1 |
root |
1.1 |
=head1 NAME |
2 |
|
|
|
3 |
|
|
AnyEvent::Impl::POE - AnyEvent adaptor for POE |
4 |
|
|
|
5 |
|
|
=head1 SYNOPSIS |
6 |
|
|
|
7 |
|
|
use AnyEvent; |
8 |
root |
1.2 |
use POE; |
9 |
root |
1.1 |
|
10 |
|
|
# this module gets loaded automatically as required |
11 |
|
|
|
12 |
|
|
=head1 DESCRIPTION |
13 |
|
|
|
14 |
|
|
This module provides transparent support for AnyEvent. You don't have to |
15 |
|
|
do anything to make POE work with AnyEvent except by loading POE before |
16 |
|
|
creating the first AnyEvent watcher. |
17 |
|
|
|
18 |
root |
1.3 |
Unfortunately, POE isn't generic enough to implement a fully working |
19 |
|
|
AnyEvent backend: POE is too badly designed, too badly documented and too |
20 |
|
|
badly implemented. |
21 |
root |
1.2 |
|
22 |
root |
1.3 |
Here are the details, and what it means to you if you want to be |
23 |
|
|
interoperable with POE: |
24 |
root |
1.2 |
|
25 |
|
|
=over 4 |
26 |
|
|
|
27 |
|
|
=item Weird messages |
28 |
|
|
|
29 |
root |
1.4 |
If you only use C<run_one_timeslice> (as AnyEvent has to for it's |
30 |
|
|
condition variables), POE will print an ugly, unsupressable, message at |
31 |
|
|
program exit: |
32 |
root |
1.2 |
|
33 |
|
|
Sessions were started, but POE::Kernel's run() method was never... |
34 |
|
|
|
35 |
|
|
The message is correct, the question is why POE prints it in the first |
36 |
|
|
place in a correct program (this is not a singular case though). |
37 |
|
|
|
38 |
|
|
The only way I found to work around this bug was to call C<< |
39 |
root |
1.5 |
->run >> at AnyEvent loading time and stop the kernel immediately |
40 |
root |
1.2 |
again. Unfortunately, due to another design bug in POE, this cannot be |
41 |
|
|
done (by documented means at least) without throwing away events in the |
42 |
|
|
event queue. |
43 |
|
|
|
44 |
|
|
This means that you will either have to live with lost events or you have |
45 |
|
|
to make sure to load AnyEvent early enough (this is usually not that |
46 |
|
|
difficult in a main program, but hard in a module). |
47 |
|
|
|
48 |
root |
1.7 |
=item One POE session per Event |
49 |
root |
1.2 |
|
50 |
|
|
AnyEvent has to create one POE::Session per event watcher, which is |
51 |
root |
1.3 |
immensely slow and makes watchers very large. The reason for this is |
52 |
|
|
lacking lifetime management (mostly undocumented, too). Without one |
53 |
|
|
session/watcher it is not possible to easily keep the kernel from running |
54 |
|
|
endlessly. |
55 |
root |
1.2 |
|
56 |
|
|
=item One watcher per fd/event combo |
57 |
|
|
|
58 |
|
|
POE, of course, suffers from the same bug as Tk and some other badly |
59 |
|
|
designed event models in that it doesn't support multiple watchers per |
60 |
root |
1.3 |
fd/poll combo. The workaround is the same as with Tk: AnyEvent::Impl::POE |
61 |
|
|
creates a separate file descriptor to hand to POE, which isn't fast and |
62 |
|
|
certainly not nice to your resources. |
63 |
root |
1.2 |
|
64 |
root |
1.7 |
Of course, without the workaround, POE also prints ugly messages again |
65 |
|
|
that say the program *might* be buggy. |
66 |
|
|
|
67 |
root |
1.3 |
=item Timing Deficiencies |
68 |
|
|
|
69 |
|
|
POE manages to not have a function that returns the current time. This is |
70 |
|
|
