> >> JB>10. Add suport for class instances (ByteCode generation and
"*.class"
> >> JB>transformations). I think this is optional.
> >>
> >> Ohh BCEL would be an option, but the community their sucks. Zero
traffic
> >in
> >> the mailing list. You don't get any help from that dudes. Am I right
Paul
> >
> >;).
> >I have this code and I think it is very good and clear, but I think it is
> >beter to generate some ASM for java and compile it in memory.
>
> Is this code written in BCEL?
> ASM?

 I saw a few ASM for java implementations, one of them use BCEL to implement
this stuff,
I don't remember, but I think it is JASMIN.
I have plans to generate byte code this way : Template -> ASM -> Byte Code


> >We will need pool for arrays, I dont know trivial and good
implementation,
> >but it is very trivial to implement (the Same as "Swap" and "Reacheble")
>
> Aha, for what do we need this pool?
This pool will be for List, Map and Set, very trivial wrapper will return
object to pool
on finalization.

> >I don have ideas abaut this :) Datasource implementation is incorrect,
but
> >it has  good performance on PostgreSQL, because driver implements pool
> >itself .... .
>
> It has now his own package, and people can use it if they want. But maybe
> we should remove this PostgreSQL javadoc, because it works for hsqldb too
;).
We will remove this Datasource itself later, it is good for some
applications and works
with most of drivers in "readonly" mode, but it will not work on simplestore
if
transactionCount > maxConectionCount, but it can be used for testing, don't
use it in production if data is not "readonly", you will have fatal errors
in transactions.







--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to