ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/AnyEvent/Changes
Revision: 1.225
Committed: Wed Apr 1 14:02:27 2009 UTC (15 years, 1 month ago) by root
Branch: MAIN
Changes since 1.224: +3 -0 lines
Log Message:
*** empty log message ***

File Contents

# Content
1 Revision history for Perl extension AnyEvent.
2
3 - set the cloexec flag on the signal pipe.
4 - actually make the signal pipe work under win32.
5
6 4.35 Fri Mar 27 11:48:20 CET 2009
7 - event models relying on AnyEvent's signal watcher emulation did
8 invoke the callback asynchronously, contradicting documentation
9 and causing signals to get lost (this includes AnyEvent's own
10 event loop). AnyEvent now uses the standard pipe trick to make
11 callback execution synchronous to the event loop.
12 - AnyEvent::Handle didn't free TLS context data on DESTROY
13 (patch by Pavel Shaydo).
14 - work around the ever-incompatibly-changing API of MakeMaker.
15 - document that changing global variables without restoring them
16 is a bad idea in Perl (noted by Adam Rosenstein).
17 - AnyEvent::Strict now barfs if ->io is passed a file.
18
19 4.34 Thu Feb 12 18:32:45 CET 2009
20 - separately buffer TLS read data, as otherwise the read queue could
21 deadlock as receiving data is not expected while draining the
22 read queue (which cna only happen with TLS).
23 - raise EBADMSG error on JSON decoding errors.
24 - fix some minor manpage bugs (reported by Maximilian Gaß).
25 - speed up select bitmask parsing quite a bit in the pure perl
26 backend.
27 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
28
29 4.331 Tue Jan 6 21:07:25 CET 2009
30 - socketpair fails on many vista machines because vista has
31 completely broken accept/getpeername and getsockname functions,
32 so we provide our own socketpair emulation that kind of works
33 (AnyEvent::Util::portable_pipe).
34 - new function: AnyEvent::Util::portable_socketpair.
35 - take advantage of the Guard module if it exists.
36
37 4.33 Fri Nov 21 02:35:40 CET 2008
38 - AnyEvent::Strict did errornously flag a fileno of 0 as
39 illegal.
40 - reduce memory usage and slightly speed up the pure perl backend
41 by only storing the file descriptor, not the file handle.
42 - add missing autocork method to AnyEvent::Handle (reported
43 by Adam Rosenstein).
44 - AnyEvent::DNS->resolve errornously documented an $rcode result
45 argument, but there isn't (spotted by Henrik Krohns).
46 - the naked truth about Tk - it's basically unmaintained.
47
48 4.32 Mon Nov 3 22:46:32 CET 2008
49 - fix AnyEvent::Socket::resolve_sockaddr to properly support
50 unix sockets again and choose a proper default for
51 the socket type.
52 - fix call to resolve_sockaddr in tcp_connect.
53
54 4.31 Thu Oct 30 04:41:48 CET 2008
55 - implemented AnyEvent::Handle->destroy method.
56 - hint about unexpected effects in TLS mode.
57 - speed up AnyEvent::Impl::Perl by using more arrays
58 instead of hashes without the slightest loss of readability :)
59 - work around a perl argument refcounting bug.
60 - with some perl patches applied, AnyEvent now seems leak-free.
61
62 4.3 Fri Oct 3 09:18:43 CEST 2008
63 - AnyEvent will now install a no-op signal handler for SIGPIPE,
64 unless one has been installed already.
65 - warn about not loaidng AnyEvent::Impl::POE early enough
66 (patch by Adam Rosenstein).
67 - fixed a great number of bugs and corner cases in AnyEvent::Handle:
68 - fix a bug in where in SSL connect mode, the
69 client would first wait for some data by the server and
70 otherwise hang. (reported and analyzed in an absolutely
71 exemplary manner by Adam Rosenstein).
72 - fix a bug in where SSL EOF would not be treated
73 as stream EOF, putting the connection into a hung state.
