This applies to P61 (for some time) and P7.
Object compile: '^'. Syntax Error popup -> close now you have two morphic UI processes running and World starts to get corrupted with artifacts around. (and it also in one case manage to actually corrupt the image to the point that I was no longer able to even launch it). Peter |
Never mind... I see now what is broken and I'll try to fix it (assuming my P7 image won't corrupt itself due to FreeType before I manage to push the changes... heh) On Sat, Nov 11, 2017 at 3:43 PM, Peter Uhnák <[hidden email]> wrote:
|
> On 11 Nov 2017, at 16:33, Peter Uhnák <[hidden email]> wrote: > > Never mind... I see now what is broken and I'll try to fix it (assuming my P7 image won't corrupt itself due to FreeType before I manage to push the changes... heh) You can now disable FreeType fonts and still use bitmapped Source Code fonts. > On Sat, Nov 11, 2017 at 3:43 PM, Peter Uhnák <[hidden email]> wrote: > This applies to P61 (for some time) and P7. > > Object compile: '^'. > > Syntax Error popup -> close > > now you have two morphic UI processes running and World starts to get corrupted with artifacts around. (and it also in one case manage to actually corrupt the image to the point that I was no longer able to even launch it). > > Peter > |
Can you share what was the original problem and how you fixed it? On Sat, Nov 11, 2017 at 4:44 PM, Sven Van Caekenberghe <[hidden email]> wrote:
|
Hi, this is a bit more confusing than I originally thought. So let me elaborate (this is not part of Pharo I've dabbled in before...) (there's an issue 20212 for this) The problem is somewhere in SyntaxErrorDebugger and probably SyntaxErrorNotification. Normally (by normally I mean according to the implementation) when an syntax debugger is created, it is created in new process. However when I close the window without saving, the process is not terminated. Some one has decided that SyntaxErrorDebugger>>release was not good enough, so they've added a WindowClosed catch (in SyntaxErrorDebugger class>>buildMorphicViewOn: ) that in turn calls SyntaxErrorDebugger>>handleWindowClosed. That probably fixed whatever the author was trying to fix, but it broke overall. The expected behavior should be: 1) when a debugger window appears and I close it, the process should terminate -- it doesn't 2) when a debugger window appears and I enter a valid text and save it, 2.i) the process should terminate -- it doesn't 2.ii) the method should be compiled -- it compiles the original broken code So something like this can happen Something compile: '^^'. * enter some correct code, e.g. 'method ^ 1' * Something selectors "-> #(#method)" (Something>>#method) sourceCode "-> '^^'" I don't understand how this is even possible to have sourceCode that doesn't include the selector. 3) when a debugger window appears and I enter invalid code, then it should 2.a) either terminate the previous process and launch a new process containing the updated, but broken code 2.b) or not terminate and reuse the current process but display the new broken code right now it creates a new process, but doesn't terminate. Peter On Mon, Nov 13, 2017 at 9:42 PM, Guillermo Polito <[hidden email]> wrote:
|
Free forum by Nabble | Edit this page |