[GR] update Qt6Base_jll

[t-bltg]

In Plots CI, this was the reason for the ColorXXX and ImageXXX stack being left behind latest versions (see below).

Cc @jheinen, since changing compat bounds are always risky in Yggdrasil.

From PlotsBase test suite, using Pkg.status(; outdated = true, mode = Pkg.PKGMODE_MANIFEST):
before PR

⌃ [35d6a980] ColorSchemes v3.26.0 (<v3.29.0)
⌅ [3da002f7] ColorTypes v0.11.5 (<v0.12.1): ColorSchemes, ColorVectorSpace, Colors, TiffImages
⌅ [c3611d14] ColorVectorSpace v0.9.10 (<v0.11.0): ColorSchemes, ImageCore
⌅ [5ae59095] Colors v0.12.11 (<v0.13.0): ColorSchemes, ImageCore
⌃ [2803e5a7] ImageAxes v0.6.11 (<v0.6.12)
⌃ [c817782e] ImageBase v0.1.5 (<v0.1.7)
⌅ [a09fc81d] ImageCore v0.9.4 (<v0.10.5): ImageBase, ImageMagick, ImageSegmentation, QuartzImageIO
⌃ [6a3955dd] ImageFiltering v0.7.6 (<v0.7.9)
⌃ [82e4d734] ImageIO v0.6.8 (<v0.6.9)
⌃ [6218d12a] ImageMagick v1.2.1 (<v1.4.1)
⌃ [bc367c6b] ImageMetadata v0.9.9 (<v0.9.10)
⌃ [80713f31] ImageSegmentation v1.8.1 (<v1.8.4)
⌃ [916415d5] Images v0.26.1 (<v0.26.2)
⌅ [033835bb] JLD2 v0.4.54 (<v0.5.13): MetaGraphs
⌅ [626554b9] MetaGraphs v0.7.2 (<v0.8.0): ImageSegmentation
⌅ [aea7be01] PrecompileTools v1.2.1 (<v1.3.2): julia
⌃ [dca85d43] QuartzImageIO v0.7.4 (<v0.7.5)
⌅ [df47a6cb] RData v0.8.3 (<v1.0.0): RDatasets
⌅ [731e570b] TiffImages v0.10.2 (<v0.11.3): ImageIO
⌃ [7b86fcea] ATK_jll v2.38.0+0 (<v2.38.1+0)
⌃ [83423d85] Cairo_jll v1.18.4+0 (<v1.18.5+0)
⌅ [b22a6f82] FFMPEG_jll v4.4.4+1 (<v7.1.0+0): FFMPEG
⌃ [7746bdde] Glib_jll v2.82.4+0 (<v2.84.0+0)
⌅ [c73af94c] ImageMagick_jll v6.9.13+0 (<v7.1.1+1): ImageMagick
⌅ [e9f186c6] Libffi_jll v3.2.2+2 (<v3.4.7+0): Glib_jll, HarfBuzz_jll, Wayland_jll
⌅ [c0090381] Qt6Base_jll v6.7.1+1 (<v6.8.2+1): GR_jll, Qt6Declarative_jll, Qt6ShaderTools_jll, Qt6Wayland_jll
⌅ [629bc702] Qt6Declarative_jll v6.7.1+2 (<v6.8.2+1): Qt6Wayland_jll
⌅ [ce943373] Qt6ShaderTools_jll v6.7.1+1 (<v6.8.2+1): Qt6Declarative_jll
⌃ [e99dba38] Qt6Wayland_jll v6.7.1+1 (<v6.8.2+0)
⌃ [a2964d1f] Wayland_jll v1.21.0+2 (<v1.23.1+0)
⌅ [02c8fc9c] XML2_jll v2.13.6+1 (<v2.14.1+0): Gettext_jll, Wayland_jll, at_spi2_atk_jll
⌅ [b437f822] adwaita_icon_theme_jll v3.33.93+0 (<v43.0.1+0): Gtk
⌅ [1270edf5] x264_jll v2021.5.5+0 (<v10164.0.1+0): FFMPEG_jll
⌅ [dfaa095f] x265_jll v3.5.0+0 (<v4.1.0+0): FFMPEG_jll

