**** BEGIN LOGGING AT Tue Apr 19 02:59:57 2022
Apr 19 07:26:34 <Herrie> Morning!
Apr 19 07:39:50 <Herrie> I can confirm it's failing for Hammerhead too
Apr 19 07:45:07 <Herrie> And for Mako
Apr 19 07:45:10 <Herrie> Same error
Apr 19 08:01:49 <JaMa> it's TUNE_PKGARCH so not surprising as it's build just once for all these 3
Apr 19 08:26:20 <Tofe> Morning!
Apr 19 08:27:14 <Tofe> JaMa: for armv7 I don't think it's worth including any gl3 :)
Apr 19 08:33:50 <JaMa> agreed, I just wonder since when it is broken, did anyone build luna-surfacemanger with LuneOS and arm libhybris before?
Apr 19 08:34:22 <JaMa> without jenkins it's difficutl to keep track what was actually built completely
Apr 19 08:34:31 <HerrieTP> JaMa: I think we might have not tried it for armv7
Apr 19 08:34:48 <Tofe> no, so far we only built for pinephone (mesa, aarch64) and qemu
Apr 19 08:34:56 <HerrieTP> I've only built PPP and qemux86-64 at my end
Apr 19 08:35:00 <JaMa> I haven't completely built any arm target in 2022 until yesterday (because my multilib in gentoo was a bit broken and I had to fix it to build webruntime for 32bit)
Apr 19 08:35:20 <HerrieTP> I'm not sure we tried Halium aarch64 even
Apr 19 08:35:34 <Tofe> I don't think we tried yet
Apr 19 08:35:43 <JaMa> and I'm also building only kirkstone
Apr 19 08:35:57 <Herrie> When my PPP is done, will kick off a Tissot or Mido build to see
Apr 19 08:44:19 <Herrie> webruntime almost done here for PPP, so I guess I can kick off a Mido build soon
Apr 19 10:11:16 <Herrie> Also failing for mido
Apr 19 10:13:12 <Tofe> ok so it's generic for halium targets
Apr 19 10:42:52 <Herrie> Yup seems so
Apr 19 10:43:15 <Herrie> Could be we need to add a DEPENDS on libhybris for Halium targets to luna-surfacemanager simply?
Apr 19 10:50:20 <Tofe> Herrie: should already be the case, indirectly; but we should check that the GL headers are all taken from android-headers
Apr 19 11:20:48 <Herrie> Tofe: When I do bitbake -g luna-surfacemanger I do get android-headers-halium in my pn-depends file
Apr 19 11:20:59 <Herrie> ehm pn-buildlist I mean
Apr 19 11:27:32 <Herrie> And my task-depends.dot: https://paste.ubuntu.com/p/rp9C8MSDBB/
Apr 19 11:49:40 <JaMa> it's not missing dependency see https://github.com/webOS-ports/meta-webos-ports/commit/a294c8a2028a3bfacc0657a46c9a1337a211af90
Apr 19 11:50:46 <JaMa> qtbase has gles 3.2 enabled:
Apr 19 11:50:48 <JaMa>   OpenGL:
Apr 19 11:50:48 <JaMa>     Desktop OpenGL ....................... no
Apr 19 11:50:48 <JaMa>     OpenGL ES 2.0 ........................ yes
Apr 19 11:50:48 <JaMa>     OpenGL ES 3.0 ........................ yes
Apr 19 11:50:51 <JaMa>     OpenGL ES 3.1 ........................ yes
Apr 19 11:50:53 <JaMa>     OpenGL ES 3.2 ........................ yes
Apr 19 11:51:18 <JaMa> will check pp build after finishing tissot, I guess it uses 2.0
Apr 19 11:58:10 <Herrie> JaMa: Thxn testing for Mido here now
Apr 19 12:00:49 <Tofe> JaMa: well, libhybris does provide GLES 3 if need be
Apr 19 12:01:02 <Tofe> it's just that the headers don't match
Apr 19 12:01:18 <Tofe> Maybe we should patch android-headers or something like that
Apr 19 12:03:01 <JaMa> qtbase is already using GLES 3, just luna-surfacemanager wasn't respecting that, now that's fixed and switching to GLES 2 would work as well
Apr 19 12:09:55 <JaMa> ./hybris/include/GLES2/gl2.h:GL_APICALL void GL_APIENTRY glShaderSource (GLuint shader, GLsizei count, const GLchar** string, const GLint* length);
Apr 19 12:09:58 <JaMa> ./hybris/include/GLES3/gl3.h:GL_APICALL void GL_APIENTRY glShaderSource (GLuint shader, GLsizei count, const GLchar *const*string, const GLint *length);
Apr 19 12:10:59 <Tofe> JaMa: I wonder which one is the most "correct" one
Apr 19 12:11:36 <Tofe> https://www.khronos.org/registry/OpenGL-Refpages/gl4/html/glShaderSource.xhtml this matches gl2.h 
Apr 19 12:12:23 <JaMa> I'm checking mesa now
Apr 19 12:13:42 <JaMa> the gl3.h version is in mesa-22.0.1
Apr 19 12:13:44 <JaMa> mesa $ grep glShaderSource include/GLES2/gl2.h include/GLES3/gl3.h include/GL*/gl*.h
