07:11 nickynick: Hi all, where can I found the official wayland source code?
07:12 wb9688: nickynick: https://gitlab.freedesktop.org/wayland
07:12 wb9688: But what are you looking for?
07:13 nickynick: thanks, I was Just searching and got that link, but didn't sure that it is official or not
07:13 nickynick: I am trying to look into the source code, I am only an intermediate programmer in C
07:14 nickynick: so just looking into that source code, it seems to be very very small that I expected
07:16 wb9688: Yes, that's basically just the protocol
07:16 wb9688: There are then many compositors that implement it, which have have a lot more code
07:19 nickynick: oh! ok
07:23 wb9688: Sway, Wayfire, etc use the wlroots library to help implement that, but e.g. GNOME has their completely own implementation
07:31 Company: Wayland itself is about 2 things:
07:31 Company: 1. the pretty basic mechanism to send messages between apps and compositors
07:32 Company: 2. a tool to generate source code from XML protocol definitions that define those messages
07:32 Company: so that sending and receiving messages can be done correctly by apps and compositors
08:45 Company: swick[m], zamundaaa[m]: a random thing I just noticed is color management for yuv fourccs - is that outlined somewhere how those should be handled?
08:45 Company: I mean, I think it's pretty straightforward, but it might warrant spelling out
09:00 Company: after checking in detail, I suppose that whole topic goes in color-representation?
09:01 Company: because YUV ends up as regular sRGB once it reaches the color-management parts?
09:26 swick[m]: not sRGB but RGB, yes
09:37 Company: I tend to say sRGB as a shorthand for RGB with sRGB transfer function
09:37 Company: actually, I even mean sRGB primaries in this case
09:38 Company: but those match the non-YUV fourcss
10:51 wb9688: In KDE server_decoration, is a server always supposed to directly emit a mode event after a request_mode request?
10:55 wb9688: "The client can request a different mode through the decoration request. The server will acknowledge this by another event with the same mode." seems so
10:56 wb9688: However I noticed with GTK 4 on Wayfire as well as Sway (both wlroots compositors), that the mode event is not sent after the request_mode request
14:28 d_ed[m]: <wb9688> "In KDE server_decoration, is a..." <- this shouldn't be used anymore since xdg_decoration
14:29 wb9688: d_ed[m]: GTK only supports KDE server_decoration, not xdg_decoration
14:30 d_ed[m]: yeah :(
14:34 wb9688: So we have to keep supporting it for now and the bug should still be fixed somehow
14:44 Company: somebody could go and update GTK to support the new one
14:45 Company: but afaik all attempts so far included changing things to "GTK MUST ALWAYS DO SSD!!!!111eleven" and then got rejected
14:45 alice: you mean csd? :D
14:46 Company: no, I mean ssd
14:46 Company: the MRs wanted to force SSD on GTK
14:56 ifreund: Company: I wrote the first xdg-deco patch, and I abused the xdg-deco protocol in gtk's favor to force CSD if gtk wanted
14:56 ifreund: (by destroying the xdg decoration object)
14:56 ifreund: I had all the GDK parts merge ready but struggled to get the GTK bits working)
14:56 ifreund: and then I lost interest, this was like 4 years ago
14:57 Company: that sounds like it was during or right after the 4.0 "crunch"
17:26 swick[m]: Company: converting YCbCr doesn't necessarily result in sRGB
17:27 Company: yeah, but I was talking about the default behavior
17:27 Company: and by default everyone assumes you end up with srgb
17:27 swick[m]: there is no default behavior. it depends on what the content is.
17:28 Company: the default behavior is what you get when you create a buffer with fourcc NV12 and send it over
17:28 swick[m]: that's undefined behavior, not default behavior
17:28 Company: every compositor treats it as narrow range, BT601 matrix, srgb primaries, srgb tf
17:29 Company: I would file bugs against compositors that don't, but yes - technically it's undefined behavior
17:31 Company: I would also define it, or at least recommend it
17:32 Company: like the cm spec does for untagged surfaces