extremely problematic, as POE can use different time functions, which can |
71 |
|
|
differ by more than a second. In addition, most timer functions in POE |
72 |
|
|
want an absoltue timestamp, which is hard to create if all you have is a |
73 |
|
|
relative time and no function to return the "current time". |
74 |
|
|
|
75 |
|
|
AnyEvent works around this by using relative timer fucntions, in the hope |
76 |
|
|
that POE gets it right at least internally. |
77 |
|
|
|
78 |
|
|
=item Event Non-Ordering |
79 |
|
|
|
80 |
|
|
POE cannot guarentee the order of callback invocation for timers, and |
81 |
|
|
usually gets it wrong. That is, if you have two timers, one timing out |
82 |
|
|
after another, the callbacks might be called in reverse order. |
83 |
|
|
|
84 |
|
|
How one manages to even implement stuff that way escapes me. |
85 |
|
|
|
86 |
|
|
=item Child Watchers |
87 |
|
|
|
88 |
root |
1.7 |
POE offers child watchers - which is a laudable thing, as few event loops |
89 |
root |
1.3 |
do. Unfortunately, they cannot even implement AnyEvent's simple child |
90 |
|
|
watchers: they are not generic enough. |
91 |
|
|
|
92 |
root |
1.6 |
Of course, if POE reaps an unrelated child it will also output a message |
93 |
|
|
for it. Very professional. |
94 |
|
|
|
95 |
root |
1.7 |
As a workaround, AnyEvent::Impl::POE will take advantage of undocumented |
96 |
|
|
behaviour in POE::Kernel to catch the status fo all child processes. |
97 |
|
|
|
98 |
|
|
Unfortunately, POE's child handling is racy: if the child exits before the |
99 |
|
|
handler is created (which is impossible to guarantee...), one has to wait |
100 |
|
|
for another event to occur, which can take an indefinite time (apparently |
101 |
|
|
POE does a busy-waiting loop every second, but this is not guarenteed |
102 |
|
|
or documented, so in practise child status events can be delayed for a |
103 |
|
|
second). |
104 |
|
|
|
105 |
|
|
How one manages to have such a glaring bug in an event loop after ten |
106 |
|
|
years of development escapes me. |
107 |
root |
1.3 |
|
108 |
|
|
=item Documentation Quality |
109 |
|
|
|
110 |
|
|
At the time of this writing, POE was in its tenth year. Still, its |
111 |
|
|
documentation is extremely lacking, making it impossible to implement |
112 |
root |
1.10 |
stuff as trivial as AnyEvent watchers without having to resort to |
113 |
root |
1.3 |
undocumented behaviour or features. |
114 |
|
|
|
115 |
|
|
For example, the POE::Kernel manpage has nice occurances of the word TODO |
116 |
root |
1.10 |
with an explanation of whats missing. In general, the POE manpages are |
117 |
|
|
littered with comments like "section not yet written". |
118 |
|
|
|
119 |
|
|
Some other gems: |
120 |
root |
1.3 |
|
121 |
|
|
This allows many object methods to also be package methods. |
122 |
|
|
|
123 |
|
|
This is nice, but since it doesn't document I<which> methods these are, |
124 |
|
|
this is utterly useless information. |
125 |
|
|
|
126 |
|
|
Terminal signals will kill sessions if they are not handled by a |
127 |
root |
1.10 |
"sig_handled"() call. The OS signals that usually kill or dump a |
128 |
|
|
process are considered terminal in POE, but they never trigger a |
129 |
root |
1.3 |
coredump. These are: HUP, INT, QUIT and TERM. |
130 |
|
|
|
131 |
root |
1.10 |
