Bob, > There are a few things in John Barton write up that we didn't implement > such an adding annotations or an optional parameter to the eval() to > described. ATF doesn't parse the JavaScript, so it can't pick up an > annotations though we may be able to do something now that we have JSDT.
The piece of metadata describes the uri location for what was put in the eval, which I think is good data to have so you can get back from the debugger to the actual source file and associate it with the right buffer. It is basically a "//" style comment at the end of the file (placed at the end of the file so that it does not ruin the line offset of the rest of the code) It should be trivial to get this data, or at least make an honest attempt to get it, without a full-blown Javascript parser. > Things like this should probably be standardized maybe through the Open > Ajax Alliance. It doesn't help to implement something that no Ajax Runtime > plans to implement. Well, you've got to start somewhere. Given that this is a niche problem, I think coming up with a solution in open source and getting multiple implementations is a perfectly reasonable first step, and we can then take this to the OAA. The OAA is wrestling with larger issues right now, I think. Cheers, Adam _______________________________________________ atf-dev mailing list atf-dev@eclipse.org https://dev.eclipse.org/mailman/listinfo/atf-dev
