Change Details

Apparently the item model changes in Qt 5.11 cause our quickopen list to be empty. I do not know why the old code does not work, to be honest, to me it seems okay. The only thing which is definitely wrong is the call to mapToSource; the index passed to this belongs to the model which has since been replaced, and not to the actual model attached to the proxy model. That issue is unrelated though, commenting out that code (and thus removing the feature) doesn't fix it. I like the suggested implementation better anyways, so I think we should change this in either case, but does anyone have an idea why the old code is broken? Is that maybe even a regression in Qt itself? @mwolff?

Apparently the item model changes in Qt 5.11 cause our quickopen list to be empty. I do not know why the old code does not work, to be honest, to me it seems okay. The only thing which is definitely wrong is the call to mapFromSource; the index passed to this belongs to the model which has since been replaced, and not to the actual model attached to the proxy model. That issue is unrelated though, commenting out that code (and thus removing the feature) doesn't fix it. With the suggested patch it's not an issue any more either because we don't change the model. I like the suggested implementation better anyways, so I think we should change this in either case, but does anyone have an idea why the old code is broken? Is that maybe even a regression in Qt itself? @mwolff?

Apparently the item model changes in Qt 5.11 cause our quickopen list to be empty. I do not know why the old code does not work, to be honest, to me it seems okay. The only thing which is definitely wrong is the call to mapTomapFromSource; the index passed to this belongs to the model which has since been replaced, and not to the actual model attached to the proxy model. That issue is unrelated though, commenting out that code (and thus removing the feature) doesn't fix it. commenting out that code (and thus removing the feature) doesn't fix itWith the suggested patch it's not an issue any more either because we don't change the model. I like the suggested implementation better anyways, so I think we should change this in either case, but does anyone have an idea why the old code is broken? Is that maybe even a regression in Qt itself? @mwolff?