Hi Andrea - I would like to do that ... question first.
Is this idea anything different then the Java Assert keyword? It is a developer only message we are trying to reproduce here ... it has a documented set of procedures for how to run with asserts on and asserts off ... looks like a good fit. I am reminded that we *used* to have hard failures on the development branch; and James had us make them "soft" for the 2.0 release. I would actually like to see if we can return to that pattern - hard errors on the milestone releases (so we can catch problems) and we relax a bit when making final releases (so we can be kinder to client code). A "strict" mode as it were. Back to what I learned from the 2.4-M0 experience: In general I would like us to have a plan for "data store conformance", recent experience has shown it needs to be done - and that the work is not that hard. Turning on a test such as the above is an example of a small easy step we can take to improve the code base. Let's round up some more hands at todays meeting .. and see if it is the kind of thing we can do as a Europe / North America tag team. In testing the last release we found a problem; with property datastore! I know I did change things last iteration with respect to Query - but the problem encountered actually had to do with transaction support. As you saw on email the test cases was clear as mud - and I have spent some time renaming the variables. And in an amazing fashion the problem is now totally clear - transaction independence has been broken! (That is two clients threads talking to the same datastore on different transactions are going to trip on each other - only the first commit will work properly). Since this test used to pass; and the code is just testing transaction / abstract DataStore functionality I am forced to conclude that transaction state diff has been broken :-( Cheers, Jody Andrea Aime wrote: > Jody Garnett ha scritto: >> Going over the release ntoes I see a lot to like (thanks especially >> to Andrea for the "Unclosed transactions should close automatically >> on finalize and complain loudly") > My pleasure. Btw, are you still wanting a switch to make it throw hard > exceptions instead of just logging stuff? > We could use a system variable for that. > We could share the work: I do the coding, you update the docs. > How does that sound? > Cheers > Andrea ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys-and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ Geotools-devel mailing list Geotools-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geotools-devel
