https://bugs.kde.org/show_bug.cgi?id=447293
--- Comment #4 from amyspark <a...@amyspark.me> --- Definitely a bug in our image conversion code. Currently, we run the G'MIC->Krita layer name parsing only if there is an existing node that is going to be overwritten (via KisImportQmicProcessingVisitor traversal). In this case, there are more output than input layers, so the previous step (KisQmicSynchronizeLayersCommand) will take care of the excess nodes... except that it never runs the name parsing step, so the appropriate layer modes are not set. While at it, I found another bug regarding Krita->G'MIC layer naming; when I implemented G'MIC's format for layer names, I actually never packed it inside the "message", so G'MIC returns nodes that use it as "[unnamed]". I'll address both bugs in the MR. -- You are receiving this mail because: You are watching all bug changes.