… | |
… | |
14 | |
14 | |
15 | AnyEvent::Handle - non-blocking I/O on file handles via AnyEvent |
15 | AnyEvent::Handle - non-blocking I/O on file handles via AnyEvent |
16 | |
16 | |
17 | =cut |
17 | =cut |
18 | |
18 | |
19 | our $VERSION = 4.232; |
19 | our $VERSION = 4.234; |
20 | |
20 | |
21 | =head1 SYNOPSIS |
21 | =head1 SYNOPSIS |
22 | |
22 | |
23 | use AnyEvent; |
23 | use AnyEvent; |
24 | use AnyEvent::Handle; |
24 | use AnyEvent::Handle; |
… | |
… | |
49 | |
49 | |
50 | This module is a helper module to make it easier to do event-based I/O on |
50 | This module is a helper module to make it easier to do event-based I/O on |
51 | filehandles. For utility functions for doing non-blocking connects and accepts |
51 | filehandles. For utility functions for doing non-blocking connects and accepts |
52 | on sockets see L<AnyEvent::Util>. |
52 | on sockets see L<AnyEvent::Util>. |
53 | |
53 | |
|
|
54 | The L<AnyEvent::Intro> tutorial contains some well-documented |
|
|
55 | AnyEvent::Handle examples. |
|
|
56 | |
54 | In the following, when the documentation refers to of "bytes" then this |
57 | In the following, when the documentation refers to of "bytes" then this |
55 | means characters. As sysread and syswrite are used for all I/O, their |
58 | means characters. As sysread and syswrite are used for all I/O, their |
56 | treatment of characters applies to this module as well. |
59 | treatment of characters applies to this module as well. |
57 | |
60 | |
58 | All callbacks will be invoked with the handle object as their first |
61 | All callbacks will be invoked with the handle object as their first |
… | |
… | |
100 | occured, such as not being able to resolve the hostname, failure to |
103 | occured, such as not being able to resolve the hostname, failure to |
101 | connect or a read error. |
104 | connect or a read error. |
102 | |
105 | |
103 | Some errors are fatal (which is indicated by C<$fatal> being true). On |
106 | Some errors are fatal (which is indicated by C<$fatal> being true). On |
104 | fatal errors the handle object will be shut down and will not be usable |
107 | fatal errors the handle object will be shut down and will not be usable |
105 | (but you are free to look at the current C< ->rbuf >). Examples of fatal |
108 | (but you are free to look at the current C<< ->rbuf >>). Examples of fatal |
106 | errors are an EOF condition with active (but unsatisifable) read watchers |
109 | errors are an EOF condition with active (but unsatisifable) read watchers |
107 | (C<EPIPE>) or I/O errors. |
110 | (C<EPIPE>) or I/O errors. |
108 | |
111 | |
109 | Non-fatal errors can be retried by simply returning, but it is recommended |
112 | Non-fatal errors can be retried by simply returning, but it is recommended |
110 | to simply ignore this parameter and instead abondon the handle object |
113 | to simply ignore this parameter and instead abondon the handle object |
… | |
… | |
149 | =item timeout => $fractional_seconds |
152 | =item timeout => $fractional_seconds |
150 | |
153 | |
151 | If non-zero, then this enables an "inactivity" timeout: whenever this many |
154 | If non-zero, then this enables an "inactivity" timeout: whenever this many |
152 | seconds pass without a successful read or write on the underlying file |
155 | seconds pass without a successful read or write on the underlying file |
153 | handle, the C<on_timeout> callback will be invoked (and if that one is |
156 | handle, the C<on_timeout> callback will be invoked (and if that one is |
154 | missing, an C<ETIMEDOUT> error will be raised). |
157 | missing, a non-fatal C<ETIMEDOUT> error will be raised). |
155 | |
158 | |
156 | Note that timeout processing is also active when you currently do not have |
159 | Note that timeout processing is also active when you currently do not have |
157 | any outstanding read or write requests: If you plan to keep the connection |
160 | any outstanding read or write requests: If you plan to keep the connection |
158 | idle then you should disable the timout temporarily or ignore the timeout |
161 | idle then you should disable the timout temporarily or ignore the timeout |
159 | in the C<on_timeout> callback. |
162 | in the C<on_timeout> callback, in which case AnyEvent::Handle will simply |
|
|
163 | restart the timeout. |
160 | |
164 | |
161 | Zero (the default) disables this timeout. |
165 | Zero (the default) disables this timeout. |
162 | |
