In accordance with the design (approximately), he mentions the separation of modules, where this makes sense, since it technically does not support prefetching; which, as said, I can only assume that GMail shows progress based on module loading progress verified through callbacks (i.e. GWT.runAsync() ). Not all modules are the same size, but you can “evaluate” and assign a weighted percentage for each module (see GWT compile report )
If your initial page size loading is large (> = 1 MB), I would recommend looking into refactoring and optimizing your design to be a bit easier (scaffolding). This would mean more trips to the server, but limiting the initial download size. This would provide the basis for more accurate user feedback that the page is “working” (ie, an Undefined progress bar) and avoids unnecessary accuracy.
The overall page size is a difficult thing to really measure the route along the route to the client / browser, this would be more trouble than it would be worth displaying to the user. You might be able to accomplish this with a few lightweight modules, but you will have to compensate for all generated resources, such as ClientBundle , because the GWT creates a specific set for each browser swap.
Sidenote : In real time, for example, to achieve file size accuracy (for example, 80 KB 1.29 MB (6% included)) is usually used when downloading a file (for example, images, video, music, etc.) on user hard drive. Data at this point is no longer temporary, and data storage space can be a problem, caches are cleared and cleared, so it usually doesn't matter how big the web page is.
source share