Community
Participate
Working Groups
We seem to have a general problem with fetching navigation twice. you can see double GETS in the console when navigating around the tree. I suspect it's some kind of hashchange duplication. There's code to detect whether a hashchange is happening due to something we've just programmatically set, but it must be wrong.
I think it's best to revisit this after addressing bug 334195. I spent a few hours on this prior to M4, and each time I found a problem, it pointed to a follow-on issue. I have a patch in my eclipse workspace with some experiments that can be revisited.
I'll look at this for M5
There are triple GETS now
I've committed the first part of this change, which eliminates the double gets from the breadcrumb navigation, favorites, when you first land, browser history, and user updating of the hash. These were problems in both the tree and table. The fix for these was to ensure that the post-processing, including parent/child caching, that we do in fileClient.getchildren was also done when loading workspaces. In the navigator load workspace code, we were calling a subsequent getChildren (presumably to cause the post-processing fito happen). Now fileClient calls the same post-processing code when a workspace is loaded. The remaining problem is specific to nav-table. We still do a double GET when the user navigates to a folder using the folder link. I'm looking at that now
(In reply to comment #4) > The remaining problem is specific to nav-table. We still do a double GET when > the user navigates to a folder using the folder link. I'm looking at that now The problem here was that the "onclick" which expands the tree was attached to the column that contains the expando, rather than just the expando. So we would do a first get based on the onclick in the expando (to show children) but then immediately navigate the link, which changed the hash and caused the second get. Fixed >20110127.
