Summary: | tao_ft_interception_point has to be removed | ||
---|---|---|---|
Product: | TAO | Reporter: | Johnny Willemsen <jwillemsen> |
Component: | Portable Interceptors | Assignee: | DOC Center Support List (internal) <tao-support> |
Status: | RESOLVED INVALID | ||
Severity: | normal | CC: | sm |
Priority: | P3 | ||
Version: | 1.4.6 | ||
Hardware: | All | ||
OS: | All | ||
Bug Depends on: | 1369 | ||
Bug Blocks: | 2181 |
Description
Johnny Willemsen
2005-06-27 07:43:27 CDT
added depends I don't know why the work on bug 1369 was considered to render this interception point unnecessary (perhaps you should svn ann the comment and go and ask the comitter ?). I do know that this interception point is currently required for successful server side FT operation. We have customers actively using it and so it should not be zapped unless someone can explain to me how else we can return cached results from repeated invocations. Simon, if we require this and there is no other way, shouldn't we then raise an OMG issue to this interception point added to spec? Hi Johnny - sorry I missed your comment ages back. If you build this extension in I think you trample all over the interceptor flow rules so the PI chapter (and hence all compliant implementations) would need a rewrite. I can't see anyone having the will to do that in all honesty. It's an important feature if (and only if) you are trying to put together a 'plugged in' (by means of PI as opposed as hardwired into the ORB) server side FT implementation. TAOs own original FT prototype from a bit back is like this and we have one customer with a similar bit of kit - it's still a bit of a niche market though. For this reason I'd oppose removal without a more compelling motivation than source tidying but think that it staying an optional proprietary extension is fine for us and them right now. I'd vote resolve invalid to get it off the list. Maybe remove the comment if it's still there. closing as invalid, seems to be good reason to keep this |