74 (reported and analyzed in an absolutely exemplary manner by Adam
75 Rosenstein).
76 - fix a potential segfault when the TLS context
77 would go missing in a read callback (e.g. due to stoptls) -
78 Net::SSLeay of course makes no type checking whatsoever.
79 - AnyEvent::Handle will not stop the read watcher in TLS mode.
80 - AnyEvent::Handle->stoptls will send a TLS close notify
81 instead of simply destroying the stream now.
82 - fix a bug where large blocks of data written in TLS mode
83 would not be sent unless triggered by receives.
84 - on_drain will now take the tls write buffer into account.
85 - SSL operations have been streamlined a bit and should be faster
86 now, more to come, though.
87 - added a FAQ section, will document nontrivial issues in there.
88
89 4.234 Mon Sep 29 04:08:13 CEST 2008
90 - fix child watcher documentation: only child exits will be reported,
91 no trace events.
92 - mention SIGPIPE in AnyEvent::Handle.
93 - perl backend will now use POSIX::times as monotonic clock
94 when available and the monotonic clock option is not.
95 - run even without Time::HiRes in some cases now, but do not
96 provide sub-second accuracy (all *supported* perl versions have
97 Time::HiRes).
98
99 4.233 Fri Aug 22 01:48:05 CEST 2008
100 - fix a bug in the testsuite, causing a hang.
101 - clarified lots of AnyEvent::Handle constructor arguments.
102
103 4.232 Thu Aug 21 20:44:25 CEST 2008
104 - fix a bug in AnyEvent::Handle that could cause two fatal error
105 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
106 (testcase by Leon Brocard).
107 - clarified on_eof/on_error documentation a bit.
108
109 4.231 Tue Jul 29 13:12:15 CEST 2008
110 - remove some debugging code left in AnyEvent::Util::fork_call
111 (and no, it's impossible to implement with the broken windows
112 perls without resource leaks or worse).
113
114 4.23 Tue Jul 29 12:19:59 CEST 2008
115 - document the first parameter passed to condvar callbacks to be
116 the callback.
117 - add AnyEvent::Socket::{ntoa,aton} aliases.
118 - optimize the AE::Handle->push_read (line) for the default
119 eol marker.
120 - optimize push_read (packstring|storable) for small packets.
121 - invoke on_error callback when no on_eof callback is set.
122 - fix a bug in push_read (storable) of unknown impact.
123
124 4.22 Sun Jul 20 16:34:13 CEST 2008
125 - new function AnyEvent::Socket::parse_hostport.
126 - as the bulkheads at microsoft can't even get getprotobyname reliably
127 working on their shitty fucking broken os we need to hardcode
128 some common protocol numbers in AnyEvent::Socket. How can
129 people even bother with such a pile of shit as windows.
130
131 4.21 Thu Jul 17 14:40:05 CEST 2008
132 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
133 which is incompatible to previous invocation syntax.
134 - work around more windows perl bugs in fork_call. I cannot imagine
135 why anybody would bother implementing fork in such an obviously
136 unusable way.
137 - avoid calling $do_retry if the request has already finished in
138 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
139 K. Sharifulin).
140
141 4.2 Sat Jul 12 22:42:11 CEST 2008
142 - support an "interval" argument for repeating timers.
143 - fix a bug in the pure perl event loop that caused timers to
144 be delayed under some circumstances.
145 - implement optional argument checking via PERL_ANYEVENT_STRICT.
146 - not importing anything from Event causes it to malfunction
147 and not use Time::HiRes, work around this bug.
148 - more bug workarounds against the endlessly broken Tk module,
149 seems to work now on 32 bit machines, but less so on 64 bit.
150 - minor optimisations applied to most backends.
151
152 4.161 Fri Jul 4 14:24:48 CEST 2008
153 - fixed the t/handle/02_write.t, it read after write, with the
154 assumption that all data could be written before read
155 (reported by Jim Razmus).
156
157 4.160 Thu Jul 3 04:02:21 CEST 2008
158 - re-registering signal watchers after unregistering one did
159 not work if a backend falls back on AnyEvent's default
160 implementation (also affected child watchers).