166 | |
163 | =item on_timeout => $cb->($handle) |
167 | =item on_timeout => $cb->($handle) |
164 | |
168 | |
… | |
… | |
168 | |
172 | |
169 | =item rbuf_max => <bytes> |
173 | =item rbuf_max => <bytes> |
170 | |
174 | |
171 | If defined, then a fatal error will be raised (with C<$!> set to C<ENOSPC>) |
175 | If defined, then a fatal error will be raised (with C<$!> set to C<ENOSPC>) |
172 | when the read buffer ever (strictly) exceeds this size. This is useful to |
176 | when the read buffer ever (strictly) exceeds this size. This is useful to |
173 | avoid denial-of-service attacks. |
177 | avoid some forms of denial-of-service attacks. |
174 | |
178 | |
175 | For example, a server accepting connections from untrusted sources should |
179 | For example, a server accepting connections from untrusted sources should |
176 | be configured to accept only so-and-so much data that it cannot act on |
180 | be configured to accept only so-and-so much data that it cannot act on |
177 | (for example, when expecting a line, an attacker could send an unlimited |
181 | (for example, when expecting a line, an attacker could send an unlimited |
178 | amount of data without a callback ever being called as long as the line |
182 | amount of data without a callback ever being called as long as the line |
179 | isn't finished). |
183 | isn't finished). |
180 | |
184 | |
181 | =item autocork => <boolean> |
185 | =item autocork => <boolean> |
182 | |
186 | |
183 | When disabled (the default), then C<push_write> will try to immediately |
187 | When disabled (the default), then C<push_write> will try to immediately |
184 | write the data to the handle if possible. This avoids having to register |
188 | write the data to the handle, if possible. This avoids having to register |
185 | a write watcher and wait for the next event loop iteration, but can be |
189 | a write watcher and wait for the next event loop iteration, but can |
186 | inefficient if you write multiple small chunks (this disadvantage is |
190 | be inefficient if you write multiple small chunks (on the wire, this |
187 | usually avoided by your kernel's nagle algorithm, see C<low_delay>). |
191 | disadvantage is usually avoided by your kernel's nagle algorithm, see |
|
|
192 | C<no_delay>, but this option can save costly syscalls). |
188 | |
193 | |
189 | When enabled, then writes will always be queued till the next event loop |
194 | When enabled, then writes will always be queued till the next event loop |
190 | iteration. This is efficient when you do many small writes per iteration, |
195 | iteration. This is efficient when you do many small writes per iteration, |
191 | but less efficient when you do a single write only. |
196 | but less efficient when you do a single write only per iteration (or when |
|
|
197 | the write buffer often is full). It also increases write latency. |
192 | |
198 | |
193 | =item no_delay => <boolean> |
199 | =item no_delay => <boolean> |
194 | |
200 | |
195 | When doing small writes on sockets, your operating system kernel might |
201 | When doing small writes on sockets, your operating system kernel might |
196 | wait a bit for more data before actually sending it out. This is called |
202 | wait a bit for more data before actually sending it out. This is called |
197 | the Nagle algorithm, and usually it is beneficial. |
203 | the Nagle algorithm, and usually it is beneficial. |
198 | |
204 | |
199 | In some situations you want as low a delay as possible, which cna be |
205 | In some situations you want as low a delay as possible, which can be |
200 | accomplishd by setting this option to true. |
206 | accomplishd by setting this option to a true value. |
201 | |
207 | |
202 | The default is your opertaing system's default behaviour, this option |
208 | The default is your opertaing system's default behaviour (most likely |
203 | explicitly enables or disables it, if possible. |
209 | enabled), this option explicitly enables or disables it, if possible. |
204 | |
210 | |
205 | =item read_size => <bytes> |
211 | =item read_size => <bytes> |
206 | |
212 | |
207 | The default read block size (the amount of bytes this module will try to read |
213 | The default read block size (the amount of bytes this module will |
208 | during each (loop iteration). Default: C<8192>. |
214 | try to read during each loop iteration, which affects memory |
|
|
215 | requirements). Default: C<8192>. |
