Closed GoogleCodeExporter closed 9 years ago
Original comment by goudreau...@gmail.com
on 26 May 2010 at 7:54
Original comment by goudreau...@gmail.com
on 27 May 2010 at 4:09
Original comment by goudreau...@gmail.com
on 27 May 2010 at 5:36
There may be loose of information:
- in PlaceManagerImpl, the revelErrorPlace is called with "request.toString()"
as the
argument
- in ProxyPlaceAbstract, reavealUnauthorizedPlace is called with
"request.getNameToken"
May be changing both to request and so the both method will take a PlaceRequest.
Original comment by olivier....@free.fr
on 3 Jun 2010 at 9:14
So, you're proposing to give the complete PlaceRequest as an argument ?
Else, there's a reason for that behavior.
Reveal error place take the Parameters into consideration while
revealUnauthorizedPlace only tell which place was unauthorized.
But I kinda love your idea of giving the complete PlaceRequest in fault and
give the
user the choice on how to use it.
Original comment by goudreau...@gmail.com
on 3 Jun 2010 at 11:38
Yes, I propose to give the PlaceRequest. You've build it so why not. If someone
need
the parameters, he will build a PlaceRequest from the token.
For revealErrorPlace, you build a PlaceRequest from String, transform it to
String,
call revealErrorPlace with the String. If I need a PlaceRequest, I transform the
String to a PlaceRequest... It's just like we love to do things again and again
;).
For revealUnhautorizedPlace, I'm not sure about the parameters but there's a
consistency problem. The default revealUnhautorizedPlace call revealErrorPlace,
so
sometimes revealErrorPlace have the full token, sometimes not...
Original comment by olivier....@free.fr
on 3 Jun 2010 at 1:58
Right. (Didn't think this morning about the case where only revealErrorPlace
was overwritten. The real default
behavior doesn't use the string at all and reveal defaultPlace.)
I'll open a new issue.
Original comment by goudreau...@gmail.com
on 3 Jun 2010 at 2:04
What if the history token can't be parsed into a PlaceRequest? I think we catch
an exception in this case
and so have no placerequest? Should we pass both the string and placerequest?
Original comment by philippe.beaudoin
on 3 Jun 2010 at 2:29
I came across this while working on issue 108 and seriously, I doubt that this
exception is useful. The only case
where this exception would be thrown is when historyToken.split(
HIERARCHY_PATTERN ); fail... so if REGEX
syntax is invalid... and that's a private final static constant... or if
HistoryToken is null... but then, even passing
the string would be for nothing.
What I've done, is to return an empty PlaceRequest instead.
I'm ready to review issue 108 btw.
Original comment by goudreau...@gmail.com
on 3 Jun 2010 at 2:50
BTW, in updateHistory and onPlaceChanged you simply return nothing for that
exception.
Original comment by goudreau...@gmail.com
on 3 Jun 2010 at 2:51
Ok sorry, forget what I said about the exception.
In kinda stepped over toPlaceRequest.
Original comment by goudreau...@gmail.com
on 3 Jun 2010 at 2:57
Original issue reported on code.google.com by
philippe.beaudoin
on 6 May 2010 at 4:25