On Tue, Jan 25, 2011 at 2:58 PM, Michael Roberts <mike@mjr104.co.uk> wrote:
yes as Adrian pointed out debugger highlighting is broken in Pharo. Has been since the very beginning. what we need ideally is a set of test cases describing the desired behaviour. i struggled to simulate this, i am not sure what a good test harness is for the debugger. ideally the highlighting can be tested without the debugger (i.e without the necessary processes and activations), just triggering the necessary execution machinery that calculates highlight offsets (intervals). With a good set of tests we could then better compare Squeak and Pharo implementations because a lot of fixes go into squeak and we need to harvest them over here. any help of course welcome in this area. I poke at it now and then....
Take a look at the ClosureCompilerTest>>testSourceRangeAccessForClosureBytecodeInjectInto which tests the source ranges (that the debugger uses to highlight) for all pcs in the inject:into: method.
cheers, Mike
On Tue, Jan 25, 2011 at 2:20 PM, Adrian Lienhard <adi@netstyle.ch> wrote:
On Jan 25, 2011, at 14:05 , Tudor Girba wrote:
I am not alone! :)
Anyone else noticed this?
yes
Is it http://code.google.com/p/pharo/issues/detail?id=709 ?
Adrian