Apr 19 12:13:48 <JaMa> include/GLES2/gl2.h:GL_APICALL void GL_APIENTRY glShaderSource (GLuint shader, GLsizei count, const GLchar *const*string, const GLint *length);
Apr 19 12:13:51 <JaMa> include/GLES3/gl3.h:GL_APICALL void GL_APIENTRY glShaderSource (GLuint shader, GLsizei count, const GLchar *const*string, const GLint *length);
Apr 19 12:13:54 <JaMa> include/GL/glcorearb.h:GLAPI void APIENTRY glShaderSource (GLuint shader, GLsizei count, const GLchar *const*string, const GLint *length);
Apr 19 12:13:57 <JaMa> include/GL/glext.h:GLAPI void APIENTRY glShaderSource (GLuint shader, GLsizei count, const GLchar *const*string, const GLint *length);
Apr 19 12:14:00 <JaMa> include/GL/glext.h:GLAPI void APIENTRY glShaderSourceARB (GLhandleARB shaderObj, GLsizei count, const GLcharARB **string, const GLint *length);
Apr 19 12:14:03 <JaMa> include/GLES2/gl2.h:GL_APICALL void GL_APIENTRY glShaderSource (GLuint shader, GLsizei count, const GLchar *const*string, const GLint *length);
Apr 19 12:14:06 <JaMa> include/GLES3/gl3.h:GL_APICALL void GL_APIENTRY glShaderSource (GLuint shader, GLsizei count, const GLchar *const*string, const GLint *length);
Apr 19 12:14:09 <JaMa> include/GLES3/gl31.h:GL_APICALL void GL_APIENTRY glShaderSource (GLuint shader, GLsizei count, const GLchar *const*string, const GLint *length);
Apr 19 12:14:12 <JaMa> include/GLES3/gl32.h:GL_APICALL void GL_APIENTRY glShaderSource (GLuint shader, GLsizei count, const GLchar *const*string, const GLint *length);
Apr 19 12:14:27 <Tofe> :/
Apr 19 12:14:51 <Tofe> so either we align with Khronos specs, or with Mesa
Apr 19 12:15:32 <Tofe> I would go with Mesa, to have the most uniform headers possible across our targets
Apr 19 12:15:34 <JaMa> https://github.com/mesa3d/mesa/commit/117d8ce27b7929c42f46f818ac038580877b3086
Apr 19 12:18:05 <JaMa> mesa uses khronos as in https://www.khronos.org/registry/OpenGL/api/GLES2/gl2.h
Apr 19 12:19:17 <Herrie> I can confirm mido now builds
Apr 19 12:19:19 <Tofe> ok, so then we patch gl2.h
Apr 19 12:20:26 <Tofe> though, if your LSM patch is enough, that can wait a bit :)
Apr 19 12:20:52 <JaMa> yes, LSM patch is enough and should be merged independently on fixing libhybris
Apr 19 12:20:55 <Herrie> Testing Tenderlon now
Apr 19 12:22:07 <JaMa> PPP qtbase uses GLES 3.2 as well, but doesn't fail as mesa headers are consistent
Apr 19 12:55:22 <JaMa> the gl2.h in libhybris was imported from rev 10602 $ on $Date:: 2010-03-04 22:35:34 -0800 #$ */ which was later renamed as api/old/2.0/gl2.h in old subversion repo in rev 23325 https://github.com/shr-project/khronos-gles-svn/commit/ff2be2e31c8c91fe1fd287cb011dc089ef1a53db#diff-0e54e9f14fa48a644d7173432e4cc776cb2597ac691a648851a281d90eaef709R571 there is a lot of changes in libhybris headers when trying 
Apr 19 12:55:28 <JaMa> to refresh from latest https://github.com/KhronosGroup/OpenGL-Registry
Apr 19 12:56:50 <JaMa> glShaderSource was updated in rev 16803 https://github.com/shr-project/khronos-gles-svn/commit/5d52a99b2f8a0e97599151294cace9c44776408c
Apr 19 12:58:29 <Tofe> wow, what a mess
Apr 19 12:58:45 <Tofe> but then we just have to update the libhybris header
Apr 19 12:59:46 <JaMa> unfortunately https://www.khronos.org/bugzilla is now just redirect to github, so not sure what bug 7157 was about
Apr 19 14:05:11 <Herrie> tenderloin built OK too
Apr 19 14:09:20 <Herrie> mako too
Apr 19 14:19:59 <Herrie> hammerhead too
Apr 19 14:25:44 <Herrie> Tissot too
Apr 19 14:26:33 <JaMa> I've already merged it after building tenderloin, hammerhead, tissot, ppp, qemux8-64
Apr 19 14:27:29 <Herrie> OK well this is with honister, you tested kirkstone, so we should be good
Apr 19 15:13:45 <Herrie> Tofe: the core-apps and app-services PR's can be merged or ?
Apr 19 15:14:24 <Tofe> yes, that should be fine
Apr 19 16:12:35 <HerrieTP_> OK merged
Apr 19 16:12:52 <Tofe> thx
Apr 19 16:13:17 <Tofe> I rework a bit the other one, not sure why it didn't work
**** ENDING LOGGING AT Wed Apr 20 02:59:56 2022