl 5.7 development
152 cycle that culminated in the release of Perl 5.8.0.
153
154 - add "require 5.005" to the Makefile.PL
155 - remove the REVISION section (CVS log) from HiRes.pm
156 - add jhi's copyright alongside Douglas'
157 - move HiRes.pm to lib/Time/
158 - move HiRes.t to t/
159 - modify HiRes.t to use $ENV{PERL_CORE}
160 - modify the original Time::HiRes version 1.20 Makefile.PL
161 to work both with Perl 5.8.0 and the new code with pre-5.8.0
162 Perls (tried with 5.6.1)
163 - tiny tweaks and updates in README and TODO
164 - bump the VERSION to 1.29
165
dcf686c9 1661.20 Wed Feb 24 21:30 1999
167 - make our usleep and ualarm substitutes into hrt_usleep
168 and hrt_ualarm. This helps static links of Perl with other
169 packages that also have usleep, etc. From
170 Ilya Zakharevich <ilya@math.ohio-state.edu>
171 - add C API stuff. From Joshua Pritikin
172 <joshua.pritikin@db.com>
3f2ee006 173 - VMS Makefile.PL fun. From pvhp@forte.com (Peter Prymmer)
dcf686c9 174 - hopefully correct "-lc" fix for SCO.
175 - add PPD stuff
176
1771.19 Tue Sep 29 22:30 1998
178 - put VMS gettimeofday() in. Patch is from Sebastian Bazley
179 <seb@stian.demon.co.uk>
180 - change GIMME_V to GIMME to help people with older versions of
181 Perl.
182 - fix Win32 version of gettimeofday(). It didn't affect anything,
183 but it confuses people reading the code when the return value
184 is backwards (0 is success).
185 - fix Makefile.PL (more) so that detection of gettimeofday is
186 more correct.
187
1881.18 Mon Jul 6 22:40 1998
189 - add usleep() for Win32.
190 - fix Makefile.PL to fix reported HP/UX feature where unresolved
191 externals still cause an executable to be generated (though no
192 x bit set). Thanks to David Kozinn for report and explanation.
193 Problems with the fix are mine :)
194
1951.17 Wed Jul 1 20:10 1998
196 - fix setitimer calls so microseconds is not more than 1000000.
197 Hp/UX 9 doesn't like that. Provided by Roland B Robert, PhD.
3f2ee006 198 - make Win32. We only get gettimeofday (the select hack doesn't
dcf686c9 199 seem to work on my Win95 system).
3f2ee006 200 - fix test 4 on 01test.t. add test to see if time() and
dcf686c9 201 Time::HiRes::time() are close.
202
2031.16 Wed Nov 12 21:05 1997
204 - add missing EXTEND in new gettimeofday scalar code.
205
2061.15 Mon Nov 10 21:30 1997
207 - HiRes.pm: update pod. Provided by Gisle Aas.
208 - HiRes.xs: if gettimeofday() called in scalar context, do
209 something more useful than before. Provided by Gisle Aas.
210 - README: tell of xsubpp '-nolinenumber' woes. thanks to
211 Edward Henigin <ed@texas.net> for pointing out the problem.
212
2131.14 Wed Nov 5 9:40 1997
214 - Makefile.PL: look for setitimer
215 - HiRes.xs: if missing ualarm, but we have setitimer, make up
216 our own setitimer. These were provided by Gisle Aas.
217
2181.13 Tue Nov 4 23:30 1997
219 - Makefile.PL: fix autodetect mechanism to do try linking in addition
220 to just compiling; should fix Linux build problem. Fix was provided
221 by Gisle Aas.
222
2231.12 Sun Oct 12 12:00:00 1997
224 - Makefile.PL: set XSOPT to '-nolinenumbers' to work around xsubpp bug;
225 you may need to comment this back out if you have an older xsubpp.
226 - HiRes.xs: set PROTOTYPES: DISABLE
227
2281.11 Fri Sep 05 16:00:00 1997
229 - Makefile.PL:
230 Had some line commented out that shouldn't have been (testing
231 remnants)
232 - README:
233 Previous version was corrupted.
234
2351.10 Thu May 22 20:20:00 1997
236 - HiRes.xs, HiRes.pm, t/*:
237 - only compile what we have OS support for (or can
238 fake with select())
239 - only test what we compiled
240 - gross improvement to the test suite
241 - fix EXPORT_FAIL.
242 This work was all done by Roderick Schertler
243 <roderick@argon.org>. If you run Linux or
244 one of the other ualarm-less platoforms, and you like this
245 module, let Roderick know; without him, it still wouldn't
246 be working on those boxes...
247 - Makefile.PL: figure out what routines the OS has and
248 only build what we need. These bits were written by Jarkko
249 Hietaniemi <jhi@iki.fi>. Again, gratitude is due...
250
2511.02 Mon Dec 30 08:00:00 1996
252 - HiRes.pm: update documentation to say what to do when missing
253 ualarm() and friends.
254 - README: update to warn that ualarm() and friends need to exist
255
2561.01 Fri Oct 17 08:00:00 1996
257 - Makefile.PL: make XSPROTOARGS => '-noprototyopes'
258 - HiRes.pm: put blank line between __END__ and =head1 so that
259 pod2man works.
260
2611.00 Tue Sep 03 13:00:00 1996
262 - original version; created by h2xs 1.16