| 1 | Revision history for AnyEvent::DBI |
1 | Revision history for AnyEvent::DBI |
| 2 | |
2 | |
| 3 | TODO: use util for fd_setfd |
3 | 3.03 Thu Apr 19 06:24:14 CEST 2018 |
|
|
4 | - the main process socket was not actually set to nonblocking mode, |
|
|
5 | causing lockups under load (reported by Felix Ostmann). |
|
|
6 | |
|
|
7 | 3.02 Mon Sep 4 13:45:56 CEST 2017 |
|
|
8 | - fixed module version mismatch (reported by Martin Barth). |
|
|
9 | |
|
|
10 | 3.01 Sun Aug 27 11:53:26 CEST 2017 |
|
|
11 | - add missing dependency to AnyEvent::Fork. |
|
|
12 | |
|
|
13 | 3.0 Tue Aug 15 09:05:30 CEST 2017 |
|
|
14 | - the exec_server constructor argument is now ignored, fork vs. exec |
|
|
15 | is now decided by availability of AnyEvent::Fork::Early or not. |
|
|
16 | - experimental support for fetching (but not setting) statement handle |
|
|
17 | attributes: ->stattr. |
|
|
18 | - use AnyEvent::Fork for process management. |
|
|
19 | - switch to CBOR::XS and Convert::Scalar for further speed and memory |
|
|
20 | improvements - a loop doing "select 0" is now almost twice as fast, |
|
|
21 | more impressive speedups are expected for larger results. |
|
|
22 | This also considerably simplified the code. |
|
|
23 | - t/02_sql_lite.t wrongly assumed sqlite doesn't implement instr(), but |
|
|
24 | nowadays, it does. |
|
|
25 | - added stability canary support. |
|
|
26 | - no longer load Scalar::Util, Fcntl, Socket and POSIX modules. |
|
|
27 | - move child code into it's own module, AnyEvent::DBI::Slave. |
|
|
28 | - no longer load DBI (except in slaves). |
|
|
29 | |
|
|
30 | 2.3 Tue Apr 2 02:59:12 CEST 2013 |
|
|
31 | - fix begin_work, commit and rollback, that were broken due |
|
|
32 | to a precedence problem (analyzed by Jerry Lundström). |
|
|
33 | - slightly improve tests. |
|
|
34 | |
|
|
35 | 2.2 Thu May 17 04:13:27 CEST 2012 |
|
|
36 | - timeouts didn't trigger (Adam Rosenstein). |
|
|
37 | - do not leave zombie processes (Adam Rosenstein). |
|
|
38 | |
|
|
39 | 2.1 Sat Oct 30 22:22:05 CEST 2010 |
|
|
40 | - no longer kill -9 the child, just kill -TERM and close the connection, |
|
|
41 | to avoid database corruption - this also works around a perl bug |
|
|
42 | (global destruction). |
|
|
43 | - convert to the faster AE API - major speedup. |
|
|
44 | - use common::sense. |
| 4 | |
45 | |
| 5 | 2.0 Mon Jun 29 10:18:58 CEST 2009 |
46 | 2.0 Mon Jun 29 10:18:58 CEST 2009 |
| 6 | - INCOMPATIBLE CHANGE: callbacks now are passed the $dbh |
47 | - INCOMPATIBLE CHANGE: callbacks now are passed the $dbh |
| 7 | as first argument, always. |
48 | as first argument, always. |
| 8 | - INCOMPATIBLE CHANGE: really pass rv as third argument to exec |
49 | - INCOMPATIBLE CHANGE: really pass rv as third argument to exec |
| 9 | callbacks, as documented but not implemented for 1.1. |
50 | callbacks, as documented but not implemented for 1.1. |
| 10 | - thanks to viacheslav.t for testing. |
51 | - thanks to viacheslav.t for testing. |
| 11 | - major patch by Adam Rosenstein: |
52 | - major patch by Adam Rosenstein: |
| 12 | - add an on_connect callback (for reliable failure notification) |
53 | - add an on_connect callback (for reliable failure notification) |
| 13 | - add timeouts to connect and exec. |
54 | - add timeouts to connect and exec. |
| 14 | - reaps server processes if you undef your AE::DBI handle and keep |
55 | - reaps server processes if you undef your AE::DBI handle and keep |
| 15 | running. |
56 | running. |
| 16 | - option to start the server via exec, otherwise try to emulate |
57 | - option to start the server via exec, otherwise try to emulate |
| 17 | close-on-exec behavior. |
58 | close-on-exec behavior. |
