Upstream changes:
0.56 2016-03-16T17:07:05Z
- Added Firebird support, thanks to Stefan Suciu.
- Fixed SQLite savepoint support to properly detect SQlite versions 3.9
and higher.
- Restored MySQL savepoint testing when the DBICTEST_* environment
variables are set.
0.55 2016-02-05T18:59:24Z
- Added versions to the RollbackError exception classes to make PAUSE
happy.
0.54 2016-02-05T18:52:44Z
- Removeed the Pod tests from the distribution.
- Fixed the example code for handling `rollback_error` exceptions in the
documentation to properly wrap the transaction in an exception
handler. Patch from Perlover (issue #32).
- Improved handling of connection failures when RaiseError (or
HandleError) is not set. Thanks to Andreas Huber for the report and
fix.
- Document that `$_` is set locally, not globally, in the methods that
set it. Suggested by William Lindley.
- The disconnect method no longer longer remove entries from the DBI
handle's CachedKids attribute. That behavior appears to be a
workaround for a database we don't (yet) support.
- Fixed a test failure where the test system has the `$DBI_DSN` or
`$DBI_DRIVER` environment variable set. Thanks to Erik Rijkers for the
patch.
- Added recommendation to use DBD::Pg 3.5.0 or later to the Pg driver.
Earlier versions had an incorrect implementation of the `ping()`
method (Issue #41).
Problems found with existing distfiles:
distfiles/D6.data.ros.gz
distfiles/cstore0.2.tar.gz
distfiles/data4.tar.gz
distfiles/sphinx-2.2.7-release.tar.gz
No changes made to the cstore or mariadb55-client distinfo files.
Otherwise, existing SHA1 digests verified and found to be the same on
the machine holding the existing distfiles (morden). All existing
SHA1 digests retained for now as an audit trail.
Do it for all packages that
* mention perl, or
* have a directory name starting with p5-*, or
* depend on a package starting with p5-
like last time, for 5.18, where this didn't lead to complaints.
Let me know if you have any this time.
a) refer 'perl' in their Makefile, or
b) have a directory name of p5-*, or
c) have any dependency on any p5-* package
Like last time, where this caused no complaints.
Upstream changes:
0.53 2013-03-20T06:04:34Z
- Fixed some documentation typos, thanks to Mike O'Regan (Issue #22).
- Fixed issue where an connection failure caused an unhelpful error
(Issue #26).
0.52 2012-05-29T18:32:42Z
- The DBI params are now encapsulated in a code reference, rather than
stored as the passed array, so that the password is less likely to be
displayed in a dump. Idea borrowed from Rose::DB. Patch from Brad
Bowman.
- Eliminated warning about the non-portability of a v-string on older
Perls. Thanks to Mark Lawrence for the report (Issue #17).
- Removed a couple of leftover examples of the `catch` feature removed
in v0.50. Thanks to Randy Stauner for the patch!
- Eliminated more "Use of qw(...) as parentheses" syntax errors in tests
when running on Perl 5.17.
- Add mention of DBI Callbacks parameter to the docs, as folks often ask
for this functionality, not realizing that the DBI already provides
it. Randy Stauner.
0.51 2012-02-18T00:14:48Z
- Fixed internal exception handling on Perls less than 5.14, where some
exceptions woult not be propagated to the caller.
0.50 2012-02-14T18:25:07Z
- The `catch` functionality has been completely removed. Any `catch`
block passed to `run()`, `txn()`, or `svp()` will be ignored. Errors
will trigger fatal exceptions.
- Removed the `with` method, which was deprecated in 0.34. Use `mode()`
instead.
- Fixed bad method call attempted when an `svp()` block failed. Thanks
to Ricardo SIGNES for the regression test and the fix.
- Fixed creation of the SQLite driver savepoint methods so that they
exist and work even if the driver is loaded before DBD::SQLite. Thanks
to Ricardo SIGNES for the regression test and the fix.
Changes from previous:
0.47 2011-09-26T17:12:07Z
- Use of the deprecated `catch` functionality now warns on every call,
rather than just the first call from a given caller.
0.46 2011-07-17T00:47:49
- Eliminated "Use of qw(...) as parentheses is deprecated" warning in
test when running on Perl 5.14.
- Properly `local`ing `$$` in the `t/base.t` test so that it doesn't die
on Perl 5.15. Thanks to Andreas J. Koenig for the report and diagnosis
and to Nicholas Clark for the fix.
- Duplicate paragraphs removed from `README.md` thanks to Ask Bjørn
Hansen.
- The `catch` functionality is deprecated. It will warn once for each
caller to keep log verbosity down. In the next release, it will warn
for every call. The release after that, it will be removed altogether.
0.45 2011-05-10T21:44:08
- Fixed crash when `in_txn()` was called before an actual connection
was established.
- Strongly recommend setting `AutoCommit` to true in the documentation.
