09:02 MoeIcenowy: I have a weird question about whether to creating a directory in drivers/gpu/drm/bridge for one or two vendor: The T-Head TH1520 SoC design is transferred from T-Head to Zhihe, and Zhihe created A210 SoC, which reuses TH1520's HDMI and DSI bridge, but added a new DP bridge
09:02 MoeIcenowy: (and TH1520 is technically rebranded to Zhihe A200, but no devices with A200 is seen in public
09:02 MoeIcenowy: If I am going to create a directory, should it be named thead (in this case Zhihe A210 DP is odd here) or zhihe (Technically TH1520 isn't a Zhihe thing although A200 is)
11:28 simon-perretta-img: Maybe naming it zhihead would be a good compromise /j
13:24 MoeIcenowy: simon-perretta-img: best cold joke today
14:38 dcbaker: gfxstrand, anholt, karolherbst: I was just looking at some of the Rust Meson code. It seems like there's some improvements that we could make on the Meson side around bindgen. I threw https://github.com/mesonbuild/meson/pull/16312 together, I'm wondering if any of that looks useful?
14:42 karolherbst: dcbaker: nooot sure... long-term I want to move to our custom bindgen binary, because the API is more powerful than what the binary exposes
14:43 karolherbst: not sure if wrapping the allowlist/blocklist is really helping there anyway?
15:05 dcbaker: I just sorta peaked at where there was a lot of duplication. There's a lot of --allowlist-* in our Meson, so being able to just write allowlist : {'var' : ['NIR_.*', 'nir_.*_infos', 'rust_.*']} seemed like an improvment over the existing API of passing --allowlist-var over and over
15:06 karolherbst: but then the ordering gets weird
15:07 dcbaker: That's true, I guess the question is whether less LOC is better, or grouping things together is better
15:19 dcbaker: eric_engestrom: gentle ping on https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/45449, if you're interested in looking