any chance of adding filters liblensfun, libtensorflow ?

[hydra3333]

Would you consider there's any chance of adding code to incorporate these new up-and-coming filters

1. liblensfun (apparently depends on 'glib2', 'png', 'zlib')

2. libtensorflow (A.I. based filtering)

I'm having no end of trouble building and/or linking in both, on my own, and your expertise (and/or that of your compatriots) would be vastly appreciated.

[rdp]

started on glib2. not easy...

On Mon, Aug 13, 2018 at 7:07 PM hydra3333 [email protected] wrote:

Would you consider there's any chance of adding code to incorporate these new up-and-coming filters

liblensfun (apparently depends on 'glib2', 'png', 'zlib') 2.

libtensorflow (A.I. based filtering)

I'm having no end of trouble building and/or linking in both, on my own, and your expertise (and/or that of your compatriots) would be vastly appreciated.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/rdp/ffmpeg-windows-build-helpers/issues/341, or mute the thread https://github.com/notifications/unsubscribe-auth/AAAw0NV_cgsBgaa65dFtJ3GEC-yLBFsbks5uQiLDgaJpZM4V7on9 .

[hydra3333]

Thank you. In case it helps, this from a deadsix27 derivative, based on AlexPux mingw works, appears to build static x64 glib2 ...

	'libglib2' : { # from Alexpux https://github.com/Alexpux/MINGW-packages/tree/master/mingw-w64-glib2
		'repo_type' : 'git',
		'url' : 'https://gitlab.gnome.org/GNOME/glib.git',
		'configure_options' : '--host={target_host} --prefix={target_prefix} --disable-shared --enable-static --disable-libelf --with-threads=posix --enable-gc-friendly --disable-fam --disable-man --disable-gtk-doc --with-pcre=system --with-libiconv --enable-libmount=no --disable-libmount --disable-selinux ',
		'patches' : [
			('https://raw.githubusercontent.com/hydra3333/h3333_python_cross_compile_script/master/patches/libglib2-from-Alexpux-2_58_0/0001-Use-CreateFile-on-Win32-to-make-sure-g_unlink-always.patch', '-Np1'), 
			('https://raw.githubusercontent.com/hydra3333/h3333_python_cross_compile_script/master/patches/libglib2-from-Alexpux-2_58_0/0001-win32-Make-the-static-build-work-with-MinGW-when-pos.patch', '-Np1'), 
			('https://raw.githubusercontent.com/hydra3333/h3333_python_cross_compile_script/master/patches/libglib2-from-Alexpux-2_58_0/disable_libmount-make-UTF-yes.patch', '-Np0' ), # 2018.08.10 mine # TODO: CHECK THIS PATCH
			# ('https://raw.githubusercontent.com/hydra3333/h3333_python_cross_compile_script/master/patches/libglib2-from-Alexpux-2_58_0/0001-glib2-mr-226.patch', '-Np1' ), # 2018.08.31 this patch is already implemented in 2.58.0 ... for same as -D_FILE_OFFSET_BITS=64 per alexpux
		],
		'run_post_patch' : [
			'rm -fv ./configure',
			'./autogen.sh NOCONFIGURE=1',
			'autoreconf -fiv',
		],
		'run_post_install' : [
			'sed -s -i.bak1 \'s/-lintl/-lintl -liconv/\' "glib-2.0.pc"',
			'sed -s -i.bak2 \'s/ -lgiowin32//g\' "glib-2.0.pc"',
			'sed -s -i.bak3 \'s/ -llgnulib//g\' "glib-2.0.pc"',
			'sed -s -i.bak4 \'s/ -lcharset//g\' "glib-2.0.pc"',
			#
			'sed -s -i.bak1 \'s/-lintl/-lintl -liconv/\' "{pkg_config_path}/glib-2.0.pc"',
			'sed -s -i.bak2 \'s/ -lgiowin32//g\' "{pkg_config_path}/glib-2.0.pc"',
			'sed -s -i.bak3 \'s/ -llgnulib//g\' "{pkg_config_path}/glib-2.0.pc"',
			'sed -s -i.bak4 \'s/ -lcharset//g\' "{pkg_config_path}/glib-2.0.pc"',
			#
			#'sed -i.bak \'s/$(cygpath -m \${MINGW_PREFIX})/\${MINGW_PREFIX}/g\' "{pkg_config_path}/glib-2.0.pc"', # TODO: CHANGE MINGW_PREFIX TO SOMETHING ELSE ...
			#'sed -i.bak \'s/$(cygpath -m \${MINGW_PREFIX})/\${MINGW_PREFIX}/g\' "glib-2.0.pc"', # TODO: CHANGE MINGW_PREFIX TO SOMETHING ELSE ...
			#'sed -i.bak \'s/$(cygpath -m ${MINGW_PREFIX})/${MINGW_PREFIX}/g" "{target_prefix}/bin/glib-gettextize" # TODO: CHANGE MINGW_PREFIX TO SOMETHING ELSE ...
		],
		'depends_on' : [ 'iconv', 'gettext', 'pcre', 'pcre2', 'libffi', 'zlib', 'python36_libs' ], # , 'libmount'
		'_info' : { 'version' : '2.58.0', 'fancy_name' : 'glib2_deadsix27' },
	},

[hydra3333]

PS I'd probably give up on liblensfun since didn't mabs have trouble building it ?

[rdp]

attempted lensfun in 7d1b6e1 since I was close on glib. tensorflow looks...scary...