The MySQL client libraries normally retry system calls when
interrupted by signals. The lone exception I've found is in the
(infrequent) connection setup where it'll propagate the
connect(2) syscall error all the way back up to the caller.
Fortunately inspection of the MySQL client library reveals it
properly preserves the global "errno" variable even with
debugging enabled.
Note that the net.last_errno member does NOT correspond to the
system "errno".
This reverts the single-threaded select() optimization from
commits 9a63a587c0 and
0190457dbd.
Under Ruby 1.9, the reverted optimization caused signal handlers
to be delayed until the socket became readable.
Under Ruby 1.8, matters are worse as receiving a signal during
select() causes Errno::EINTR to be raised.
There is now a (somewhat fragile) spec for testing signal
handling during a "SELECT sleep(2)" query.
Furthermore, the performance difference I measured on
benchmark/thread_alone.rb was negligible (over 1000 iterations)
between the two:
Ruby 1.8.7-p249
Rehearsal ----------------------------------------------------
select 0.040000 0.020000 0.060000 ( 0.119448)
rb_thread_select 0.050000 0.020000 0.070000 ( 0.132091)
------------------------------------------- total: 0.130000sec
user system total real
select 0.030000 0.030000 0.060000 ( 0.116471)
rb_thread_select 0.050000 0.020000 0.070000 ( 0.119874)
Ruby 1.9.2-p0
Rehearsal ----------------------------------------------------
select 0.050000 0.030000 0.080000 ( 0.134208)
rb_thread_select 0.080000 0.000000 0.080000 ( 0.141316)
------------------------------------------- total: 0.160000sec
user system total real
select 0.050000 0.020000 0.070000 ( 0.123325)
rb_thread_select 0.060000 0.010000 0.070000 ( 0.124075)
Benchmarks were performed on an _empty_ mysql2_test table to
maximize the relative time for the select(2)-related code path
(vs reading data). The test was run on Debian Lenny, x86_64 and
a stock 2.6.35.2 Linux kernel. Hardware was a Core2 Duo @
1.6GHz with the performance CPU governor while on AC power
to eliminate CPU speed fluctuations during the test.
1) let mysql_init/mysql_close take care of any/all allocation, thread state and freeing
2) for faster access to the encoding and active state variables for the connection
* master:
remove Sequel adapter as it's now in Sequel core :)
move -Wextra to development flags area
update AR adapter to reflect timezone setting update
application_timezone is allowed to be nil
default application_timezone to nil
sync up with sequel adapter from my Sequel fork until it's officially merged in
convert :timezone option into two new ones :database_timezone - the timezone (:utc or :local) Mysql2 will assume time/datetime fields are stored in the db. This modifies what initial timezone your Time objects will be in when creating them from libmysql in C and :application_timezone - the timezone (:utc or :local) you'd finally like the Time objects converted to before you get them
can't call literal here because it'll try to join it's own thread
Mysql2::Client uses the :username key, set it to :user if that was used instead
heh
fix typo in comment
major refactor of Sequel adapter - it's now green in Sequel
add :cast_booleans option for automatically casting tinyint(1) fields into true/false for ruby
move most previously global symbols to static to prevent conflicts (thanks for catching this Eric)
respect :symbolize_keys option for Mysql2::Result#fields if it's called before the first row is built
initialize @active early on to prevent warnings later
let's try that again - libmysql only allows one query be sent at a time per connection, bail early if that's attempted
Revert "libmysql only allows one query be sent at a time per connection, bail early if that's attempted"
libmysql only allows one query be sent at a time per connection, bail early if that's attempted
no need to carry over options twice as we're already doing it up in rb_mysql_client_async_result