after PR

⌅ [aea7be01] PrecompileTools v1.2.1 (<v1.3.2): julia
⌅ [df47a6cb] RData v0.8.3 (<v1.0.0): RDatasets
⌃ [7b86fcea] ATK_jll v2.38.0+0 (<v2.38.1+0)
⌃ [83423d85] Cairo_jll v1.18.4+0 (<v1.18.5+0)
⌅ [b22a6f82] FFMPEG_jll v4.4.4+1 (<v7.1.0+0): FFMPEG
⌃ [7746bdde] Glib_jll v2.82.4+0 (<v2.84.0+0)
⌅ [88015f11] LERC_jll v3.0.0+1 (<v4.0.1+0): Libtiff_jll
⌅ [e9f186c6] Libffi_jll v3.2.2+2 (<v3.4.7+0): Glib_jll, HarfBuzz_jll, Wayland_jll
⌅ [89763e89] Libtiff_jll v4.4.0+0 (<v4.7.1+0): ImageMagick_jll, LittleCMS_jll, OpenJpeg_jll, gdk_pixbuf_jll
⌃ [d3a379c0] LittleCMS_jll v2.12.0+0 (<v2.16.0+0)
⌃ [643b3616] OpenJpeg_jll v2.4.0+0 (<v2.5.4+0)
⌃ [a2964d1f] Wayland_jll v1.21.0+2 (<v1.23.1+0)
⌅ [02c8fc9c] XML2_jll v2.13.6+1 (<v2.14.1+0): Gettext_jll, Wayland_jll, at_spi2_atk_jll
⌅ [b437f822] adwaita_icon_theme_jll v3.33.93+0 (<v43.0.1+0): Gtk
⌃ [da03df04] gdk_pixbuf_jll v2.42.8+0 (<v2.42.12+0)
⌃ [c5f90fcd] libwebp_jll v1.4.0+0 (<v1.5.0+0)
⌅ [1270edf5] x264_jll v2021.5.5+0 (<v10164.0.1+0): FFMPEG_jll
⌅ [dfaa095f] x265_jll v3.5.0+0 (<v4.1.0+0): FFMPEG_jll

Tested locally (GR test suite + Plots test suite) on linux.
Might be worth testing on macos & windows as well ...

[t-bltg]

Crap, I'm clueless / stuck with this libQt6EntryPoint.a error, on windows x32 :

[11:27:26] /opt/i686-w64-mingw32/bin/../lib/gcc/i686-w64-mingw32/10.2.0/../../../../i686-w64-mingw32/bin/ld: /workspace/destdir/lib/libQt6EntryPoint.a: error adding symbols: file format not recognized
[11:27:28] make[2]: *** [CMakeFiles/gksqt.dir/build.make:184: gksqt.exe] Error 1
[11:27:28] make[2]: Leaving directory '/workspace/srcdir/gr/build'
[11:27:28] make[1]: *** [CMakeFiles/Makefile2:742: CMakeFiles/gksqt.dir/all] Error 2
[11:27:28] make[1]: *** Waiting for unfinished jobs....

EDIT: as pointed out in #11052 (comment), it looks like an ICE.

[barche]

I think this also needs a version bump on GR itself, due to the changing compat bounds on the dependencies. Best to do this when there is a new version of GR.

[giordano]

Version number must change.

[barche]

Just tried locally, you need to switch to GCC 13 on Windows.

[barche]

Note that this still can't be merged without changing the JLL version, which is currently in sync with the upstream GR version. I think it's best to hold off merging this until the next upstream version is released, @jheinen ?

[eschnett]

Keeping upstream versions and jll versions in sync has mostly cosmetic value for the jll package maintainers. Waiting for a new upstream release is fine, but so is artificially increasing the version number of the jll package. Look e.g. at the recipe for Ncurses where a new version might take years.