I'm using v3, and the #Spotlight (class name) that comes with it. I didn't know about Spotter of v4 before your comment. I've just tried it, it could be handy for some specific situations, after i manually tweak its non-sense whiteness which somewhy escapes global theme defaults (lovely dark theme), i'll give it a try. Yet to the end of shortcuting accessing methods/classes i found Spotlight to be more direct and keystroke cheap.
Saludos On Tue, Feb 3, 2015 at 1:28 PM, Guillermo Polito <guillermopol...@gmail.com> wrote: > Hi! > > Spotlight or Spotter? Which version of Pharo are you using? > > I ask because the old Spotlight should be using the same code completion > mechanism => it should appear/disappear from both at the same time... > > As for spotter, I don't have an answer now :) > > Saludos, > Guille > > El Tue Feb 03 2015 at 5:24:26 PM, Laura Risani <laura.ris...@gmail.com> > escribió: > > Thank you all for your answers. >> >> Hi Guille , >> >> Nice solution! It worked perfectly for Code Completition! >> >> Yet the obsolete symbols (class names and selectors) keep showing in >> Spotlight, but i guess this is too implementation specific. I will step >> through its execution to try to find out how it works. >> >> Best, >> Laura >> >> >> >> On Mon, Feb 2, 2015 at 8:39 AM, Guillermo Polito < >> guillermopol...@gmail.com> wrote: >> >>> Well yes, but GCs are happening all the time :). So the problem is that >>> someone is keeping a strong reference on the symbol. >>> >>> I tried the following: >>> >>> - create a method named #unusedUnexpectedMessage: That came up in the >>> auto completion, ok >>> - remove it: Still in auto completion >>> - force GC a looot: still in auto completion >>> >>> After chasing strong references I could clean up my image by doing: >>> >>> ChangeSet cleanUp: true. >>> RecentMessageList cleanUp. >>> 10 timesRepeat: [Smalltalk garbageCollect] >>> >>> First obvious thing: changesets and friends (lets also think about >>> nautilus history that may do that) could keep strong references on symbols. >>> >>> Then the question is if that is correct or not... To me the problem is >>> that the auto completion mechanism is pretty primitive and depends on all >>> existing symbols instead of <the subset of symbols that could be messages>. >>> >>> Guille >>> >>> El Mon Feb 02 2015 at 12:05:49 PM, Sven Van Caekenberghe <s...@stfx.eu> >>> escribió: >>> >>> But weak refs are only killed after GC, right ? >>>> >>>> And even then... >>>> >>>> > On 02 Feb 2015, at 12:03, Guillermo Polito <guillermopol...@gmail.com> >>>> wrote: >>>> > >>>> > It's the table that keeps the symbols and checks their uniqueness. >>>> > >>>> > but AFAIK the symbol table is weak. So probably it's the completion >>>> mechanism that is keeping extra references... >>>> > El Mon Feb 02 2015 at 11:52:28 AM, p...@highoctane.be < >>>> p...@highoctane.be> escribió: >>>> > Got the same problem here. >>>> > >>>> > Annoying when third parties have to do something in the environment >>>> as they are shown things that do not exist. >>>> > >>>> > Where to look? What's the symbol intern table? >>>> > >>>> > Phil >>>> > >>>> > >>>> > On Mon, Feb 2, 2015 at 11:44 AM, stepharo <steph...@free.fr> wrote: >>>> > Probably by resetting the Symbol intern table. but no time to dive >>>> into it now. >>>> > Hi all, >>>> > >>>> > Code completition tools show class/method old names i've changed and >>>> no longer use. How can i remove all unused (occurring nowhere in source >>>> code) or unimplemented names? >>>> > >>>> > Best, >>>> > Laura >>>> > >>>> > >>>> > >>>> > >>>> > >>>> > >>>> > >>>> >>>> >>>> >>
