> [EMAIL PROTECTED] wrote: >> Hi everybody, >> >> Jerome suggested yesterday that we could handle the upgrade as follows: >> - use MyGWT widgets library for the web-gwt functionalities (modal >> dialog >> boxes) since it is a pure java library and its usage only involves >> changes >> in the web-gwt code >> - if needed, gwt-ext could be used for particular projects using web-gwt >> api and each project's code could be changed accordingly (this would be >> the case for watch). >> >> Both libraries seem pretty reliable, with active developers and nice >> community, so, from this point of view, they are both superior to gwttk >> that we currently use. >> >> Here's my +1 for this approach. >> >> WDYT? >> >> >>> Hi all, >>> >>> we want to upgrade XWiki Web gwt to gwt 1.4 for the 1.3 version so, >>> since >>> gwttk does not have a version for gwt 1.4 and we depend on it for >>> creating >>> modal dialogs, we need to replace tk with something else. A good option >>> is >>> gwt-ext, which also can replace gwt-widgets and provides some more nice >>> ui >>> objects and client functionality (like date parsing -- which we get >>> from >>> gwt-widgets for the moment). >>> The trouble with gwt-ext is that it requires ext javascript library to >>> run, which means that any gwt application needs to import, besides the >>> gwt.js file, some ext javascript files. Since the modal dialogs are >>> defined in the web-gwt module (so our gwt-ext dependency is there) and >>> we >>> cannot import the ext javascript files at that level, the only solution >>> is >>> to rely on the application using web-gwt to include right all required >>> files. It doesn't seem to me as good practice but I cannot figure out >>> how >>> big of an issue it is (since that application already has some rules to >>> obey, js files to include, etc to have gwt working). >>> >>> WDYT? >>> >>> > > Hi, > > I think that the most important think is that updating the products > using gwt should be as easy as possible. The two products that > intensively use gwt are Watch and Curriki, and since you are a Watch > developer, can you try the upgrade and see how much time it takes to > update the failing code?
I already did that with gwt-ext, the changes don't take long at all (6-8 hours, maybe) and they can be done in a transparent manner, without impacting for the code that uses them, except, of course for some appearance changes which are inevitable. There is also some testing to be done to ensure that no behaviour was changed which might also take some time. The problem with the projects using gwt code is that we would also upgrade gwt to gwt 1.4 and they must handle their own dependencies and resolve any code failing due to API breaking changes (which are not too many, actually). While XWiki Watch is OK from this point of view I don't know anything about Curriki yet. > Curriki is already on a tight deadline, so I'm > sure they won't like long delays (however, they are still using an older > branch, so they won't be affected by the trunk update yet). > > Sergiu > _______________________________________________ > devs mailing list > [email protected] > http://lists.xwiki.org/mailman/listinfo/devs > _______________________________________________ devs mailing list [email protected] http://lists.xwiki.org/mailman/listinfo/devs