161 - new AnyEvent::Handle options: autocork and no_delay.
162
163 4.152 Sun Jun 22 14:15:44 CEST 2008
164 - allow for 32-bit perls that implement shifts differently
165 on different architectures in parse_ipv4 (reported and
166 analysed by Keiichi DAIBA).
167
168 4.151 Fri Jun 6 17:34:24 CEST 2008
169 - make sure specifying _only_ on_read and never pushing reads
170 works.
171
172 4.15 Fri Jun 6 13:00:46 CEST 2008
173 - the pure perl backend would keep some watchers alive when more than
174 one watcher was registered for the same fd.
175 - new "packstring" and "storable" read and write types
176 for AnyEvent::Handle.
177 - allow on_eof handler to be called after on_error with EPIPE returns.
178 - do not immediately call on_read callback in handle constructor.
179
180 4.14 Thu Jun 5 20:29:31 CEST 2008
181 - Fixed a bug in DNS SRV priority sorting.
182 - AnyEvent::Util::guard now reports runtime errors while
183 executing the guard block as warnings.
184 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
185
186 4.13 Thu Jun 5 00:47:59 CEST 2008
187 - AnyEvent::DNS only followed cname chains with length 2,
188 contrary to documentation. bumped it up to 10, thanks to
189 microsoft, the current broken-dns-config-king.
190 - AnyEvent::DNS didn't check for socket return status
191 but instead relied on perl not creating filehandles in that
192 case - too bad it gives you a bogus file handle (reported
193 and analysed by Vladimir Timofeev).
194 - fix queue management logic in AnyEvent::Handle:
195 when on_read was registered, the queue was empty
196 and no progress could be made AnyEvent::Handle would
197 enter an endless loop.
198 - correctly start reading again when the handle became
199 busy again after idling.
200 - correctly treat tls shut-downs as EOF condition
201 (for the time being, would be nice to have a callback for that).
202 - correctly call eof callback under all conditions
203 (hopefully).
204 - the timeout callback did not expect that $self can go
205 away any time.
206
207 4.12 Tue Jun 3 10:58:04 CEST 2008
208 - include AnyEvent::Intro, a tutorial for anyevent,
209 anyevent::socket and anyevent::handle.
210 - allow more options in on_error.
211
212 4.11 Fri May 30 23:42:25 CEST 2008
213 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
214 reverse_lookup and reverse_verify to replace it, support
215 v4mapped and v4compat addresses.
216 - provide more documentation for the resolver class.
217 - really replace longest run of :0: by :: in format_address,
218 also properly convert :: and ::1 again.
219 - support NAPTR record name and decode it.
220 - implement random weight sampling for SRV records, as per
221 rfc 2782.
222 - correctly abort on srv-record targets of ".".
223 - added AnyEvent::DNS::wait_for_slot.
224 - in the unlikely event of a virtual circuit connection
225 being invalidated by a delayed udp reply, AnyEvent::DNS
226 could die. this has been fixed.
227 - plug a probable memleak in the DNS vc code.
228 - use configured timeout also for VC request phase.
229 - implement timeout and max_outstanding methods
230 for resolver class.
231 - update version numbers in all modules.
232
233 4.1 Thu May 29 05:45:40 CEST 2008
234 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
235 which mirrors their purpose better (old names still available
236 for a while). Also moved AnyEvent::DNS::addr to
237 AnyEvent::Socket::resolve_sockaddr.
238 - implement AnyEvent->time and AnyEvent->now.
239 - fix IPv6 support in pack_sockaddr.
240 - officially un-experimentalise all newly introduced code.
241 - support unix domain sockets everywhere by specifying
242 a host of "unix/" and the pathname as service.
243 - implement an activity timeout in AnyEvent::Handle.
244 - added a regex read type to AnyEvent::Handle.
245 - added a json read/write type to AnyEvent::Handle.
246 - always croak in on_error in AnyEvent::Handle.
247 - document how to subclass AnyEvent::Handle.
248 - implement AnyEvent::Util::fork_call.
