Added small blurb to the readme
This commit is contained in:
23
README.md
23
README.md
@@ -13,10 +13,9 @@ in the nif library or the sqlite database can crash the entire Erlang
|
||||
VM. If you do not want to take this risk, it is always possible to
|
||||
access the sqlite nif from a separate erlang node.
|
||||
|
||||
Special care has been taken not to block the scheduler of the calling
|
||||
process. This is done by handling all commands from erlang within a
|
||||
lightweight thread. The erlang scheduler will get control back when
|
||||
the command has been added to the command-queue of the thread.
|
||||
Special care has been taken not to block the normal erlang scheduler
|
||||
of the calling process. This is done by handling neccesary commands
|
||||
from erlang by using a dirty scheduler.
|
||||
|
||||
SQLite Compile Options
|
||||
----------------------
|
||||
@@ -66,3 +65,19 @@ INSERT INTO table VALUES("abcd", 1234);
|
||||
Will not work, because it sees the value 'abcd' as a SQL object
|
||||
value, and not a string literal.
|
||||
|
||||
Version 0.8.0
|
||||
-------------
|
||||
|
||||
This version is a major derivation from previous versions. When
|
||||
I started with this library it was implemented by using a separate
|
||||
os level thread per connection. At the time this was the only way
|
||||
to use functions in C which take longer to process than 1ms.
|
||||
A lot has changed since then. The VM now has dirty schedulers
|
||||
which make it possible to remove the thread per connection.
|
||||
This makes it possible to open a lot more connections. On the
|
||||
SQLite side some things have also changed. Extended error codes,
|
||||
introspection into the internals. This release modernizes the
|
||||
integration. In some places the API is no longer compatible and
|
||||
will require small changes.
|
||||
|
||||
|
||||
|
||||
Reference in New Issue
Block a user