David Gewirtz wrote:
John Stanton wrote:
What language are you using? Certain language environments would make
implementing the server clumsy.
I'm using an environment called the Frontier Kernel
(http://www.frontierkernel.org), which is the open source version of the
original UserLand Frontier. As obscure as that may sound, Frontier's been
around for a very long time, and was the tool used for the first blogging
and RSS feeds.
That said, I recently finished coding an SQLite wrapper into the source of
the Frontier Kernel. So far, I've only implemented a very few functions, but
wanted to pull them into my application before adding more. The wiki on this
is at http://frontierkernel.org/docs/verbs/sqlite.
You can feed the the DB thread with a FIFO buffer sync'd on events,
which are an efficient mechanism in a threaded environment.
That does make sense. The only gotcha is stuff like long, multipage search
responses (something, honestly, I'm not doing now), and there then needs to
be some artificial tracking added to the app to keep track of which user
request caused which database result. Not impossible, but I thought this
might add almost as much (or maybe more, since this part's scripted)
overhead as just opening and closing the database a lot, and semaphoring
availability.
-- David
-----------------------------------------------------------------------------
To unsubscribe, send email to [EMAIL PROTECTED]
-----------------------------------------------------------------------------
Sqlite is essentially single user, so long searches are going to always
be an issue if you want to write to the DB simultaneously.
I don't understand how you apply Frontier. To build a multi-threaded
web server hosting Sqlite looks like a project to replace Frontier,
which already has embedded Sqlite.
-----------------------------------------------------------------------------
To unsubscribe, send email to [EMAIL PROTECTED]
-----------------------------------------------------------------------------