249 - add support for IPv6 nameservers and nameserver statements.
250 - work around _yet_ _another_ windows perl bug in where empty
251 select masks cause errors in the pure perl backend (this
252 workaround cannot be implemented for the other event loops,
253 but EV also works around this).
254 - supply AnyEvent::Socket::address_type and make good use of it.
255 - clarify and fix the Handle documentation w.r.t $self vs. $handle
256 vs. callback arguments.
257 - add some recommends to the META.yml.
258
259 4.05 Mon May 26 19:44:06 CEST 2008
260 - some platforms ignore the rfc and prepend an extra sa_len member to
261 the sockaddr structure, cater for those. also use sockaddr_family,
262 didn't know it was there, but comes in handy for the workaround.
263 - undo the import WIN32 hack.
264
265 4.04 Mon May 26 08:03:31 CEST 2008
266 - try to work around yet another windows bug: failed connects are reported
267 as if it were out-of-band data. windows users: you suck. Only
268 EV, Glib, Event and the pure perl backend can handle this condition.
269 - optimize write algorithm in AnyEvent::Handle.
270 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
271 - tcp_connect now properly iterates through all targets.
272 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
273 support entirely. also hardcode some AF_INET6 constants for the
274 sake of perl 5.8.
275 - vastly improved nameserver/suffix detection algorithm for windows
276 (still a hack...)
277 - try to use a fallback dns server if no dns servers could be found.
278 - splatter around some local $SIG{__DIE__}'s to work around bad
279 die handlers.
280
281 4.03 Sun May 25 05:05:57 CEST 2008
282 - work around perl bugs on windows, where perl returns undocumented
283 error codes for sysread, syswrite etc.
284 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
285 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
286 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
287 - updated documentation.
288 - lots of workarounds for perl 5.8 quirks.
289 - implement a more flexible type system for both reads and writes
290 in AnyEvent::Handle.
291 - support IPv6 in tcp_server.
292 - pass local hostname and port to the prepare callback in tcp_server.
293 - make types extendable in AnyEvent::Handle.
294 - croak, not die, in AnyEvent::Handle;
295
296 4.0 Sat May 24 19:58:08 CEST 2008
297 - added (experimental) AnyEvent::DNS module.
298 - added (experimental) AnyEvent::Socket module.
299 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle.
300 - make condvar->end call ->send as documented.
301 - make condvar constructor support the documented cb => argument.
302 - overload condvars so you can use them instead of code references.
303 - support more of the POSIX forms instead of a pure dotted quad for
304 verbatim ip addresses.
305 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
306 - try very hard not to freeze on broken windows perls.
307 - make on_eof optional in AnyEvent::Handle.
308 - support service names in tcp_connect and tcp_server.
309 - work around netbsd/openbsd bug causing unix sockets not to
310 become writable until empty.
311 - only make a dummy read for error detection when getpeername
312 returns ENOTCONN.
313 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
314 override.
315 - new AnyEvent::Guard::cancel.
316
317 3.5 Sat May 17 23:17:14 CEST 2008
318 - gracefully deal with systems not defining CLOCK_MONOTONIC
319 (as opposed to not supporting the clock itself, which worked in 3.41).
320 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server
321 helpers, to replace the ill-designed AnyEvent::Socket class.
322 - added some examples (eg/connect, eg/handle).
323 - added AnyEvent::Util::guard.
324 - AnyEvent::Util now exports most of its functions.
325 - fixed a bug in *_read_chunk, where the callback was called with $self as
326 first and second argument.
327 - simplified and fixed regex handling in AnyEvent::Handle::*_read_line.
328 - fix manpage condvars examples to use new syntax.
329 - considerably improve inet_aton with EV::ADNS backend to support
330 cname chains (common dns configuration bug, e.g. with www.google.de)
331 and aaaa records if no a record could be found.
332
333 3.41 Sun May 11 19:53:13 CEST 2008
334 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
335
336 3.4 Sun May 11 00:29:25 CEST 2008
337 - complete re-implementation of condvars with a hopefully much more useful
338 API (backward compatible functions exist but are undocumented).
339 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
340 (still subject to change).
