The way I would envisage a PL/SQL type capability existing in Sqlite is
by adding the grammar to the existing parser and adding the code
generation to the existing code generator. The existing VDBE engine
might need to be slightly extended.
Apparently a version of Sqlite which uses stored VDB
At 17:00 24/12/2006, you wrote:
On Sun, Dec 24, 2006 at 09:35:01AM -0600, John Stanton wrote:
> >There is even a starting grammar for you:
> > http://www.antlr.org/grammar/1107752678378/PLSQLGrammar.g
> A compiler for a subset of PL/SQL would not be too arduous a project,
If what you want is
Andrew Piskorski wrote:
On Sun, Dec 24, 2006 at 09:35:01AM -0600, John Stanton wrote:
There is even a starting grammar for you:
http://www.antlr.org/grammar/1107752678378/PLSQLGrammar.g
A compiler for a subset of PL/SQL would not be too arduous a project,
If what you want is something l
--- Andrew Piskorski <[EMAIL PROTECTED]> wrote:
> If what you want is something like PL/SQL, it might be both easier and
> better to retarget PostgreSQL's PL/pgSQL for SQLite opcodes, rather
> than starting from scratch with your own Oracle PL/SQL style language.
> (And PL/pgSQL's BSD license is co
On Sun, Dec 24, 2006 at 09:35:01AM -0600, John Stanton wrote:
> >There is even a starting grammar for you:
> > http://www.antlr.org/grammar/1107752678378/PLSQLGrammar.g
> A compiler for a subset of PL/SQL would not be too arduous a project,
If what you want is something like PL/SQL, it might b
A compiler for a subset of PL/SQL would not be too arduous a project,
and a code generator into VDBE would, as you suggest, give an insight
into what extra opcodes would be desirable to produce dense PL/SQ.
Obviously the SQL component and optimization is already handled.
I can see that the maj
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Eduardo Morras wrote:
| but don't know much about the VDBE code.
The opcodes and operation (stacks etc) are all well documented:
~ http://www.sqlite.org/opcode.html
All the necessary opcodes are already present in that you can do checks
(ge, eq, le
At 03:41 22/12/2006, you wrote:
There has been discussion about extending Sqlite to have more
functions, but the risk is creating Sqlite-Bloat and losing the most
endearing feature of the product, its light weight and simplicity.
Here is an interesting and thought provoking discussion on the
There has been discussion about extending Sqlite to have more functions,
but the risk is creating Sqlite-Bloat and losing the most endearing
feature of the product, its light weight and simplicity.
Here is an interesting and thought provoking discussion on the general
subject. A good case is
9 matches
Mail list logo
