09:25 ofourdan: MrCooper: hey, is it worth waiting for https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1086 for xwayland 23.1 rc2? what do you reckon?
09:26 MrCooper: if you want to cut rc2 today, I wouldn't wait
09:26 ofourdan: well, rc2 was due yesterday :)
09:28 MrCooper: I'd rather not rush that MR
09:28 ofourdan: right
09:28 ofourdan: and we have a number of fixes already piled up for rc2
09:28 ofourdan: so rc2 it will be
09:28 MrCooper: cool
09:39 pq: swick[m], we standardised the term "primary color volume", but I noticed libdisplay-info (EDID) has the term "primary color space". What's that?
09:58 mceier: wikipedia describes that bit as "Standard sRGB colour space. Bytes 25–34 must contain sRGB standard values." ( https://en.wikipedia.org/wiki/Extended_Display_Identification_Data#Structure,_version_1.4 )
10:06 mceier: primary color space probably comes from: https://glenwing.github.io/docs/VESA-EEDID-A2.pdf - "Standard Default Color Space - sRGB: If this bit is set to 1, the display uses the sRGB standard default color space as its primary color space. [...]" and it doesn't define primary color space ;)
10:08 pq: cpli, you should not be able to call wl_log() at all, it is not exported.
10:09 pq: mceier, yes, the doc for 'bool srgb_is_primary;' is what made me ask. :-)
11:31 cpli: thanks pq, https://git.sr.ht/~cpli/anime/tree/dev/item/src/log.c
12:45 swick[m]: pq: looks like we copied that from edid-decode
12:46 pq: swick[m], mceier, pointed out the wording is in a spec, too.
12:46 swick[m]: jup, just checked it myself...
12:46 swick[m]: but it's the only mention of it
12:47 pq: maybe it means "when nothing else is indicated" or something
12:48 swick[m]: I think it's just badly worded in the spec and what it wants to say is that the default colorimetry matches the sRGB colorimetry
12:48 pq: so a shortcut to not need to compare all the parameters?
12:49 swick[m]: yes
12:49 swick[m]: i.e., it also says 'If this bit is set, the color information in section 3.7 must match the sRGB standard values.' and 3.7 carries RGBW chromaticities
12:49 pq: alright
12:50 vsyrjala: and probably half the displys will fill in the data with contradicting data. which one do you trust then?
12:50 pq: nothing new :-)
12:52 pq: maybe we should trust the bit, since it's easier to fail to set the colorimetry parameters
12:52 pq: and if the bit is unset, then look at the parameters which could still be sRGB
20:26 DodoGTA: Is there a good way to improve the protocol fragmentation problem?