**** BEGIN LOGGING AT Thu Jul 03 03:00:01 2014 Jul 03 07:39:25 ant_work: hi Jul 03 07:39:39 gm Jul 03 07:39:47 ant_work: I've fixed opie calibration issue, wanna patch to try? :) Jul 03 07:40:10 yes please ;) Jul 03 07:40:21 were my results consistent? Jul 03 07:40:36 same mismatches as yours? Jul 03 07:42:44 ant_work: well, just see the patch :) Jul 03 07:43:27 basically, qte should use tslib_read_raw when uncalibrated, but it used tslib_read instead Jul 03 07:43:53 and did something fancy in interpolateSample() method Jul 03 07:45:04 how old is that code? Jul 03 07:46:30 anarsoul: some drivers do already interpolate Jul 03 07:46:46 but didn't years ago Jul 03 07:47:00 so maybe that's the reason for the interp. Jul 03 07:47:18 anyway, great job, thx Jul 03 07:47:44 I just didn't find the time to examine the logs attentively Jul 03 07:47:55 ant_work: it just read several samples than then fall through to tslib_read() Jul 03 07:48:14 if you remember there was an issue with ads7846 jitter Jul 03 07:48:31 I don't have any devices with ads7846 Jul 03 07:48:36 mostly evident in x Jul 03 07:49:08 but there's a jitter with s3c2410 ts, and it's workarounded in a driver, it just averages several samples Jul 03 07:50:01 anyway, it's either tslib or driver job to provide accurate values, end user (i.e. qte) should not be aware of differences in ts drivers Jul 03 07:52:13 right Jul 03 07:59:07 btw, how do I rebuild image? bitbake opie-image doesn't install new libqte into rootfs... Jul 03 08:01:48 try -c cleansstate Jul 03 08:06:39 ant_work: doesn't help Jul 03 08:07:05 libqte.so in ipk and in rootfs differs Jul 03 08:07:59 ok, remove tmp and keep sstate-cache Jul 03 08:08:11 it will rebuild from sstate packages Jul 03 08:09:57 ant_work: hm, I think there should be a way to avoid full rebuild Jul 03 08:10:17 afair there is some chain of packages in Opie Jul 03 08:10:48 you have to cleansstate the parent Jul 03 08:11:16 I've cleansstated opie-image Jul 03 08:14:26 anarsoul: in those cases just rebuild from sstate. Takes a couple of mins Jul 03 08:14:39 sstate? Jul 03 08:15:02 yes, is a cache of the built packages Jul 03 08:15:07 and their sig Jul 03 08:15:24 and how to do that? Jul 03 08:15:25 so if the checksum matches it just unpack it Jul 03 08:16:03 it is the revised 'packaged staging' Jul 03 08:16:44 there is a task populating the sstate cache of each recipe you build Jul 03 08:16:52 it's automatic Jul 03 08:17:35 if you change smthg in the recipe, it is invalidated and rebuilt Jul 03 08:18:03 that's why we don't need to bump PR anymore Jul 03 08:18:16 ant_work: but libqte was rebuilt correctly, it's just a matter of re-installing it into rootfs Jul 03 08:19:58 I think you have to rebuild the full packagegroup the recipe is belonging to Jul 03 08:20:13 now, wich one..I can't remember offhand :p Jul 03 08:21:47 I wonder where it takes old library from Jul 03 09:02:01 anarsoul: is it ok now? Jul 03 09:02:16 ant_work: I'm just rebuilding from scratch Jul 03 09:02:45 ah, haven't you kept the sstate-cache? Jul 03 12:57:38 what is the task that strips debug information from binaries Jul 03 13:22:24 heen why? Jul 03 13:22:44 I think do_package does it Jul 03 14:08:41 have there been some recent changes to the bitbake error handling? Jul 03 14:08:45 I just got this error Jul 