From 5d6e7115a965f1c079e5fc872d89d077ac174ed3 Mon Sep 17 00:00:00 2001 From: Pekka Paalanen Date: Tue, 20 Feb 2024 13:54:06 +0200 Subject: [PATCH] color-lcms: make identity_pipeline comment more detailed This perhaps better explains what the issue here is. Signed-off-by: Pekka Paalanen --- libweston/color-lcms/color-lcms.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/libweston/color-lcms/color-lcms.c b/libweston/color-lcms/color-lcms.c index 5fa01af04..4bad306fd 100644 --- a/libweston/color-lcms/color-lcms.c +++ b/libweston/color-lcms/color-lcms.c @@ -104,9 +104,12 @@ cmlcms_get_surface_color_transform(struct weston_color_manager *cm_base, surf_xform->transform = &xform->base; /* - * When we introduce LCMS plug-in we can precisely answer this question - * by examining the color pipeline using precision parameters. For now - * we just compare if it is same pointer or not. + * TODO: Instead of this, we should create the INPUT_TO_OUTPUT color + * transformation and check if that is identity. Comparing just the + * profiles will miss image adjustments if we add some. + * OTOH, that will only be useful if DRM-backend learns to do + * opportunistic direct scanout without KMS blending space + * transformations. */ if (xform->search_key.input_profile == xform->search_key.output_profile) surf_xform->identity_pipeline = true;