209 | |
216 | |
210 | =item low_water_mark => <bytes> |
217 | =item low_water_mark => <bytes> |
211 | |
218 | |
212 | Sets the amount of bytes (default: C<0>) that make up an "empty" write |
219 | Sets the amount of bytes (default: C<0>) that make up an "empty" write |
213 | buffer: If the write reaches this size or gets even samller it is |
220 | buffer: If the write reaches this size or gets even samller it is |
214 | considered empty. |
221 | considered empty. |
215 | |
222 | |
|
|
223 | Sometimes it can be beneficial (for performance reasons) to add data to |
|
|
224 | the write buffer before it is fully drained, but this is a rare case, as |
|
|
225 | the operating system kernel usually buffers data as well, so the default |
|
|
226 | is good in almost all cases. |
|
|
227 | |
216 | =item linger => <seconds> |
228 | =item linger => <seconds> |
217 | |
229 | |
218 | If non-zero (default: C<3600>), then the destructor of the |
230 | If non-zero (default: C<3600>), then the destructor of the |
219 | AnyEvent::Handle object will check wether there is still outstanding write |
231 | AnyEvent::Handle object will check whether there is still outstanding |
220 | data and will install a watcher that will write out this data. No errors |
232 | write data and will install a watcher that will write this data to the |
221 | will be reported (this mostly matches how the operating system treats |
233 | socket. No errors will be reported (this mostly matches how the operating |
222 | outstanding data at socket close time). |
234 | system treats outstanding data at socket close time). |
223 | |
235 | |
224 | This will not work for partial TLS data that could not yet been |
236 | This will not work for partial TLS data that could not be encoded |
225 | encoded. This data will be lost. |
237 | yet. This data will be lost. |
226 | |
238 | |
227 | =item tls => "accept" | "connect" | Net::SSLeay::SSL object |
239 | =item tls => "accept" | "connect" | Net::SSLeay::SSL object |
228 | |
240 | |
229 | When this parameter is given, it enables TLS (SSL) mode, that means it |
241 | When this parameter is given, it enables TLS (SSL) mode, that means |
230 | will start making tls handshake and will transparently encrypt/decrypt |
242 | AnyEvent will start a TLS handshake as soon as the conenction has been |
231 | data. |
243 | established and will transparently encrypt/decrypt data afterwards. |
232 | |
244 | |
233 | TLS mode requires Net::SSLeay to be installed (it will be loaded |
245 | TLS mode requires Net::SSLeay to be installed (it will be loaded |
234 | automatically when you try to create a TLS handle). |
246 | automatically when you try to create a TLS handle): this module doesn't |
|
|
247 | have a dependency on that module, so if your module requires it, you have |
|
|
248 | to add the dependency yourself. |
235 | |
249 | |
236 | For the TLS server side, use C<accept>, and for the TLS client side of a |
250 | Unlike TCP, TLS has a server and client side: for the TLS server side, use |
237 | connection, use C<connect> mode. |
251 | C<accept>, and for the TLS client side of a connection, use C<connect> |
|
|
252 | mode. |
238 | |
253 | |
239 | You can also provide your own TLS connection object, but you have |
254 | You can also provide your own TLS connection object, but you have |
240 | to make sure that you call either C<Net::SSLeay::set_connect_state> |
255 | to make sure that you call either C<Net::SSLeay::set_connect_state> |
241 | or C<Net::SSLeay::set_accept_state> on it before you pass it to |
256 | or C<Net::SSLeay::set_accept_state> on it before you pass it to |
242 | AnyEvent::Handle. |
257 | AnyEvent::Handle. |
243 | |
258 | |
244 | See the C<starttls> method if you need to start TLS negotiation later. |
259 | See the C<< ->starttls >> method for when need to start TLS negotiation later. |
245 | |
260 | |
246 | =item tls_ctx => $ssl_ctx |
261 | =item tls_ctx => $ssl_ctx |
247 | |
262 | |
248 | Use the given Net::SSLeay::CTX object to create the new TLS connection |
263 | Use the given C<Net::SSLeay::CTX> object to create the new TLS connection |
249 | (unless a connection object was specified directly). If this parameter is |
264 | (unless a connection object was specified directly). If this parameter is |
250 | missing, then AnyEvent::Handle will use C<AnyEvent::Handle::TLS_CTX>. |
265 | missing, then AnyEvent::Handle will use C<AnyEvent::Handle::TLS_CTX>. |
