Re: [gwt-contrib] Re: More control over split point fragments

2011-10-07 Thread John Patterson
On 08/10/2011 01:56, Alan Leung wrote: I don't really understand this point. Smaller fragments are still strongly cachable - there are just more download requests, no? In a multi-page app it is likely that most of the code in the left overs fragment (used in other pages) will

Re: [gwt-contrib] Re: More control over split point fragments

2011-10-07 Thread Alan Leung
> > > I don't really understand this point. Smaller fragments are still strongly > cachable - there are just more download requests, no? In a multi-page app > it is likely that most of the code in the left overs fragment (used in other > pages) will not be needed. > > What Lex described was havin

Re: [gwt-contrib] Re: More control over split point fragments

2011-10-06 Thread John Patterson
On 06/10/2011 07:27, Alan Leung wrote: I have studied Lex Spoon's writeup extensively a few weeks back. While I do believe it is beneficial to leftover code size, when I bought it up with some of the internal projects, they believe the cache-ability lost with multiple leftover fragments is not

Re: [gwt-contrib] Re: More control over split point fragments

2011-10-06 Thread John Patterson
On Thursday, 6 October 2011 01:02:10 UTC+7, John A. Tamplin wrote: > > The basic code splitting algorithm can handle this, > Can you give some more details? I have not seen a solution to this problem yet. -- http://groups.google.com/group/Google-Web-Toolkit-Contributors

Re: [gwt-contrib] Re: More control over split point fragments

2011-10-05 Thread Alan Leung
I think Ray described it perfectly. I am current in process of make the change he described. While I have the basic idea implemented (minus some bugs I am current fixing), the trick is to know what are the right parameters for doing these kinda of things. Example 1) How many Async point do we me

Re: [gwt-contrib] Re: More control over split point fragments

2011-10-05 Thread Ray Cromwell
The problem of detecting common code and putting them into shared fragments of certain size is related to the Graph Partitioning Problem, like many compiler problems, which is NP-Complete. That doesn't mean it can't be done, Alan Leung is working on this for the next release, it just means it has t

Re: [gwt-contrib] Re: More control over split point fragments

2011-10-05 Thread John Tamplin
On Wed, Oct 5, 2011 at 12:52 PM, John Patterson wrote: > That seems likes a step in the right direction - but ideally there would be > some algorithm that can put common code into new shared fragments. I'm sure > its a lot harder than I imagine to handle all the permutations and load > sequences.

Re: [gwt-contrib] Re: More control over split point fragments

2011-10-05 Thread John Patterson
On 05/10/2011 23:35, Jens wrote: I think currently the left over fragment will grow forever if you introduce new commonly used classes (used in more than one split point). Maybe a potential solution will make it into one of the next GWT releases. Take a look at: https://groups.google.com/d/ms

[gwt-contrib] Re: More control over split point fragments

2011-10-05 Thread Jens
I think currently the left over fragment will grow forever if you introduce new commonly used classes (used in more than one split point). Maybe a potential solution will make it into one of the next GWT releases. Take a look at: https://groups.google.com/d/msg/google-web-toolkit-contributors