I noticed that in 4.0 one gets a (resume-able) exception when trying to define a class in category ''. Why? Is it true, as the error message states, that "the Category should not be empty"? One of our key principles inherited from Smalltalk is to treat the programmer like they know what they're doing. Although only a minor annoyance in this case, I often create classes in empty categories specifically because they are throwaways - like a temp directory, and was a bit confused by the debugger popping up, especially before I realized that the exception was resume-able... Also, that made me think about something else... What if the debugger somehow made it immediately obvious whether the exception was resume-able? That could be useful in cases like these to make the user's options clearer. ----- Cheers, Sean -- View this message in context: http://forum.world.st/Empty-Package-Names-tp4803642.html Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.
On 04 Feb 2015, at 15:14, Sean P. DeNigris <sean@clipperadams.com> wrote:
I noticed that in 4.0 one gets a (resume-able) exception when trying to define a class in category ''. Why? Is it true, as the error message states, that "the Category should not be empty"? One of our key principles inherited from Smalltalk is to treat the programmer like they know what they're doing. Although only a minor annoyance in this case, I often create classes in empty categories specifically because they are throwaways - like a temp directory, and was a bit confused by the debugger popping up, especially before I realized that the exception was resume-able...
Also, that made me think about something else... What if the debugger somehow made it immediately obvious whether the exception was resume-able? That could be useful in cases like these to make the user's options clearer.
We could solve that together with this: https://pharo.fogbugz.com/f/cases/14097/generated-code-non-existing-package-... <https://pharo.fogbugz.com/f/cases/14097/generated-code-non-existing-package-...> if you generate code via meta programming, the resulting methods now have to be in a package. Which is not good, as it will be dirty. Maybe we could have a general place for all these? Marcus
Marcus Denker-4 wrote
Maybe we could have a general place for all these?
_UnpackagedPackage? ----- Cheers, Sean -- View this message in context: http://forum.world.st/Empty-Package-Names-tp4803642p4803660.html Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.
2015-02-04 17:13 GMT+01:00 Sean P. DeNigris <sean@clipperadams.com>:
Marcus Denker-4 wrote
Maybe we could have a general place for all these?
_UnpackagedPackage?
Yes ! Esteban, this package is there for that reason, isn't it? Do we have the special logic which makes it a fall-through for code created without package for whatever reason? Thierry
----- Cheers, Sean -- View this message in context: http://forum.world.st/Empty-Package-Names-tp4803642p4803660.html Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.
participants (3)
-
Marcus Denker -
Sean P. DeNigris -
Thierry Goubier