There was a thread earlier about this same one-line fix. At the time the first question which went unanswered was why is it necessary -- if we are convinced it is, then I think there is a lot more work to do besides this fix.
http://gwt-code-reviews.appspot.com/44801/diff/1/2 File dev/core/src/com/google/gwt/core/ext/linker/impl/hosted.html (right): http://gwt-code-reviews.appspot.com/44801/diff/1/2#newcode254 Line 254: $hosted = decodeURIComponent($hosted); On 2009/06/22 13:40:18, jgw wrote: > Pretend these tabs are fixed. I will have to dig up the thread from the last time this one-line fix was proposed, but basically we need to: * decode it earlier -- for example, the indexOfs just above this line assume that it can search for decoded characters * encode the URL properly in the shells so the decoding is balanced, otherwise we risk breaking things. * does the browser do decoding to split it up into the components (ie, do we get location.search set properly at all if the ? is encoded) or do we need to split it ourselves in that case? I also wonder about doing the encoding in the shells and interaction with the history token -- ie, are we encoding it twice? http://gwt-code-reviews.appspot.com/44801 --~--~---------~--~----~------------~-------~--~----~ http://groups.google.com/group/Google-Web-Toolkit-Contributors -~----------~----~----~----~------~----~------~--~---