Setting `AutoCommit` to false defeats the scoping behavior of `txn()`
and therefore should not be used.
- Nested exception handling now works properly in nested calls to
`run()` in fixup mode and in nested calls to `txn()` in all modes.
Thanks to Mark Lawrence for the report (RT #66974).
0.44 2011-03-20T01:04:59
- Fixed bug with the MySQL driver introduced by the auto-reconnection
fix in 0.43. Sorry for the lame mistake. [Lee Aylward]
0.43 2011-03-17T20:03:16
- DBIx::Connector now sets the DBI `RaiseError` parameter to true in
`new()` if neither it nor `HandleError` has been specified. This is to
increase the likelihood that exception handling will be properly
triggered in `run()`, `txn()`, and `svp()`. Documentation has also
been added to emphasize the importance of setting `RaiseError` or
`HandleError` appropriately.
- Documented that `AutoInactiveDestroy` is set to true in `new()` if it
is not specified. It's important tht this attribute be true in forking
environments.
- After connecting to the database, the MySQL driver,
DBIx::Connector::Driver::mysql, now always sets the
`mysql_auto_reconnect` attribute to false. This is to prevent MySQL's
auto-reconnection feature from interfering with DBIx::Connector's
auto-reconnection functionality in `fixup` mode. Thanks to Karen
Etheridge and Peter Rabbitson for the report.
- Removed mention of the use of the `catch` function from Try::Tiny,
since it is no longer compatible to use passing the exception-handling
function. Just using `catch =>` instead, which is cleaner-looking
anyway (RT #65196).
0.42 2010-12-17T18:35:54
- If a catch block died, the exception was not being propagated. That
is, if a catch block threw an exception, DBIx::Connector ate it, and
any calling code would not be able to catch it. This was a pretty
serious bug; upgrading is strongly recommended for anyone using catch
blocks.
- When `run()`, `txn()`, or `svp()` was called recursively from within a
second fixup execution, it was not respecting the fact that it was
recursive and could try to start a transaction again. This happened
*only* when a fixup run found that the database was disconnected and
successfully re-connected, so it's a pretty rare condition.
0.41 2010-12-08T21:02:32
- `connect()` no longer returns a disconnected database handle. Thanks
to John Siracusa for the spot (Issue #6).
- Added `disconnect_on_destroy()`, which can be used to disable
disconnecting the database handle when the connector object is
destroyed. Suggested by John Siracusa.
pkgsrc changes:
- bump required DBI version
Upstream changes:
0.40 2010-09-17T18:44:01
- The code refs passed to `run()`, `txn()`, and `svp()` now know their
contexts, so that `wantarray` can be used to decide what to return.
Patch from Yaroslav Korshak.
- Set `AutoInactiveDestroy` on connect with DBI 1.614 and higher, unless
it is explicitly set in the attributes. This makes things even safer
in a forking environment, preventing a parent process from getting
disconnected when a child exits without using the connection. The
reports from Peter Rabbitson and Aran Deltac and subsequent discussion
with Tim Bunce led to the addition of this attribute in DBI 1.614,
which is now the recommended version of DBI.
- `DESTROY()` no longer pings the database or rolls back transactions.
It now simply calls `disconnect`. This avoids warnings during global
destruction, and doesn't seem necessary anyway, as the DBI does these
things during global destruction (and always has). Thanks to Matt
Trout for the heads-up.
- `DESTROY()` now clears `CachedKids`, following the precedent of
DBIx::Class. May not be needed for recent-ish drivers, but seems
harmless and it's nice to avoid warnings were possible. Reported by
Matt Trout.
- The `connected()` method no longer `local`ly sets `RaiseError`. It
instead leaves that to the drivers (currenly only Driver::Oracle).
- The exception classes `DBIx::Connector::TxnRollbackError` and
`DBIx::Connector::SvpRollbackError` now use `our @ISA =` instead of
`use base` to inherit from `DBIx::Connector::RollbackError. This is to
avoid failures from mod_perl restarts. Suggested by Matt Trout.
- Require Test::Pod 1.41 for POD tests so that `L<text|url>` is
considered valid.
to trigger/signal a rebuild for the transition 5.10.1 -> 5.12.1.
The list of packages is computed by finding all packages which end
up having either of PERL5_USE_PACKLIST, BUILDLINK_API_DEPENDS.perl,
or PERL5_PACKLIST defined in their make setup (tested via
"make show-vars VARNAMES=..."), minus the packages updated after
the perl package update.
sno@ was right after all, obache@ kindly asked and he@ led the
way. Thanks!
DBIx::Connector provides a simple interface for fast and safe DBI connection
and transaction management. Connecting to a database can be expensive; you
don't want your application to re-connect every time you need to run a query.
The efficient thing to do is to hang on to a database handle to maintain a
connection to the database in order to minimize that overhead.
DBIx::Connector lets you do that without having to worry about dropped or
corrupted connections.
