Hi Chris, it is much easier to interpret your results using the Story of Your Compile (soyc) report feature built into GWT 2.0. To activte that you need to add -compileReport to the arguments of the GWT compiler (if you are using the Eclipse plugin, the compiler arguments are hidden under the "Advanced" panel in the compile dialogue). It will show you exactly which functions take what amount of space etc.. You will then find your compile report in the subdirectory extras/ in your project root.
Anyway, from the looks of it in both of your tests the vast majority of the code is downloaded right at the beginning (328/329 k), then the code for your first panel is downloaded, which is about 30k, as well as a leftover fragment (0.5/7.6k k), which includes code that is specific to no particular split point. It is also possible, that the 30k are the leftover and the 0.5k are the split point code. The remaining code required for the second click seems to be pretty smallish at less than 1k. There definitely seems to be room for improvement in your setup, but it is really hard to tell how exactly based on your analysis. You can find out how to optimize your splitpoints in the developers guide http://code.google.com/webtoolkit/doc/latest/DevGuideCodeSplitting.html . Concerning your second question, my preliminary tests have shown that static member functions and variables do not cause the entire class to be loaded, so you should be okay :) Cheers and happy coding, Jonas On Dec 12, 4:46 pm, chris <stocker.christ...@gmail.com> wrote: > Hi, > > if have a GWT application using the GXT library (http://www.extjs.com). My > Application has on the top a toolbar with a lot of > menu-entries. every menu-entry opens a Window (from GXT) with serveral > UI-elements on it. i want to prevent loading all the Windows from my > app at startup. so i built in code splitting - only a click on the > menu-entry creates the window with new() and shows it. > > i built the code-splitting in different ways to see, how it is > working, and therefore i get different results. With safari i > monitored my results, but wat is the best result? in every case i see > dynamic loading for the javascript with different file sizes. can > someone help me to interpret the results? > > for testing purpose i have only to menu-entries. for the test, i click > the two in series. > > TEST 1 > ----------- > startup: > - file 1: 328 KB > click 1: > - file 2: 7,6 KB > - file 3: 25,2 KB > click 2: > - file 4: 0,8 KB > > TEST 2 > ----------- > startup: > - file 1: 329 KB > click 1: > - file 2: 0,5 KB > - file 3: 32,6 KB > click 2: > - file 4: 0,5 KB > > Another question i have is the following: If i use the Window before > code-splitting (e.g. in the onLoad()) with "MyWindow.class" or > "MyWindow.class.getName()" - not new() - does the code-splitting for > the Window still working? Or do these code-statements also load the > Window-JavaScript? -- You received this message because you are subscribed to the Google Groups "Google Web Toolkit" group. To post to this group, send email to google-web-tool...@googlegroups.com. To unsubscribe from this group, send email to google-web-toolkit+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/google-web-toolkit?hl=en.
