I just added a perf branch to the main sequel repository (http://
github.com/jeremyevans/sequel/commits/perf).  There are a couple of
commits in the perf branch that aren't in the master branch.  I'd like
some testing of them before I merge the perf branch into master.

The first commit is a 20-25% performance increase to Dataset#literal.
This is probably only about a 2% increase to Sequel code in general,
but the internals are also made nicer in this commit.  All adapters
had to be modified for this commit, and I only tested on SQLite and
PostgreSQL.  If you use any other adapters, please test to make sure I
didn't break anything.  I'll be testing MySQL and H2 databases, and
JDBC and DO adapters later in the week, but I'm not going to be doing
any testing of the other adapters.  If you use the Firebird or ODBC
adapters, I would especially like testing of those, as the changes
were non trivial.

The second commit is a 30-40% performance increase to Model.[], at
least for models with single table, single primary key datasets. This
is potentially much more significant, as Model.[] is a high level
method called directly by user code (and often the most significant or
only method call in a web app action).  It works by using a static SQL
dataset instead of going through Dataset#select_sql.

Both of these performance increases were measure against the SQLite
and PostgreSQL databases.  I'd appreciate feedback on these before
merging, especially to make sure that nothing got broken.  Please test
and report any breakage here.  If you have a performance test suite
for your app, I'd appreciate before and after results.

Thanks,
Jeremy
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"sequel-talk" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/sequel-talk?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to