341 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
342 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which
343 allows module authors to avoid forcing event loop detection.
344 (used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example).
345 - remove coro backends: Coro now provides generic support for AnyEvent
346 via Coro::AnyEvent.
347
348 3.3 Mon Apr 28 09:51:06 CEST 2008
349 - added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules.
350 - fix a bug in the pure perl backend that kept watchers alive
351 when multiple watchers were registered for the same fh/poll combo.
352 - add a benchmark section showing AnyEvent overhead and comparing
353 the different event loops with each other.
354 - prefer pure perl over tk when autoprobing, as it's about as fast,
355 but doesn't crash with many watchers.
356 - declare Qt support non-experimental.
357 - clarify the confusing section about the file descriptor being
358 kept alive.
359 - document the race between loading of an event module and
360 child processes exiting.
361 - support POE as "backend" (with some caveats, POE is not generic
362 enough, and darn slow).
363 - support Wx and Prima through POE.
364 - optimise perl backend to use 20% less memory and take advantage
365 of typical timeout behaviour. It can now compete with
366 select/poll-based C event loops in most cases (it is usually
367 faster than Event and Glib :).
368 - roughly cut EV memory use in half and increase its speed by 30%,
369 by removing undocumented functionality. Did something similar
370 to other event loops where possible.
371
372 3.2 Thu Apr 24 10:10:40 CEST 2008
373 - do not die when anyevent watchers are destroyed while running callbacks
374 in the pureperl backend (could only happen when two watchers are registered
375 for the same fh/poll combo).
376 - support autoloading for child watchers, was broken in all versions.
377 - implement PERL_ANYEVENT_MODEL env variable.
378 - (experimental) implement interface to the (very crashy/buggy) Event::Lib module.
379 - (experimental) implement interface to the Qt module (cannot be autoprobed).
380 - this release is IO:AnyEvent-proof.
381
382 3.12 Tue Apr 22 07:11:46 CEST 2008
383 - reinstate AUTHOR section that got lost somehow.
384 - do not hang in the testsuite with badly broken perls (activestate,
385 strawberry...), but instead diagnose the problem and continue.
386 - use INT instead of CHLD in an attempt to support broken windows perls
387 better (this decreases the test quality, unfortunately).
388 - do not send the signal to the process group
389 (no problem for CHLD, bad for INT :).
390
391 3.11 Sat Apr 19 06:57:31 CEST 2008
392 - major documentation rework.
393 - document the fact that child watchers only watch for zombies.
394 - fix the child watcher example.
395
396 3.1 Wed Apr 16 17:09:01 CEST 2008
397 - work around recurring bugs in Tk by dup'ing filehandles, the
398 only method with good success chances on Tk (the bugs apparently
399 don't get fixed anytime soon).
400 - lift the restriction of only one watcher per fh direction
401 (as the Tk bug workaround also lifts it and only Tk imposed
402 such strong limits).
403 - changed probe order to prefer coro adaptors.
404 - explain why recursion into the event loop is not supported
405 unless the backend supports it (only Coro::EV does without
406 any restrictions...).
407 - add simple manpages for all backend modules.
408
409 3.0 Mon Apr 7 21:30:23 CEST 2008
410 - Coro::Signal changed semantics, roll our own, also cleaning
411 up the Coro implementation in general.
412 - rename Coro backend to CoroEvent.
413 - add some decision helping paragraph to the manpage that should
414 help people to decide whether AnyEvent is the right thing for them.
415
416 2.9 Mon Jan 28 13:31:54 CET 2008
417 - update for EV 3.0 API changes.
418
419 2.8 Sun Nov 25 15:06:03 CET 2007
420 - waitpid can validly return 0. accept this fact of life
421 instead of reporting it to any watchers.
422
423 2.7 Fri Nov 23 11:41:14 CET 2007
424 - force use of AnyEvent::Impl::Perl in testsuite, there is
425 too much breakage outside AnyEvent.
