On 22 October 2012 16:41, Phil Blundell <[email protected]> wrote: > It seems that glib-2.0 has had a dependency on zip for as far back as I > could be bothered to trace the history. However, I couldn't immediately > see any reason why this should be the case. Does anyone know where this > came from and what purpose it serves?
Well, in oe-core it came from: $ git bisect good 31180a56d1b6981fa9d5588ad3538f34e339e0fe is the first bad commit commit 31180a56d1b6981fa9d5588ad3538f34e339e0fe Author: Kevin Tian <[email protected]> Date: Fri Jun 25 15:07:01 2010 +0800 glib-2.0: upgrade to 2.24.1 (borrow from OE 2.24.0, with study about reason for detail changes) [Patches-Poky] KEEP _configure-libtool.patch_: fix cross-build env REMOVE _glibconfig-sysdefs.h_: introduced in glib-1.2.10, which is actually not valid today KEEP _glib-gettexize-dir.patch_: another quite old patch added before 2.10.3 add history info, and mark for upstream [Patches-OE] NOTAKE _gio.patch_: a private extension for a window matching library (wncksync) NOTAKE _gvavriant.patch_: in 2.24.1 now NOTAKE _gasync.patch_: in 2.24.1 now TAKE _60_wait-longer-for-threads-to-die.patch: said to fix anonying warning TAKE _gatomic-proper-get-cast.patch_: fix error about type casting. Instead of using OE's patch, I found a better one from latest repo NOTAKE _g_once_init_enter.patch_: in OE since 2.22.1, not quite sure of its purpose, also it's not in upstream yet [Recipe] - fix metadata - keep glib-2.inc which is for python optimization - remove PRIORITY since 'optional' is the default - no need to specially install glibconfig.h in ad hoc path, which is only relied by 'obexpush' which poky doesn't have and can be fixed in 'obexpush' itself as what OE does - no need to manually create glib2.0/glib, which is handled by current code - remove 'acpaths' which only applies to versions before 2.12.4 - below lines are not taken from OE * glib-mkenums tweaks, which exists in OE for a long time but never in poky. since there's no way to find out original commit, keep poky's style instead. * host libtool check. not sure about its purpose, and it's also in OE for quite long time Signed-off-by Kevin Tian <[email protected]> I don't have an oe-classic tree to hand but some casual digging doesn't reveal any zip dependencies. There's no good reason for that dependency that I'm aware of (grep certainly doesn't show anything relevant), so I say kill it. Ross _______________________________________________ Openembedded-core mailing list [email protected] http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
