**** BEGIN LOGGING AT Tue Jul 29 03:00:00 2014 Jul 29 07:41:46 morning all Jul 29 07:54:26 hi bl Jul 29 08:05:11 hi woglinde Jul 29 08:38:18 howdy good people Jul 29 08:38:25 I have a feature request Jul 29 08:38:30 awesome feature Jul 29 08:39:29 I neet automagic buttons that allow me to be able to create a recipy that basically constitutes the diff between two other packages Jul 29 08:39:49 we have devices in the field I cannot send 300Mb every time I make a minor change to a recipy Jul 29 08:40:25 isn't that what a package manager is for? per-package upgrades Jul 29 08:40:34 ...constitutes the diff between the same package different version Jul 29 08:40:47 but it sucks the entire payload again Jul 29 08:40:50 I only want a diff Jul 29 08:41:01 like for example Jul 29 08:41:12 our android used to be launched from sysvinit Jul 29 08:41:15 I changed it to systemd Jul 29 08:41:22 why do I need to send gigabytes Jul 29 08:41:44 its one minor change, delete /etc/init.d/android and insert new file at /lib/systemd/system Jul 29 08:41:53 surely I am not the first to have this problem Jul 29 08:42:13 1.5 -> 1.6 somehow killed sysvinit Jul 29 08:43:36 if your bitbake is setup correctly, you might be able to create a system where you could call build a new package but subtract files that are the same in the old package? Jul 29 08:45:14 currenly I cant even do an opkg upgrade Jul 29 08:45:20 even when I am directly connected to the package source Jul 29 08:46:14 if my rootfs is X meggerbytes now I need to have 2X space floating on my device for if in case I need to do a opkg upgrade Jul 29 08:46:30 terrible Jul 29 08:48:14 your rootfs is several gigabytes? Jul 29 08:50:50 but yeah, i agree, the upgrade path for changing to another PROVIDER of something, for example startup program, is probably not trivial Jul 29 09:27:36 my rootfs is 779 Mb Jul 29 09:27:52 basically android ownage Jul 29 09:28:40 basically technology causing these situations Jul 29 09:28:50 normally an embedded device in the field won' Jul 29 09:29:05 normally an embedded device in the field won't be a full blown linux OS, but these days you can do it no problem Jul 29 09:35:08 wow, 779 mb. that's a lot! Jul 29 12:23:01 what's the status about gcc 4.9.0 and linux kernel on OE ? :p Jul 29 16:29:33 bluelightning ping Jul 29 16:29:43 woglinde: pong Jul 29 16:30:48 bluelightning are you familiar with RREPLACES? Jul 29 16:31:26 we have a packages which is splitted into serval libs now and the main lib packages goes from libname to libname3 Jul 29 16:31:55 how do we set RREPLACES so the libname3 replaces libname Jul 29 16:32:10 I only found examples for the other way round Jul 29 16:48:01 hm maybe DEBIANNAME_${PN} helps Jul 29 16:59:13 nope Jul 29 17:57:33 Is it safe to rm the biteable build/cache dir, or would that cause BitBAke to rebuild *everything* — I made a mistake in a recipe, and it seems just changing the file isn’t making bitbake re-generate the cache. Jul 29 17:57:46 s/bitable/bitbake/ **** ENDING LOGGING AT Wed Jul 30 03:00:01 2014