forked from jasontang/net-stomp
-
Notifications
You must be signed in to change notification settings - Fork 6
/
CHANGES
171 lines (139 loc) · 6.4 KB
/
CHANGES
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
Revision history for Perl module Net::Stomp:
0.62 2024-02-08
- "failover" URL now works together with "ssl" and "ssl_options"
constructor arguments by applying the same SSL settings to all
hosts in the failover URL.
0.61 2021-04-13
- Avoid referencing $frame if it's undefined (thanks Lee Burton
N6UDP)
0.60 2019-07-17
- switch to Log::Any
- add trace / debug logging
- work sponsored by Broadbean Technology
0.57 2016-04-17
- fix reconnection problems when server dies (thanks dhnukem)
- add NACK command (thanks Doug Henderson)
0.56 2015-06-27
- fix typo (thanks POLETTIX)
- fix boundary condition when receiving frames with content-length,
RT #105500 (thanks Dryapak Grigory)
0.55 2015-03-26
- drop "use Data::Printer" leftover from debugging
0.54 2015-03-25
- when reading times out in the middle of a frame, but the connection
is still up, keep parsing the current frame instead of getting
confused, fixes #103071 thanks to colddaylike for noticing the
issue
0.53 2015-02-18
- add receipt_timeout attribute, to make sure we give up when no
receipt comes (usually because of network problems)
- test whan happens when timing out while waiting for the receipt
- make sure to never modify the passed-in hashrefs
0.52 2015-02-16
- add socket_options attribute
- add TCP keep-alive support
0.51 2015-01-26
- fix and document return values of all methods
- add global read timeout attribute
0.50 2015-01-26
- extract send_with_receipt from inside send_transactional
0.49 2014-07-14
- allow setting SSL per-server
- expose the currently connected host
0.48 2014-07-12
- fix dependency for Test::Deep RT #97167
- better diagnostics when test modules are missing
0.47 2014-07-11
- a test suite! we now have good test coverage
- configurable reconnect delay (was fixed at 5 seconds)
- configurable logger, instead of just warn/carp
- handle partial writes, fixes RT #83047
- properly reconnect on write failure
- more robust reconnecting on read failure
- Net::Stomp::Frame::parse now actually exists (it was documented
but not implemented)
- fix RT #35194 'ack' now allow arbitrary headers
- fix RT #80079 correctly pick the first value for repeated headers
- from RT #68160, allow changing the maximum number of reconnect attempts
- from RT #68160, ignore SIGPIPE while writing to the socket
- more accessors on ::Frame, fix RT #90100
0.46 Tue Jan 14 12:21:43 GMT 2014
- Recommend modules required for IPv6 support (Dagfinn Ilmari Mannsåker)
0.45 Wed Mar 14 10:25:35 GMT 2012
- reconnect on fork() (Gianni Ceccarelli)
0.44 Wed Nov 23 23:26:39 GMT 2011
- Updated CHANGES to include recent changes (Jason Tang)
0.43 Mon Nov 21 10:46:58 GMT 2011
- rt#68583: "Use of uninitialized value" warning in as_string()(Chisel Wright)
- rt#70670: Key using id rather than destination (Gianni Ceccarelli)
0.42 Sat Aug 20 07:37:31 GMT 2011
- Improve handling of disconnecting/reconnecting (Chisel Wright)
0.41 Thu Apr 28 14:56:49 GMT 2011
- Various small warning/error message fies:
- Remove undef warning from substr (Squeeks)
- rt#67160: IO::Socket::INET/SSL Error Variables (Stephen Fralich)
- rt#65979: Fix bug in failover hosts (vigith maurice)
0.41 Fri Feb 4 09:16:39 GMT 2011
- Fix silly bug in _read_body (reported by Zulf Ahmed)
- Remove uninitialized value in numeric gt warning (reported by Dave Krieger)
0.39 Fri Jan 14 15:59:27 GMT 2011
- Cope with EOF during receive_frame better
0.38_99 Fri Aug 27 00:11:08 BST 2010
- Reimplemented Net::Stomp::receive_frame() to properly handle STOMP frames
that have extraneous line feed characters after the NULL terminator, before
the next frame header (e.g., ActiveMQ). (Implemented by Michael S. Fischer)
- Improve reading performace by reading large chunks and buffering
- Read buffer size can now be specified in Net::Stomp->new() with the
'bufsize' key.
- Possible incompatible change - Net::Stomp::Frame#parse method removed.
Hopefully no one was using this. If you were, RTFS and look at
Net::Stomp#receive_frame method.
0.38 Tue Aug 03 13:58:10 BST 2010
- Fix an issue with IO::Handle buffering interfering with IO::Select. This
manifested as receive_frame hanging, or as if not all messages would get
delivered to the client. (RT 44629)
0.37 Fri May 28 15:26:17 BST 2010
- Report a proper error if non-SSL STOMP server is not listening rather than
"Can't use an undefined value as a symbol reference".
- Fix regression in can_read: it now again defaults to waiting indefinitely.
(RT 58502)
0.36 Fri May 28 16:20:15 BST 2010
- Fix behaviour regression on constructor - 0.35 would look at ->new time if
no server could be reached. We now die after trying each server once.
0.35 Tue May 25 15:55:36 BST 2010
- add some examples
- add support for SSL (thanks to Aleksandar Ivanisevic)
- add send_transactional (based on Net::Stomp::Receipt, thanks
to Hugo Salgado)
- add some convenience methods for accessing headers in a frame
(thanks to Claes Jakobsson)
- receive_frame now accepts a {timeout=>1} option
- failover support (thanks to Thom May and Ash Berlin)
- reconnect and resubscribe when connection failes (Thom May)
0.34 Fri Jun 27 09:29:13 BST 2008
- revert to 0.32's code, as the last release broke things that
I don't have time to fix right now
0.33 Wed Jun 25 08:45:59 BST 2008
- fixes from Paul Driver: can_read doesn't behave correctly. The
docs say it returns true iff there's a frame waiting to be
read, but it's just a select() call to see if there's data
on the socket, which can return both false positives and
false negatives. While investigating this, I noticed that
Net::Stomp::Frame->parse was modified at some point to take
a socket as its argument. This is inconsistent with docs and
probably not what was wanted anyway. parse and as_string
should be symmetrical. Attached is a patch that fixes both
issues and adds tests for them.
0.32 Sun Oct 29 09:06:31 GMT 2006
- rewrote the parser to add support for BytesMessage, which should
allow Java<->Perl messages (thanks to Hiram Chirino and Sileshi
Kassa)
- set binmode() on the socket so Windows should now work (thanks
to Sileshi Kassa)
0.31 Wed Oct 11 17:28:31 BST 2006
- minor docpatch noticed by Leo Lapworth
- tweak frame parser
- add unsubscribe() method
- add can_read() method
0.30 Sat Oct 7 09:47:57 BST 2006
- initial release