(I will answer all the compiler related question later this week� I am a bit snowed under
and need to finish some thing before� in general: the whole compiler API needs
work, but we need to get rid of the old compiler before, else it is very hard to do� so this
is for Pharo4)

On 14 Jan 2014, at 00:46, Nicolai Hess <nicolaihess@web.de> wrote:

Am I right that with the new compiler, the way
how analyser warnings work changed too?
see the description in
case 12512 SemanticAnalyzer ignores answer from UnknownSelectorWarning
and
case 12514
OCSemanticWarnings ignore "cancel" from a confirmation dialog

I tried to fix this but I can not figure out how the answers from an
OCSemanticWarning dialog supposed to interact with the analyser/parser/compiler.

Especially "Cancel" should cancel the current evaluation.
I tried this with evaluationg
compilationContext failBlock value
That works but I don't know if this is the way it is supposed to be.

For those answers that would change the code
(OCUnusedVariableWarning) is another problem, that the
compiler does not handle ReparseAfterSourceEditing.
(see case 12442 Nautilus brakes code when removing more than one unused local variables)

Nicolai