On Sat, Jul 4, 2009 at 12:36 PM, BobV <[email protected]> wrote: > Thanks for taking a look. I've updated my client. > > > I think the dev/ comments have been subsumed by the latest version of > the patch at: > http://gwt-code-reviews.appspot.com/46801/show >
Yeah, I think so too. All other comments fine, but > Pair: shouldn't there be a common place for this, outside the deRPC server > > context? > > There should be, but I don't really know where it would go. I would > have put this in the dev.util.collect package, but that's not bundled > with gwt-user or gwt-servlet. > Point taken, but that's not a great reason. At the least, that may argue for creating com.google.gwt.user.util[.collect] or the like, so we could have a utility place for user code as well as for compiler code? Or for bundling com.google.gwt.util.** into both gwt-dev-* and also gwt-user/gwt-servlet? (The logical extreme is to create gwt-util.jar as a requirement of either-or-both; I'm not going that far. I won't kick and scream about Pair, but it does suggest this is a general class that may be a larger issue for a more interesting example. > > RPCServlet.java: Since we need matching e.g. X-GWT-Module-Base strings in > > multiple places, should we have a shared Constants class for them? > > Right now gwt-servlet.jar has all of the client code packed into it. > If this were to change, the Constants class would cause runtime-only > breaks. Sorry, how so? (Or rather, how so in a way that wouldn't likewise break today?) Today, we have multiple copies of the constants, which if changed would have to separately be found and corrected (search-and-replace, easy, I know... if you know there are multiple copies.) With Constants, we'd have one copy, which would be replicated by the build system into gwt-user.jar, and into gwt-servlet.jar which is a strict subset of gwt-user.jar. Server code would at runtime[1] use the copy from gwt-servlet.jar; hosted or web-compiled code would use the copy from gwt-user.jar, and web-compiled would bake that value in... but in terms of changes, that's the same behavior as what we have today. [1] Actually, it's not at runtime... javac inlines final constants anyway, so the server code values are still baked in at (server) compile time. --~--~---------~--~----~------------~-------~--~----~ http://groups.google.com/group/Google-Web-Toolkit-Contributors -~----------~----~----~----~------~----~------~--~---