251 | |
266 | |
252 | =item json => JSON or JSON::XS object |
267 | =item json => JSON or JSON::XS object |
253 | |
268 | |
254 | This is the json coder object used by the C<json> read and write types. |
269 | This is the json coder object used by the C<json> read and write types. |
255 | |
270 | |
256 | If you don't supply it, then AnyEvent::Handle will create and use a |
271 | If you don't supply it, then AnyEvent::Handle will create and use a |
257 | suitable one, which will write and expect UTF-8 encoded JSON texts. |
272 | suitable one (on demand), which will write and expect UTF-8 encoded JSON |
|
|
273 | texts. |
258 | |
274 | |
259 | Note that you are responsible to depend on the JSON module if you want to |
275 | Note that you are responsible to depend on the JSON module if you want to |
260 | use this functionality, as AnyEvent does not have a dependency itself. |
276 | use this functionality, as AnyEvent does not have a dependency itself. |
261 | |
277 | |
262 | =item filter_r => $cb |
278 | =item filter_r => $cb |
263 | |
279 | |
264 | =item filter_w => $cb |
280 | =item filter_w => $cb |
265 | |
281 | |
266 | These exist, but are undocumented at this time. |
282 | These exist, but are undocumented at this time. (They are used internally |
|
|
283 | by the TLS code). |
267 | |
284 | |
268 | =back |
285 | =back |
269 | |
286 | |
270 | =cut |
287 | =cut |
271 | |
288 | |
… | |
… | |
324 | } |
341 | } |
325 | } |
342 | } |
326 | |
343 | |
327 | =item $fh = $handle->fh |
344 | =item $fh = $handle->fh |
328 | |
345 | |
329 | This method returns the file handle of the L<AnyEvent::Handle> object. |
346 | This method returns the file handle used to create the L<AnyEvent::Handle> object. |
330 | |
347 | |
331 | =cut |
348 | =cut |
332 | |
349 | |
333 | sub fh { $_[0]{fh} } |
350 | sub fh { $_[0]{fh} } |
334 | |
351 | |
… | |
… | |
352 | $_[0]{on_eof} = $_[1]; |
369 | $_[0]{on_eof} = $_[1]; |
353 | } |
370 | } |
354 | |
371 | |
355 | =item $handle->on_timeout ($cb) |
372 | =item $handle->on_timeout ($cb) |
356 | |
373 | |
357 | Replace the current C<on_timeout> callback, or disables the callback |
374 | Replace the current C<on_timeout> callback, or disables the callback (but |
358 | (but not the timeout) if C<$cb> = C<undef>. See C<timeout> constructor |
375 | not the timeout) if C<$cb> = C<undef>. See the C<timeout> constructor |
359 | argument. |
376 | argument and method. |
360 | |
377 | |
361 | =cut |
378 | =cut |
362 | |
379 | |
363 | sub on_timeout { |
380 | sub on_timeout { |
364 | $_[0]{on_timeout} = $_[1]; |
381 | $_[0]{on_timeout} = $_[1]; |
… | |
… | |
1362 | # basically, this is deep magic (because SSL_read should have the same issues) |
1379 | # basically, this is deep magic (because SSL_read should have the same issues) |
1363 | # but the openssl maintainers basically said: "trust us, it just works". |
1380 | # but the openssl maintainers basically said: "trust us, it just works". |
1364 | # (unfortunately, we have to hardcode constants because the abysmally misdesigned |
1381 | # (unfortunately, we have to hardcode constants because the abysmally misdesigned |
1365 | # and mismaintained ssleay-module doesn't even offer them). |
1382 | # and mismaintained ssleay-module doesn't even offer them). |
1366 | # http://www.mail-archive.com/openssl-dev@openssl.org/msg22420.html |
1383 | # http://www.mail-archive.com/openssl-dev@openssl.org/msg22420.html |
|
|
1384 | # |
|
|
1385 | # in short: this is a mess. |
|
|
1386 | # |
|
|
1387 | # note that we do not try to kepe the length constant between writes as we are required to do. |
|
|
1388 | # we assume that most (but not all) of this insanity only applies to non-blocking cases, |
|
|
1389 | # and we drive openssl fully in blocking mode here. |
1367 | Net::SSLeay::CTX_set_mode ($self->{tls}, |
1390 | Net::SSLeay::CTX_set_mode ($self->{tls}, |
1368 | (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ENABLE_PARTIAL_WRITE () } || 1) |
1391 | (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ENABLE_PARTIAL_WRITE () } || 1) |
1369 | | (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ACCEPT_MOVING_WRITE_BUFFER () } || 2)); |
1392 | | (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ACCEPT_MOVING_WRITE_BUFFER () } || 2)); |
1370 | |
1393 | |
1371 | $self->{_rbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ()); |
1394 | $self->{_rbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ()); |