It is really difficult to work with these, and it's been like that for a long time (I remember these issues with Pharo 3 and 4). Or at least with GLORP Proxies (I don't know if those you're working with are any different).
I put a breakpoint in the Proxy to find what was triggering it and the breakpoint was ignored. I put `1 halt` and my image hung. Pressed Alt+. and the fireworks started. ¯\_(ツ)_/¯ Esteban A. Maringolo On Fri, Aug 7, 2020 at 8:59 PM Martin McClure <mar...@hand2mouse.com> wrote: > > Hi Esteban, > > I also have trouble with this sort of thing in Pharo -- inspectors > sending messages to objects under inspection that are not understood, or > that do not respond in the way the inspector expects. > > It is possible to build a basic inspector that sends no messages at all > to the inspected object (using the mirror methods), and it would be nice > to have that option in Pharo. > > It's also possible to have a "careful" inspector that uses a mirror > method to determine the class of the inspected object, then uses either > the purely-mirroring inspector or the richer inspector, depending on > whether the object's class inherits from Object. Since the GemStone > clients make use of proxy objects, GemStone's inspectors take this > general approach. They're not perfect either, but problems are > considerably fewer. > > Regards, > -Martin > > On 8/7/20 4:33 PM, Esteban Maringolo wrote: > > Hi, > > > > I'm having a hard time trying to debug some Glorp code in Pharo 8 > > because the inspectors and the glamour things are triggering the > > instantiation of my Proxy objects. > > > > I don't know how the inspectors, variable lists, and other tool work, > > but I assume there should be a way to reimplement something in Proxy > > to avoid causing a MessageNotUnderstood that triggers the > > instantiation. > > > > If there is no way for it (I guess there is, looking at some gt* > > selectors in ProtoObject), please consider having a debugPrintString > > (and debugPrintOn:) in the upcoming Spec2 based tools. :-) > > > > Best regards, > > > > Esteban A. Maringolo > > >
