Hi Robert,
I have not noticed a major problem with performance at this point, but that is partially because I have created my own Introspector. Basically, I created an Introspector that reads a single mapping file that has the equivalent of all the .betwixt files (this allows me to go to and from multiple XML formats by just sending a different mapping file to the Introspector). As part of processing the mapping file, the XMLBeanInfo's are created and put in the registry. So I take a one time hit (at startup) to process the entire mapping file and do any reflection that is required, but then every time I read or write something, I already have a fully loaded registry of XMLBeanInfos. If you are interested in that kind of an Introspector, I have spoken to my employer and I can donate the code (although I don't have it completely stable and tested yet).
that sounds interesting. we'd need to have a signed grant form from your employers before we could accept the code but if the legalese doesn't frighten you guys, let me know and i'll
With regards to using a byte code manipulator to generate XMLBeanInfos, that is not something I could work on now (at least not on my employer's time). However, that may change in a few months if we decide to continue with betwixt (so far, so good) and are ready to start looking more closely at performance and optimizations.
fine
One question I have had is regarding the road map for betwixt. The last official release (alpha) was quite a while back. Are there plans for another release in the near future?
that's not an easy question to answer (for various reasons).
the first is the question about whether the code's ready. it's unlikely that the branch will be ready to release for a while, realistically a month and maybe more. there are a lot more interface changes i'm planning and i'm not really willing to release and support code that i know is going to be deprecated soon. CVS HEAD is stable and ready for a 0.x release (which indicates that the code's ok but the interfaces are not really stable).
the second question is whether there's a suitable release manager. i'm not on the pmc at the moment and so i can't/won't cut releases. there are other committers who are (mildly) interested in cutting a release for betwixt but it isn't a priority for them. if one of them suddenly wanted to cut a 0.5 (from HEAD) i'd probably support them.
so the best answer i can give is probably not soon with a HEAD release more likely than the branch. (it is worth remembering that lobbying will sometime make a difference to these timescales.)
- robert
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
