RequestFactory's historyToken encoding / characters

[Arash]

Hi,

I use "$" as the separator in my Tokenizer and then parse the token respectively considering the "$" location. This used to work in gwt 2.3 mainly because getHistoryToken(...) never returned "$" as part of HistoryToken string. 

In gwt 2.4 "$" character can be easily found in historyToken which contradicts my assumption. The solution is simply switch to another separator but I like to know what the contract is for historyToken encoding characters so I can safely choose let's say "!" and be sure that requestFactory.getHistoryToken(...) will never contain "!"

Thanks,

Arash

[Thomas Broyer]

IIRC, RF uses a Base64 encoded ID, except it uses chars $ and _ in place of the traditional + and /, and it uses an @ separator between ID parts.

So ! should be "safe".