Although AnyEvent calls C<sig_handled>, removing it has no apparent |
132 |
|
|
effects on POE handling SIGINT. |
133 |
root |
1.3 |
|
134 |
|
|
Furthermore, since the Kernel keeps track of everything sessions do, it |
135 |
|
|
knows when a session has run out of tasks to perform. |
136 |
|
|
|
137 |
|
|
This is impossible - how does the kernel now that a session is no longer |
138 |
|
|
watching for some (external) event (e.g. by some other session)? It |
139 |
root |
1.10 |
cannot, and therefore this is wrong - but you would be hard pressed to |
140 |
|
|
find out how to work around this and tell the kernel manually about such |
141 |
|
|
events. |
142 |
root |
1.3 |
|
143 |
|
|
It gets worse, though - the notion of "task" or "resource", although used |
144 |
|
|
throughout the documentation, is not defined in a usable way. For example, |
145 |
|
|
waiting for a timeout is considered to be a task, waiting for a signal is |
146 |
root |
1.10 |
not (a session that only waits for a signal is considered finished and |
147 |
|
|
gets removed). The user is left guessing when waiting for an event counts |
148 |
|
|
as task and when not. |
149 |
root |
1.3 |
|
150 |
root |
1.10 |
One could go on endlessly - ten years, no usable documentation. |
151 |
root |
1.3 |
|
152 |
root |
1.8 |
It is likely that differences between documentation, or the one or two |
153 |
root |
1.10 |
things I had to guess, cause unanticipated problems with this adaptor. |
154 |
root |
1.3 |
|
155 |
|
|
=item Bad API |
156 |
|
|
|
157 |
|
|
The POE API is extremely inconsistent - sometimes you have to pass a |
158 |
|
|
session argument, sometimes it gets ignored, sometimes a session-specific |
159 |
|
|
method must not use a session argument. |
160 |
root |
1.2 |
|
161 |
root |
1.10 |
Error handling is sub-standard as well: even for programming mistakes, |
162 |
|
|
POE does not C<croak> but, in most cases, just sets C<$!> or simply does |
163 |
|
|
nothing at all, leading to fragile programs. |
164 |
|
|
|
165 |
|
|
Sometimes registering a handler uses the "eventname, parameter" ordering |
166 |
|
|
(timeouts), sometimes it is "parameter, eventname" (signals). There is |
167 |
|
|
little consistency overall. |
168 |
root |
1.1 |
|
169 |
root |
1.3 |
=back |
170 |
root |
1.1 |
|
171 |
root |
1.9 |
On the good side, AnyEvent allows you to write your modules in a 100% |
172 |
|
|
POE-compatible way (bug-for-bug compatible even), without forcing your |
173 |
root |
1.3 |
module to use POE - it is still open to better event models, of which |
174 |
|
|
there are plenty. |
175 |
root |
1.1 |
|
176 |
|
|
=cut |
177 |
|
|
|
178 |
|
|
package AnyEvent::Impl::POE; |
179 |
|
|
|
180 |
|
|
no warnings; |
181 |
|
|
use strict; |
182 |
|
|
|
183 |
|
|
use POE; |
184 |
|
|
|
185 |
|
|
# have to do this to keep POE from spilling ugly messages |
186 |
|
|
POE::Session->create (inline_states => { _start => sub { @_[KERNEL]->stop } }); |
187 |
|
|
POE::Kernel->run; |
188 |
|
|
|
189 |
|
|
sub io { |
190 |
|
|
my ($class, %arg) = @_; |
191 |
|
|
my $poll = delete $arg{poll}; |
192 |
|
|
my $cb = delete $arg{cb}; |
193 |
root |
1.2 |
|
194 |
|
|
# cygwin requires the fh mode to be matching, unix doesn't |
195 |
|
|
my ($pee, $mode) = $poll eq "r" ? ("select_read" , "<") |
196 |
|
|
: $poll eq "w" ? ("select_write", ">") |