03 14:08:51 ERROR: This autoconf log indicates errors, it looked at host include and/or library paths while determining system capabilities. Jul 03 14:08:58 but when looking at the log, it was actually Jul 03 14:09:18 ah, no wait Jul 03 14:09:53 there were two errors, it does also seem to be maybe looking at host include paths Jul 03 14:13:03 that might have been upgraded to ahrd fail Jul 03 14:18:43 Crofton|work: yeah, I think it may have; it's a strange one Jul 03 14:18:58 host includes are evil though Jul 03 14:19:19 it's trying to find the cross-compile version of a needed binary by pre-pending the gcc cross-compile to it, rather than the cross cross compile sysroot path Jul 03 14:19:40 so instead of /cross/compile/sysroot/pcap-config Jul 03 14:19:45 it's going Jul 03 14:19:54 i585-oe-linux-pcap-config Jul 03 14:19:59 the joy... Jul 03 14:22:42 aww Jul 03 15:02:00 Hello Jul 03 15:05:26 Does bb.utils.which intend to locate regular file only or dir also? Jul 03 15:12:42 @vit_: def which(path, item, direction = 0, history = False): """Locate a file in a PATH""" Jul 03 15:13:19 @vit_: see https://github.com/openembedded/bitbake/blob/master/lib/bb/utils.py Jul 03 19:58:52 anarsoul: opie 1.2.5 or git? Jul 03 19:59:06 ant__: git Jul 03 19:59:10 ok Jul 03 19:59:16 building now Jul 03 20:04:31 ant__: build will probably fail due to missing fontconfig dependency Jul 03 20:18:31 well, first comes ERROR: Nothing RPROVIDES 'ntpdate' Jul 03 20:27:23 hm, I've cloned layers repos back in May, so it's something new Jul 03 20:33:02 nm, it is in meta-networking Jul 03 20:33:19 well, strange...ntp is pretty core Jul 03 21:17:35 ant__: I guess my patch is not 100% correct, ts_calibrate calculates median of several samples, I'll just take code from it. Jul 03 21:19:11 anarsoul, there was one patch from methan/cyril doing similat things Jul 03 21:19:37 for qte? Jul 03 21:20:33 hm, no iirc was for the event interface Jul 03 21:22:28 anarsoul, itwas in this git repo Jul 03 21:22:31 http://metan.lindevdoc.org/git/?p=zaurusd.git/.git;a=summary Jul 03 21:22:59 Cyril Hrubis Jul 03 21:23:55 anarsoul, found it Jul 03 21:23:55 https://github.com/lhc4/libevfilter/blob/master/src/evf_profile.h Jul 03 21:25:43 http://atrey.karlin.mff.cuni.cz/~metan/evfilter/ Jul 03 21:27:25 ant__: don't you think that it's overkill? Jul 03 21:27:33 heh Jul 03 21:28:03 I'll test with your patch, hoping the kernel driver is already dejittered Jul 03 21:29:43 I wonder why median filter is necessary for ts_read_raw(), but isn't for ts_read() Jul 03 21:32:16 ah, tslib does dejitter for ts_read() Jul 03 21:33:36 https://bugzilla.kernel.org/show_bug.cgi?id=14434 Jul 03 21:33:42 anarsoul, ^ Jul 03 21:40:17 ant__: s3c2410ts does averaging: https://github.com/anarsoul/linux-2.6/blob/v3.14-anarsoul-asoc-wip/drivers/input/touchscreen/s3c2410_ts.c#L135 Jul 03 21:41:53 so no need for userspace normalizing Jul 03 21:42:21 well, averaging is not a good filter :) Jul 03 22:32:08 anarsoul, indeed weird, maybe http://patchwork.openembedded.org/patch/68923/ Jul 03 22:32:49 we have to ask bluelightning for a fix Jul 03 23:24:52 * darknighte waves before heading out for the long holiday weekend Jul 04 00:11:11 * darknighte waves and really disconnects for real this time. really. **** ENDING LOGGING AT Fri Jul 04 02:59:58 2014