426 - deliver signals synchronously in AnyEvent::Impl::Perl
427 or any other event loop that relies on AnyEvents child
428 watcher emulation. *Could* help with hanging testsuite
429 (except when Event or EV are installed).
430
431 2.6 Fri Nov 9 20:36:35 CET 2007
432 - fix bug in testsuite.
433 - move EV adaptor modules to AnyEvent.
434 - add Coro+EV adaptor module.
435
436 2.55 Tue Nov 6 17:41:32 CET 2007
437 - add EV to the list of supported event models.
438 - do not auto-reset pid watchers, pass pid and status to them.
439 - allow a pid of zero to watch for all children in child watchers.
440
441 2.54 Wed Jul 18 17:36:23 CEST 2007
442 - work around a perl bug that results in
443 BEGIN not safe after errors--compilation aborted without
444 any discernible reason or error message by once
445 again not use'ing strict.
446
447 2.53 Sun Jul 8 10:51:53 CEST 2007
448 - make Glib interface work again (spotted by elmex).
449
450 2.52 Wed Mar 7 18:36:16 CET 2007
451 - child watchers stopped working permanently when
452 all current child watchers were destroyed.
453
454 2.51 Mon Dec 11 21:33:24 CET 2006
455 - work around bugs in perl where eval "require Module" returns
456 true even if the module couldn't be loaded :/.
457
458 2.5 Mon Dec 11 02:15:28 CET 2006
459 - avoid $AUTOLOAD because many perls corrupt it.
460 - AnyEvent::detect forces autodetection.
461 - implement signal watchers (experimental)
462 (Tk does not support async signals (<= 804.027 at least),
463 so they might get delayed indefinitely on Tk).
464 - implement child watchers (experimental).
465 - moved default condvar implementation into base module, simplifying
466 most implementation modules.
467
468 2.1 Fri Nov 24 15:50:48 CET 2006
469 - better docs.
470 - simple testsuite.
471 - added AnyEvent->one_event method for special purposes.
472
473 2.0 Wed Nov 1 02:21:30 CET 2006
474 - INCOMPATIBLE CHANGE: poll can now either be "r" or "w" but not both.
475 - INCOMPATIBLE CHANGE: io watcher callbacks have no arguments anymore.
476 - EXPERIMENTAL:
477 new pure-perl model added, AnyEvent now always finds a suitable
478 event model.
479 - improved documentation.
480 - reduced cpu overhead.
481
482 1.02 Fri Jan 13 14:15:40 CET 2006
483 - add COPYING to clarify license.
484
485 1.01 Sun Jan 8 05:48:09 CET 2006
486 - correctly cancel Event timers, avoid memory leaks.
487
488 1.0 Sun Jan 8 05:40:58 CET 2006
489 - allow third-party interfaces to be added at runtime.
490 - fix version number.
491
492 0.9 Fri Dec 30 02:25:41 CET 2005
493 - fixed errors in the documentation.
494 - going back to using $VERSION as "load check",
495 as suggested by Jörn Reder. Let's see why I stopped using
496 this method.
497 - add 'hup' to the list of events to receive for glib
498 and generate 'rw' events for it, as some versions of
499 glib seem to require this, while this breaks other
500 (older) versions of glib.
501 - implemented PERL_ANYEVENT_VERBOSE env variable.
502
503 0.3 Sun Dec 4 10:43:33 CET 2005
504 - no changes to the code.
505 - fixed documentation.
506 - clarified some corner cases.
507 - vastly improved documentation with a long example.
508
509 0.2 Thu Dec 1 22:18:49 CET 2005
510 - check whether a valid method was called, to avoid endless loops.
511 - move condvars into their own class (cleaner, safer).
512 - don't die when some modules are not available.
513 - Tk is too broken to allow for words. Will work once Tk works.
514
515 0.1 Thu Dec 1 19:44:57 CET 2005
516 - spiced up and cleaned up for release.
517
518 0.01 Tue Sep 21 18:12:09 CEST 2004
519 - original version; copied from Convert::Scalar.
520