2014-06-06 17:55 GMT+02:00 Nicolas Cellier < nicolas.cellier.aka.nice@gmail.com>:
2014-06-06 12:43 GMT+02:00 Nicolas Cellier < nicolas.cellier.aka.nice@gmail.com>:
2014-06-06 4:48 GMT+02:00 David T. Lewis lewis@mail.msen.com:
The change behaves as described, fixing testHandlerFromAction and
introducing the (now expected) failure into testHandlerRentrancy.
I am not qualified to offer any more useful critique other than to say that is clear that Nicolas has put some serious thought into this issue and I hope that others will take a close look at it.
Dave
Thanks David, Added wisdom: googling about usage of #rearmHandlerDuring: it seems required by Island mechanism in Tweak see https://code.google.com/p/pharo/issues/detail?id=2519
Maybe there is another implementation possible in Tweak, but that's definitely the main usage of #rearmHandlerDuring:
I also see this hack in VMMaker: http://forum.world.st/VM-Maker-VMMaker-oscog-eem-180-mcz-td4637135.html found after this thread: http://web.archiveorange.com/archive/v/kLJIkmCV2lbVUoHLbp3R
I'll try and give a try to above Eliot's snippets
And good news, the new implementation of rearmHandlerDuring: is still work on Eliot's snippets (#caught is correctly transcripted 5 times). Though the count of second snippet is still 1
Nah! The count is not 1 when the outer handler is rearmed. The count is 6 (1 initial ProgressInitiationException + 5 inner Notification). The return value is 1 (it is the value returned by 1 to: 5 do: ...), and that's what abused me.