Hi Davanum, On Mon, 2005-12-05 at 09:45 -0500, Davanum Srinivas wrote: > Personally, i'd like to see progress on the "VM Interface" ASAP, that > would go a long way to removing the mistrust. That would enable > harmony VM to use unmodified classpath stuff as-is for development > purposes and can act as a firewall till the licenses get sorted out.
I am afraid this is ignoring part of the community trust issues and just hoping technology will solve it all. But sure, the current VM interface for GNU Classpath is described at: http://www.gnu.org/software/classpath/docs/vmintegration.html This is a living document and interface. We learn all the time about needs of different runtimes and execution engines. And new stuff gets added with each release of course. Just today a new section on the recently added instrumentation hooks was added (already in CVS, website will update in about 2 hours). > See snippet from Leo's email: > "Mark told me someone tried something like that a year or two ago > already. I forgot whom or what it was called, but I'd suggest trying > to learn about it and if it failed, why." I believe that was after 2 beers :) And I don't know exactly which part of the conversation this was about. I know I said how interesting it was the Anthony Green came up with freevm.org and now Stefano Mazzocchi came up with openvm.org. Both wanting to bridge the gap and bringing it onto a higher level to be "above all parties". In the end freevm.org both succeeded and failed. It failed because freevm.org is no longer, it succeeded because we decided we didn't need anything "above" GNU Classpath, gcj and Kaffe, but that we would just cooperate as is. > If i see some componentization such that i can drop in say > xalan/xerces and not use classpath's built-in stuff (this would end > users can mix and match stuff to make their distribution) that would > be even better. But "VM Interface" is priority #1. gcj for example has a switch/system-property -Dendorsed.dirs which allows for this. Cheers, Mark
signature.asc
Description: This is a digitally signed message part