197 |
|
|
: Carp::croak "AnyEvent->io requires poll set to either 'r' or 'w'"; |
198 |
|
|
|
199 |
|
|
open my $fh, "$mode&" . fileno $arg{fh} |
200 |
|
|
or die "cannot dup() filehandle: $!"; |
201 |
|
|
|
202 |
root |
1.1 |
my $session = POE::Session->create ( |
203 |
|
|
inline_states => { |
204 |
|
|
_start => sub { |
205 |
root |
1.2 |
$_[KERNEL]->$pee ($fh => "ready"); |
206 |
root |
1.1 |
}, |
207 |
|
|
ready => sub { |
208 |
|
|
$cb->(); |
209 |
|
|
}, |
210 |
|
|
stop => sub { |
211 |
root |
1.2 |
$_[KERNEL]->$pee ($fh); |
212 |
root |
1.1 |
}, |
213 |
|
|
}, |
214 |
|
|
); |
215 |
|
|
bless \\$session, AnyEvent::Impl::POE:: |
216 |
|
|
} |
217 |
|
|
|
218 |
|
|
sub timer { |
219 |
|
|
my ($class, %arg) = @_; |
220 |
|
|
my $after = delete $arg{after}; |
221 |
|
|
my $cb = delete $arg{cb}; |
222 |
|
|
my $session = POE::Session->create ( |
223 |
|
|
inline_states => { |
224 |
|
|
_start => sub { |
225 |
|
|
$_[KERNEL]->delay_set (timeout => $after); |
226 |
|
|
}, |
227 |
|
|
timeout => sub { |
228 |
|
|
$cb->(); |
229 |
|
|
}, |
230 |
|
|
stop => sub { |
231 |
|
|
$_[KERNEL]->alarm_remove_all; |
232 |
|
|
}, |
233 |
|
|
}, |
234 |
|
|
); |
235 |
|
|
bless \\$session, AnyEvent::Impl::POE:: |
236 |
|
|
} |
237 |
|
|
|
238 |
|
|
sub signal { |
239 |
|
|
my ($class, %arg) = @_; |
240 |
|
|
my $signal = delete $arg{signal}; |
241 |
|
|
my $cb = delete $arg{cb}; |
242 |
|
|
my $session = POE::Session->create ( |
243 |
|
|
inline_states => { |
244 |
|
|
_start => sub { |
245 |
|
|
$_[KERNEL]->sig ($signal => "catch"); |
246 |
|
|
$_[KERNEL]->refcount_increment ($_[SESSION]->ID => "poe"); |
247 |
|
|
}, |
248 |
|
|
catch => sub { |
249 |
|
|
$cb->(); |
250 |
|
|
$_[KERNEL]->sig_handled; |
251 |
|
|
}, |
252 |
|
|
stop => sub { |
253 |
|
|
$_[KERNEL]->refcount_decrement ($_[SESSION]->ID => "poe"); |
254 |
|
|
$_[KERNEL]->sig ($signal); |
255 |
|
|
}, |
256 |
|
|
}, |
257 |
|
|
); |
258 |
|
|
bless \\$session, AnyEvent::Impl::POE:: |
259 |
|
|
} |
260 |
|
|
|
261 |
root |
1.7 |
sub child { |
262 |
|
|
my ($class, %arg) = @_; |
263 |
|
|
my $pid = delete $arg{pid}; |
264 |
|
|
my $cb = delete $arg{cb}; |
265 |
|
|
my $session = POE::Session->create ( |
266 |
|
|
inline_states => { |
267 |
|
|
_start => sub { |
268 |
|
|
$_[KERNEL]->sig (CHLD => "child"); |
269 |
|
|
$_[KERNEL]->refcount_increment ($_[SESSION]->ID => "poe"); |
270 |
|
|
}, |
271 |
|
|
child => sub { |
272 |
|
|
my ($rpid, $status) = @_[ARG1, ARG2]; |
273 |
|
|
|
274 |
|
|
$cb->($rpid, $status) if $rpid == $pid || $pid == 0; |
275 |
|
|
}, |
276 |
|
|
stop => sub { |
277 |
|
|
$_[KERNEL]->refcount_decrement ($_[SESSION]->ID => "poe"); |
278 |
|
|
$_[KERNEL]->sig ("CHLD"); |
279 |
|
|
}, |
280 |
|
|
}, |
281 |
|
|
); |
282 |
|
|
bless \\$session, AnyEvent::Impl::POE:: |
283 |
|
|
} |
284 |
|
|
|
285 |
root |
1.1 |
sub DESTROY { |
286 |
|
|
POE::Kernel->post (${${$_[0]}}, "stop"); |
287 |
|
|
} |
288 |
|
|
|
289 |
|
|
sub one_event { |
290 |
|
|
POE::Kernel->loop_do_timeslice; |
291 |
|
|
} |
292 |
|
|
|
293 |
|
|
1; |
294 |
|
|
|
295 |
|
|
=head1 SEE ALSO |
296 |
|
|
|
297 |
|
|
L<AnyEvent>, L<POE>. |
298 |
|
|
|
299 |
|
|
=head1 AUTHOR |
300 |
|
|
|
301 |
|
|
Marc Lehmann <schmorp@schmorp.de> |
302 |
|
|
http://home.schmorp.de/ |
303 |
|
|
|
304 |
|
|
=cut |
305 |
|
|
|